6660 lines
234 KiB
6660 lines
234 KiB
// Copyright (c) Facebook, Inc. and its affiliates.
|
|
// All rights reserved.
|
|
//
|
|
// Copyright 2020 Google LLC
|
|
//
|
|
// This source code is licensed under the BSD-style license found in the
|
|
// LICENSE file in the root directory of this source tree.
|
|
//
|
|
// Auto-generated file. Do not edit!
|
|
// Specification: test/f32-pavgpool-minmax.yaml
|
|
// Generator: tools/generate-avgpool-test.py
|
|
|
|
|
|
#include <gtest/gtest.h>
|
|
|
|
#include <xnnpack/common.h>
|
|
#include <xnnpack/isa-checks.h>
|
|
|
|
#include <xnnpack/avgpool.h>
|
|
#include <xnnpack/pavgpool.h>
|
|
#include "avgpool-microkernel-tester.h"
|
|
|
|
|
|
#if XNN_ARCH_ARM || XNN_ARCH_ARM64
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_eq_4_twopass_fulltile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_eq_4_twopass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_eq_4_twopass_fulltile_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_eq_4_twopass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_eq_4_twopass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_eq_4_twopass_subtile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_eq_4_twopass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_eq_4_twopass_subtile_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_div_4_twopass_fulltile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_div_4_twopass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_div_4_twopass_fulltile_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_div_4_twopass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_div_4_twopass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_div_4_twopass_subtile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_div_4_twopass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_div_4_twopass_subtile_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_lt_4_twopass_fulltile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_lt_4_twopass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_lt_4_twopass_fulltile_with_zero_index) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_lt_4_twopass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_lt_4_twopass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_lt_4_twopass_subtile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_lt_4_twopass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_lt_4_twopass_subtile_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_gt_4_twopass_fulltile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_gt_4_twopass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_gt_4_twopass_fulltile_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_gt_4_twopass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_gt_4_twopass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_gt_4_twopass_subtile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_gt_4_twopass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_gt_4_twopass_subtile_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_eq_4_multipass) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_eq_4_multipass_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_eq_4_multipass_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_eq_4_multipass_with_qmin) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_eq_4_multipass_with_qmax) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_div_4_multipass) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_div_4_multipass_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_div_4_multipass_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_div_4_multipass_with_qmin) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_div_4_multipass_with_qmax) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_lt_4_multipass) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_lt_4_multipass_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_lt_4_multipass_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(4)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_lt_4_multipass_with_qmin) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_lt_4_multipass_with_qmax) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_gt_4_multipass) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_gt_4_multipass_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_gt_4_multipass_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_gt_4_multipass_with_qmin) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, channels_gt_4_multipass_with_qmax) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, few_output_pixels) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, few_output_pixels_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, few_output_pixels_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, few_output_pixels_with_qmin) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, few_output_pixels_with_qmax) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, few_output_pixels_with_output_stride) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__NEON_C4, few_output_pixels_with_step) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t step = 2; step <= pooling_elements; step++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.step(step)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
#endif // XNN_ARCH_ARM || XNN_ARCH_ARM64
|
|
|
|
|
|
#if XNN_ARCH_X86 || XNN_ARCH_X86_64
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_eq_4_twopass_fulltile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_eq_4_twopass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_eq_4_twopass_fulltile_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_eq_4_twopass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_eq_4_twopass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_eq_4_twopass_subtile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_eq_4_twopass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_eq_4_twopass_subtile_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_div_4_twopass_fulltile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_div_4_twopass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_div_4_twopass_fulltile_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_div_4_twopass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_div_4_twopass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_div_4_twopass_subtile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_div_4_twopass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_div_4_twopass_subtile_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_lt_4_twopass_fulltile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_lt_4_twopass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_lt_4_twopass_fulltile_with_zero_index) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_lt_4_twopass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_lt_4_twopass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_lt_4_twopass_subtile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_lt_4_twopass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_lt_4_twopass_subtile_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_gt_4_twopass_fulltile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_gt_4_twopass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_gt_4_twopass_fulltile_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_gt_4_twopass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_gt_4_twopass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_gt_4_twopass_subtile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_gt_4_twopass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_gt_4_twopass_subtile_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_eq_4_multipass) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_eq_4_multipass_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_eq_4_multipass_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_eq_4_multipass_with_qmin) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_eq_4_multipass_with_qmax) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_div_4_multipass) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_div_4_multipass_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_div_4_multipass_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_div_4_multipass_with_qmin) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_div_4_multipass_with_qmax) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_lt_4_multipass) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_lt_4_multipass_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_lt_4_multipass_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(4)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_lt_4_multipass_with_qmin) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_lt_4_multipass_with_qmax) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_gt_4_multipass) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_gt_4_multipass_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_gt_4_multipass_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_gt_4_multipass_with_qmin) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, channels_gt_4_multipass_with_qmax) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, few_output_pixels) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, few_output_pixels_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, few_output_pixels_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, few_output_pixels_with_qmin) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, few_output_pixels_with_qmax) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, few_output_pixels_with_output_stride) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SSE_C4, few_output_pixels_with_step) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t step = 2; step <= pooling_elements; step++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.step(step)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
#endif // XNN_ARCH_X86 || XNN_ARCH_X86_64
|
|
|
|
|
|
#if XNN_ARCH_WASMSIMD
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_eq_4_twopass_fulltile) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_eq_4_twopass_fulltile_with_input_offset) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_eq_4_twopass_fulltile_with_zero) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_eq_4_twopass_fulltile_with_qmin) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_eq_4_twopass_fulltile_with_qmax) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_eq_4_twopass_subtile) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_eq_4_twopass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_eq_4_twopass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_div_4_twopass_fulltile) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_div_4_twopass_fulltile_with_input_offset) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_div_4_twopass_fulltile_with_zero) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_div_4_twopass_fulltile_with_qmin) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_div_4_twopass_fulltile_with_qmax) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_div_4_twopass_subtile) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_div_4_twopass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_div_4_twopass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_lt_4_twopass_fulltile) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_lt_4_twopass_fulltile_with_input_offset) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_lt_4_twopass_fulltile_with_zero_index) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_lt_4_twopass_fulltile_with_qmin) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_lt_4_twopass_fulltile_with_qmax) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_lt_4_twopass_subtile) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_lt_4_twopass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_lt_4_twopass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_gt_4_twopass_fulltile) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_gt_4_twopass_fulltile_with_input_offset) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_gt_4_twopass_fulltile_with_zero) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_gt_4_twopass_fulltile_with_qmin) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_gt_4_twopass_fulltile_with_qmax) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_gt_4_twopass_subtile) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_gt_4_twopass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_gt_4_twopass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_eq_4_multipass) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_eq_4_multipass_with_input_offset) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_eq_4_multipass_with_zero) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_eq_4_multipass_with_qmin) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_eq_4_multipass_with_qmax) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_div_4_multipass) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_div_4_multipass_with_input_offset) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_div_4_multipass_with_zero) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_div_4_multipass_with_qmin) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_div_4_multipass_with_qmax) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_lt_4_multipass) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_lt_4_multipass_with_input_offset) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_lt_4_multipass_with_zero) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(4)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_lt_4_multipass_with_qmin) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_lt_4_multipass_with_qmax) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_gt_4_multipass) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_gt_4_multipass_with_input_offset) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_gt_4_multipass_with_zero) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_gt_4_multipass_with_qmin) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, channels_gt_4_multipass_with_qmax) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, few_output_pixels) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, few_output_pixels_with_input_offset) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, few_output_pixels_with_zero) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, few_output_pixels_with_qmin) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, few_output_pixels_with_qmax) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, few_output_pixels_with_output_stride) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_ARM_C4, few_output_pixels_with_step) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t step = 2; step <= pooling_elements; step++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.step(step)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
#endif // XNN_ARCH_WASMSIMD
|
|
|
|
|
|
#if XNN_ARCH_WASMSIMD
