// // Copyright (c) 2017 The Khronos Group Inc. // // Licensed under the Apache License, Version 2.0 (the "License"); // you may not use this file except in compliance with the License. // You may obtain a copy of the License at // // http://www.apache.org/licenses/LICENSE-2.0 // // Unless required by applicable law or agreed to in writing, software // distributed under the License is distributed on an "AS IS" BASIS, // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. // See the License for the specific language governing permissions and // limitations under the License. // #include "harness/compat.h" #include #include #include #include #include "procs.h" #include "harness/testHarness.h" #include "harness/errorHelpers.h" //--- the code for the kernel executables static const char *readKernelCode[] = { "__kernel void testWritef(__global uchar *src, write_only image2d_t dstimg)\n" "{\n" " int tid_x = get_global_id(0);\n" " int tid_y = get_global_id(1);\n" " int indx = tid_y * get_image_width(dstimg) + tid_x;\n" " float4 color;\n" "\n" " indx *= 4;\n" " color = (float4)((float)src[indx+0], (float)src[indx+1], (float)src[indx+2], (float)src[indx+3]);\n" " color /= (float4)(255.f, 255.f, 255.f, 255.f);\n" " write_imagef(dstimg, (int2)(tid_x, tid_y), color);\n" "\n" "}\n", "__kernel void testWritei(__global char *src, write_only image2d_t dstimg)\n" "{\n" " int tid_x = get_global_id(0);\n" " int tid_y = get_global_id(1);\n" " int indx = tid_y * get_image_width(dstimg) + tid_x;\n" " int4 color;\n" "\n" " indx *= 4;\n" " color.x = (int)src[indx+0];\n" " color.y = (int)src[indx+1];\n" " color.z = (int)src[indx+2];\n" " color.w = (int)src[indx+3];\n" " write_imagei(dstimg, (int2)(tid_x, tid_y), color);\n" "\n" "}\n", "__kernel void testWriteui(__global uchar *src, write_only image2d_t dstimg)\n" "{\n" " int tid_x = get_global_id(0);\n" " int tid_y = get_global_id(1);\n" " int indx = tid_y * get_image_width(dstimg) + tid_x;\n" " uint4 color;\n" "\n" " indx *= 4;\n" " color.x = (uint)src[indx+0];\n" " color.y = (uint)src[indx+1];\n" " color.z = (uint)src[indx+2];\n" " color.w = (uint)src[indx+3];\n" " write_imageui(dstimg, (int2)(tid_x, tid_y), color);\n" "\n" "}\n" }; static const char *readKernelName[] = { "testWritef", "testWritei", "testWriteui" }; //--- helper functions static cl_uchar *generateImage( int n, MTdata d ) { cl_uchar *ptr = (cl_uchar *)malloc( n * sizeof( cl_uchar ) ); int i; for( i = 0; i < n; i++ ){ ptr[i] = (cl_uchar)genrand_int32( d ); } return ptr; } static char *generateSignedImage( int n, MTdata d ) { char *ptr = (char *)malloc( n * sizeof( char ) ); int i; for( i = 0; i < n; i++ ){ ptr[i] = (char)genrand_int32( d ); } return ptr; } static int verifyImage( cl_uchar *image, cl_uchar *outptr, int w, int h ) { int i; for( i = 0; i < w * h * 4; i++ ){ if( outptr[i] != image[i] ){ return -1; } } return 0; } //----- the test functions int read_image( cl_device_id device, cl_context context, cl_command_queue queue, int numElements, const char *code, const char *name, cl_image_format image_format_desc ) { cl_mem memobjs[2]; cl_program program[1]; void *inptr; void *dst = NULL; cl_kernel kernel[1]; cl_event readEvent; cl_ulong queueStart, submitStart, readStart, readEnd; size_t threads[2]; int err; int w = 64, h = 64; cl_mem_flags flags; size_t element_nbytes; size_t num_bytes; size_t channel_nbytes = sizeof( cl_uchar ); MTdata d; PASSIVE_REQUIRE_IMAGE_SUPPORT( device ) element_nbytes = channel_nbytes * get_format_channel_count( &image_format_desc ); num_bytes = w * h * element_nbytes; threads[0] = (size_t)w; threads[1] = (size_t)h; d = init_genrand( gRandomSeed ); if( image_format_desc.image_channel_data_type == CL_SIGNED_INT8 ) inptr = (void *)generateSignedImage( w * h * 4, d ); else inptr = (void *)generateImage( w * h * 4, d ); free_mtdata(d); d = NULL; if( ! inptr ){ log_error("unable to allocate inptr at %d x %d\n", (int)w, (int)h ); return -1; } dst = malloc( num_bytes ); if( ! dst ){ free( (void *)inptr ); log_error("unable to allocate dst at %d x %d\n", (int)w, (int)h ); return -1; } // allocate the input and output image memory objects flags = CL_MEM_READ_WRITE; memobjs[0] = create_image_2d( context, flags, &image_format_desc, w, h, 0, NULL, &err ); if( memobjs[0] == (cl_mem)0 ){ free( dst ); free( (void *)inptr ); log_error("unable to create Image2D\n"); return -1; } memobjs[1] = clCreateBuffer(context, CL_MEM_READ_WRITE, channel_nbytes * 4 * w * h, NULL, &err); if( memobjs[1] == (cl_mem)0 ){ free( dst ); free( (void *)inptr ); clReleaseMemObject(memobjs[0]); log_error("unable to create array\n"); return -1; } err = clEnqueueWriteBuffer( queue, memobjs[1], true, 0, num_bytes, inptr, 0, NULL, NULL ); if( err != CL_SUCCESS ){ clReleaseMemObject(memobjs[0]); clReleaseMemObject(memobjs[1]); free( dst ); free( inptr ); log_error("clWriteArray failed\n"); return -1; } err = create_single_kernel_helper( context, &program[0], &kernel[0], 1, &code, name ); if( err ){ log_error( "Unable to create program and kernel\n" ); clReleaseMemObject(memobjs[0]); clReleaseMemObject(memobjs[1]); free( dst ); free( inptr ); return -1; } err = clSetKernelArg( kernel[0], 0, sizeof( cl_mem ), (void *)&memobjs[1] ); err |= clSetKernelArg( kernel[0], 1, sizeof( cl_mem ), (void *)&memobjs[0] ); if( err != CL_SUCCESS ){ log_error( "clSetKernelArg failed\n" ); clReleaseKernel( kernel[0] ); clReleaseProgram( program[0] ); clReleaseMemObject(memobjs[0]); clReleaseMemObject(memobjs[1]); free( dst ); free( inptr ); return -1; } err = clEnqueueNDRangeKernel(queue, kernel[0], 2, NULL, threads, NULL, 0, NULL, NULL ); if( err != CL_SUCCESS ){ print_error( err, "clEnqueueNDRangeKernel failed" ); clReleaseKernel( kernel[0] ); clReleaseProgram( program[0] ); clReleaseMemObject(memobjs[0]); clReleaseMemObject(memobjs[1]); free( dst ); free( inptr ); return -1; } size_t origin[3] = { 0, 0, 0 }; size_t region[3] = { w, h, 1 }; err = clEnqueueReadImage( queue, memobjs[0], false, origin, region, 0, 0, dst, 0, NULL, &readEvent ); if( err != CL_SUCCESS ){ print_error( err, "clReadImage2D failed" ); clReleaseKernel( kernel[0] ); clReleaseProgram( program[0] ); clReleaseMemObject(memobjs[0]); clReleaseMemObject(memobjs[1]); free( dst ); free( inptr ); return -1; } // This synchronization point is needed in order to assume the data is valid. // Getting profiling information is not a synchronization point. err = clWaitForEvents( 1, &readEvent ); if( err != CL_SUCCESS ) { clReleaseEvent(readEvent); clReleaseKernel( kernel[0] ); clReleaseProgram( program[0] ); clReleaseMemObject(memobjs[0]); clReleaseMemObject(memobjs[1]); free( dst ); free( inptr ); return -1; } while( ( err = clGetEventProfilingInfo( readEvent, CL_PROFILING_COMMAND_QUEUED, sizeof( cl_ulong ), &queueStart, NULL ) ) == CL_PROFILING_INFO_NOT_AVAILABLE ); if( err != CL_SUCCESS ){ print_error( err, "clGetEventProfilingInfo failed" ); clReleaseEvent(readEvent); clReleaseKernel( kernel[0] ); clReleaseProgram( program[0] ); clReleaseMemObject(memobjs[0]); clReleaseMemObject(memobjs[1]); free( dst ); free( inptr ); return -1; } while( ( err = clGetEventProfilingInfo( readEvent, CL_PROFILING_COMMAND_SUBMIT, sizeof( cl_ulong ), &submitStart, NULL ) ) == CL_PROFILING_INFO_NOT_AVAILABLE ); if( err != CL_SUCCESS ){ print_error( err, "clGetEventProfilingInfo failed" ); clReleaseEvent(readEvent); clReleaseKernel( kernel[0] ); clReleaseProgram( program[0] ); clReleaseMemObject(memobjs[0]); clReleaseMemObject(memobjs[1]); free( dst ); free( inptr ); return -1; } err = clGetEventProfilingInfo( readEvent, CL_PROFILING_COMMAND_START, sizeof( cl_ulong ), &readStart, NULL ); if( err != CL_SUCCESS ){ print_error( err, "clGetEventProfilingInfo failed" ); clReleaseEvent(readEvent); clReleaseKernel( kernel[0] ); clReleaseProgram( program[0] ); clReleaseMemObject(memobjs[0]); clReleaseMemObject(memobjs[1]); free( dst ); free( inptr ); return -1; } err = clGetEventProfilingInfo( readEvent, CL_PROFILING_COMMAND_END, sizeof( cl_ulong ), &readEnd, NULL ); if( err != CL_SUCCESS ){ print_error( err, "clGetEventProfilingInfo failed" ); clReleaseEvent(readEvent); clReleaseKernel( kernel[0] ); clReleaseProgram( program[0] ); clReleaseMemObject(memobjs[0]); clReleaseMemObject(memobjs[1]); free( dst ); free( inptr ); return -1; } err = verifyImage( (cl_uchar *)inptr, (cl_uchar *)dst, w, h ); if( err ){ log_error( "Image failed to verify.\n" ); } else{ log_info( "Image verified.\n" ); } clReleaseEvent(readEvent); clReleaseKernel(kernel[0]); clReleaseProgram(program[0]); clReleaseMemObject(memobjs[0]); clReleaseMemObject(memobjs[1]); free(dst); free(inptr); if (check_times(queueStart, submitStart, readStart, readEnd, device)) err = -1; return err; } // end read_image() int test_read_image_float( cl_device_id device, cl_context context, cl_command_queue queue, int numElements ) { cl_image_format image_format_desc = { CL_RGBA, CL_UNORM_INT8 }; PASSIVE_REQUIRE_IMAGE_SUPPORT( device ) // 0 to 255 for unsigned image data return read_image( device, context, queue, numElements, readKernelCode[0], readKernelName[0], image_format_desc ); } int test_read_image_char( cl_device_id device, cl_context context, cl_command_queue queue, int numElements ) { cl_image_format image_format_desc = { CL_RGBA, CL_SIGNED_INT8 }; PASSIVE_REQUIRE_IMAGE_SUPPORT( device ) // -128 to 127 for signed iamge data return read_image( device, context, queue, numElements, readKernelCode[1], readKernelName[1], image_format_desc ); } int test_read_image_uchar( cl_device_id device, cl_context context, cl_command_queue queue, int numElements ) { cl_image_format image_format_desc = { CL_RGBA, CL_UNSIGNED_INT8 }; PASSIVE_REQUIRE_IMAGE_SUPPORT( device ) // 0 to 255 for unsigned image data return read_image( device, context, queue, numElements, readKernelCode[2], readKernelName[2], image_format_desc ); }