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_eq_4_twopass_fulltile) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_eq_4_twopass_fulltile_with_input_offset) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_eq_4_twopass_fulltile_with_zero) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_eq_4_twopass_fulltile_with_qmin) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_eq_4_twopass_fulltile_with_qmax) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_eq_4_twopass_subtile) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_eq_4_twopass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_eq_4_twopass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_div_4_twopass_fulltile) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_div_4_twopass_fulltile_with_input_offset) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_div_4_twopass_fulltile_with_zero) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_div_4_twopass_fulltile_with_qmin) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_div_4_twopass_fulltile_with_qmax) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_div_4_twopass_subtile) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_div_4_twopass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_div_4_twopass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_lt_4_twopass_fulltile) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_lt_4_twopass_fulltile_with_input_offset) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_lt_4_twopass_fulltile_with_zero_index) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_lt_4_twopass_fulltile_with_qmin) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_lt_4_twopass_fulltile_with_qmax) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_lt_4_twopass_subtile) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_lt_4_twopass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_lt_4_twopass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_gt_4_twopass_fulltile) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_gt_4_twopass_fulltile_with_input_offset) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_gt_4_twopass_fulltile_with_zero) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_gt_4_twopass_fulltile_with_qmin) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_gt_4_twopass_fulltile_with_qmax) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_gt_4_twopass_subtile) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_gt_4_twopass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_gt_4_twopass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_eq_4_multipass) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_eq_4_multipass_with_input_offset) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_eq_4_multipass_with_zero) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_eq_4_multipass_with_qmin) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_eq_4_multipass_with_qmax) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(4)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_div_4_multipass) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_div_4_multipass_with_input_offset) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_div_4_multipass_with_zero) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_div_4_multipass_with_qmin) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_div_4_multipass_with_qmax) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_lt_4_multipass) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_lt_4_multipass_with_input_offset) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_lt_4_multipass_with_zero) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(4)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_lt_4_multipass_with_qmin) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_lt_4_multipass_with_qmax) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_gt_4_multipass) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_gt_4_multipass_with_input_offset) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_gt_4_multipass_with_zero) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_gt_4_multipass_with_qmin) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, channels_gt_4_multipass_with_qmax) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, few_output_pixels) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, few_output_pixels_with_input_offset) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, few_output_pixels_with_zero) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, few_output_pixels_with_qmin) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, few_output_pixels_with_qmax) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, few_output_pixels_with_output_stride) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASMSIMD_X86_C4, few_output_pixels_with_step) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t step = 2; step <= pooling_elements; step++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.step(step)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
#endif // XNN_ARCH_WASMSIMD
|
|
|
|
|
|
#if XNN_ARCH_WASM || XNN_ARCH_WASMSIMD
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_eq_1_twopass_fulltile) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_eq_1_twopass_fulltile_with_input_offset) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_eq_1_twopass_fulltile_with_zero) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_eq_1_twopass_fulltile_with_qmin) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_eq_1_twopass_fulltile_with_qmax) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_eq_1_twopass_subtile) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_eq_1_twopass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_eq_1_twopass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_gt_1_twopass_fulltile) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_gt_1_twopass_fulltile_with_input_offset) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_gt_1_twopass_fulltile_with_zero) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_gt_1_twopass_fulltile_with_qmin) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_gt_1_twopass_fulltile_with_qmax) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_gt_1_twopass_subtile) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_gt_1_twopass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_gt_1_twopass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_eq_1_multipass) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_eq_1_multipass_with_input_offset) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_eq_1_multipass_with_zero) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_eq_1_multipass_with_qmin) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_eq_1_multipass_with_qmax) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_gt_1_multipass) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_gt_1_multipass_with_input_offset) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_gt_1_multipass_with_zero) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_gt_1_multipass_with_qmin) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, channels_gt_1_multipass_with_qmax) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, few_output_pixels) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, few_output_pixels_with_input_offset) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, few_output_pixels_with_zero) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, few_output_pixels_with_qmin) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, few_output_pixels_with_qmax) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, few_output_pixels_with_output_stride) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.output_stride(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__WASM_C1, few_output_pixels_with_step) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
for (size_t step = 2; step <= pooling_elements; step++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.step(step)
|
|
.channels(channels)
|
|
.output_stride(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
#endif // XNN_ARCH_WASM || XNN_ARCH_WASMSIMD
|
|
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_eq_1_twopass_fulltile) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_eq_1_twopass_fulltile_with_input_offset) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_eq_1_twopass_fulltile_with_zero) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_eq_1_twopass_fulltile_with_qmin) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_eq_1_twopass_fulltile_with_qmax) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_eq_1_twopass_subtile) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_eq_1_twopass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_eq_1_twopass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_gt_1_twopass_fulltile) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_gt_1_twopass_fulltile_with_input_offset) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_gt_1_twopass_fulltile_with_zero) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 17; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_gt_1_twopass_fulltile_with_qmin) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_gt_1_twopass_fulltile_with_qmax) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(17)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_gt_1_twopass_subtile) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_gt_1_twopass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_gt_1_twopass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 10; pooling_elements < 17; pooling_elements++) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_eq_1_multipass) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_eq_1_multipass_with_input_offset) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_eq_1_multipass_with_zero) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_eq_1_multipass_with_qmin) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_eq_1_multipass_with_qmax) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(1)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_gt_1_multipass) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_gt_1_multipass_with_input_offset) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_gt_1_multipass_with_zero) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_gt_1_multipass_with_qmin) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, channels_gt_1_multipass_with_qmax) {
|
|
for (size_t pooling_elements = 18; pooling_elements <= 33; pooling_elements += 3) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, few_output_pixels) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, few_output_pixels_with_input_offset) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, few_output_pixels_with_zero) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, few_output_pixels_with_qmin) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, few_output_pixels_with_qmax) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, few_output_pixels_with_output_stride) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.channels(channels)
|
|
.output_stride(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9P8X__SCALAR_C1, few_output_pixels_with_step) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{10, 16, 18}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
for (size_t step = 2; step <= pooling_elements; step++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 8)
|
|
.step(step)
|
|
.channels(channels)
|
|
.output_stride(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9p8x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
#if XNN_ARCH_ARM || XNN_ARCH_ARM64
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_eq_4_unipass_fulltile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_eq_4_unipass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_eq_4_unipass_fulltile_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_eq_4_unipass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_eq_4_unipass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_eq_4_unipass_subtile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_eq_4_unipass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_eq_4_unipass_subtile_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_div_4_unipass_fulltile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_div_4_unipass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_div_4_unipass_fulltile_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_div_4_unipass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_div_4_unipass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_div_4_unipass_subtile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_div_4_unipass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_div_4_unipass_subtile_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_lt_4_unipass_fulltile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_lt_4_unipass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_lt_4_unipass_fulltile_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_lt_4_unipass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_lt_4_unipass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_lt_4_unipass_subtile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_lt_4_unipass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_lt_4_unipass_subtile_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_gt_4_unipass_fulltile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_gt_4_unipass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_gt_4_unipass_fulltile_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_gt_4_unipass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_gt_4_unipass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_gt_4_unipass_subtile) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_gt_4_unipass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, channels_gt_4_unipass_subtile_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, few_output_pixels) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, few_output_pixels_with_input_offset) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, few_output_pixels_with_zero) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, few_output_pixels_with_qmin) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, few_output_pixels_with_qmax) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, few_output_pixels_with_output_stride) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__NEON_C4, few_output_pixels_with_step) {
|
|
TEST_REQUIRES_ARM_NEON;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t step = 2; step <= pooling_elements; step++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.step(step)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__neon_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
#endif // XNN_ARCH_ARM || XNN_ARCH_ARM64
|
|
|
|
|
|
#if XNN_ARCH_X86 || XNN_ARCH_X86_64
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_eq_4_unipass_fulltile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_eq_4_unipass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_eq_4_unipass_fulltile_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_eq_4_unipass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_eq_4_unipass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_eq_4_unipass_subtile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_eq_4_unipass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_eq_4_unipass_subtile_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_div_4_unipass_fulltile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_div_4_unipass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_div_4_unipass_fulltile_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_div_4_unipass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_div_4_unipass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_div_4_unipass_subtile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_div_4_unipass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_div_4_unipass_subtile_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_lt_4_unipass_fulltile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_lt_4_unipass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_lt_4_unipass_fulltile_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_lt_4_unipass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_lt_4_unipass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_lt_4_unipass_subtile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_lt_4_unipass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_lt_4_unipass_subtile_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_gt_4_unipass_fulltile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_gt_4_unipass_fulltile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_gt_4_unipass_fulltile_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_gt_4_unipass_fulltile_with_qmin) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_gt_4_unipass_fulltile_with_qmax) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_gt_4_unipass_subtile) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_gt_4_unipass_subtile_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, channels_gt_4_unipass_subtile_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, few_output_pixels) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, few_output_pixels_with_input_offset) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, few_output_pixels_with_zero) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, few_output_pixels_with_qmin) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, few_output_pixels_with_qmax) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, few_output_pixels_with_output_stride) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SSE_C4, few_output_pixels_with_step) {
|
|
TEST_REQUIRES_X86_SSE;
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t step = 2; step <= pooling_elements; step++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.step(step)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__sse_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
#endif // XNN_ARCH_X86 || XNN_ARCH_X86_64
|
|
|
|
|
|
#if XNN_ARCH_WASMSIMD
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_eq_4_unipass_fulltile) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_eq_4_unipass_fulltile_with_input_offset) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_eq_4_unipass_fulltile_with_zero) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_eq_4_unipass_fulltile_with_qmin) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_eq_4_unipass_fulltile_with_qmax) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_eq_4_unipass_subtile) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_eq_4_unipass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_eq_4_unipass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_div_4_unipass_fulltile) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_div_4_unipass_fulltile_with_input_offset) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_div_4_unipass_fulltile_with_zero) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_div_4_unipass_fulltile_with_qmin) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_div_4_unipass_fulltile_with_qmax) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_div_4_unipass_subtile) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_div_4_unipass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_div_4_unipass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_lt_4_unipass_fulltile) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_lt_4_unipass_fulltile_with_input_offset) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_lt_4_unipass_fulltile_with_zero) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_lt_4_unipass_fulltile_with_qmin) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_lt_4_unipass_fulltile_with_qmax) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_lt_4_unipass_subtile) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_lt_4_unipass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_lt_4_unipass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_gt_4_unipass_fulltile) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_gt_4_unipass_fulltile_with_input_offset) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_gt_4_unipass_fulltile_with_zero) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_gt_4_unipass_fulltile_with_qmin) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_gt_4_unipass_fulltile_with_qmax) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_gt_4_unipass_subtile) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_gt_4_unipass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, channels_gt_4_unipass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, few_output_pixels) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, few_output_pixels_with_input_offset) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, few_output_pixels_with_zero) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, few_output_pixels_with_qmin) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, few_output_pixels_with_qmax) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, few_output_pixels_with_output_stride) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_ARM_C4, few_output_pixels_with_step) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t step = 2; step <= pooling_elements; step++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.step(step)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_arm_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
#endif // XNN_ARCH_WASMSIMD
|
|
|
|
|
|
#if XNN_ARCH_WASMSIMD
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_eq_4_unipass_fulltile) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_eq_4_unipass_fulltile_with_input_offset) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_eq_4_unipass_fulltile_with_zero) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_eq_4_unipass_fulltile_with_qmin) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_eq_4_unipass_fulltile_with_qmax) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_eq_4_unipass_subtile) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_eq_4_unipass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_eq_4_unipass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(4)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_div_4_unipass_fulltile) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_div_4_unipass_fulltile_with_input_offset) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_div_4_unipass_fulltile_with_zero) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_div_4_unipass_fulltile_with_qmin) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_div_4_unipass_fulltile_with_qmax) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_div_4_unipass_subtile) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_div_4_unipass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_div_4_unipass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 8; channels < 32; channels += 4) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(37)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_lt_4_unipass_fulltile) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_lt_4_unipass_fulltile_with_input_offset) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_lt_4_unipass_fulltile_with_zero) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_lt_4_unipass_fulltile_with_qmin) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_lt_4_unipass_fulltile_with_qmax) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_lt_4_unipass_subtile) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_lt_4_unipass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_lt_4_unipass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 1; channels < 4; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(5)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_gt_4_unipass_fulltile) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_gt_4_unipass_fulltile_with_input_offset) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_gt_4_unipass_fulltile_with_zero) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_gt_4_unipass_fulltile_with_qmin) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_gt_4_unipass_fulltile_with_qmax) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_gt_4_unipass_subtile) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_gt_4_unipass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, channels_gt_4_unipass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 5; channels < 8; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(11)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, few_output_pixels) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, few_output_pixels_with_input_offset) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, few_output_pixels_with_zero) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.input_offset(23)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, few_output_pixels_with_qmin) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, few_output_pixels_with_qmax) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, few_output_pixels_with_output_stride) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASMSIMD_X86_C4, few_output_pixels_with_step) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 20; channels += 3) {
|
|
for (size_t step = 2; step <= pooling_elements; step++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.step(step)
|
|
.channels(channels)
|
|
.output_stride(23)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasmsimd_x86_c4);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
#endif // XNN_ARCH_WASMSIMD
|
|
|
|
|
|
#if XNN_ARCH_WASM || XNN_ARCH_WASMSIMD
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_eq_1_unipass_fulltile) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_eq_1_unipass_fulltile_with_input_offset) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_eq_1_unipass_fulltile_with_zero) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_eq_1_unipass_fulltile_with_qmin) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_eq_1_unipass_fulltile_with_qmax) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_eq_1_unipass_subtile) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_eq_1_unipass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_eq_1_unipass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_gt_1_unipass_fulltile) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_gt_1_unipass_fulltile_with_input_offset) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_gt_1_unipass_fulltile_with_zero) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_gt_1_unipass_fulltile_with_qmin) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_gt_1_unipass_fulltile_with_qmax) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_gt_1_unipass_subtile) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_gt_1_unipass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, channels_gt_1_unipass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, few_output_pixels) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, few_output_pixels_with_input_offset) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, few_output_pixels_with_zero) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, few_output_pixels_with_qmin) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, few_output_pixels_with_qmax) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, few_output_pixels_with_output_stride) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.output_stride(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__WASM_C1, few_output_pixels_with_step) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
for (size_t step = 2; step <= pooling_elements; step++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.step(step)
|
|
.channels(channels)
|
|
.output_stride(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__wasm_c1);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
#endif // XNN_ARCH_WASM || XNN_ARCH_WASMSIMD
|
|
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_eq_1_unipass_fulltile) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_eq_1_unipass_fulltile_with_input_offset) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_eq_1_unipass_fulltile_with_zero) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_eq_1_unipass_fulltile_with_qmin) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_eq_1_unipass_fulltile_with_qmax) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_eq_1_unipass_subtile) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_eq_1_unipass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_eq_1_unipass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(1)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_gt_1_unipass_fulltile) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_gt_1_unipass_fulltile_with_input_offset) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_gt_1_unipass_fulltile_with_zero) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
for (size_t zero_index = 0; zero_index < 9; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_gt_1_unipass_fulltile_with_qmin) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_gt_1_unipass_fulltile_with_qmax) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(9)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_gt_1_unipass_subtile) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_gt_1_unipass_subtile_with_input_offset) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, channels_gt_1_unipass_subtile_with_zero) {
|
|
for (size_t pooling_elements = 2; pooling_elements < 9; pooling_elements++) {
|
|
for (size_t channels = 2; channels < 10; channels++) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9)
|
|
.channels(channels)
|
|
.input_offset(3)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, few_output_pixels) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, few_output_pixels_with_input_offset) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.input_offset(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, few_output_pixels_with_zero) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
for (size_t zero_index = 0; zero_index < pooling_elements; zero_index++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.input_offset(7)
|
|
.zero_index(zero_index)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, few_output_pixels_with_qmin) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.qmin(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, few_output_pixels_with_qmax) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.qmax(128)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, few_output_pixels_with_output_stride) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.channels(channels)
|
|
.output_stride(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
TEST(F32_PAVGPOOL_MINMAX_9X__SCALAR_C1, few_output_pixels_with_step) {
|
|
for (size_t output_pixels = 2; output_pixels <= 5; output_pixels++) {
|
|
for (size_t pooling_elements : std::vector<size_t>{{2, 8, 9}}) {
|
|
for (size_t channels = 1; channels <= 5; channels += 1) {
|
|
for (size_t step = 2; step <= pooling_elements; step++) {
|
|
AvgPoolMicrokernelTester()
|
|
.output_pixels(output_pixels)
|
|
.pooling_elements(pooling_elements)
|
|
.pooling_tile(9, 0)
|
|
.step(step)
|
|
.channels(channels)
|
|
.output_stride(7)
|
|
.Test(xnn_f32_pavgpool_minmax_ukernel_9x__scalar_c1, AvgPoolMicrokernelTester::Variant::Scalar);
|
|
}
|
|
}
|
|
}
|
|
}
|
|
} |