You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

94 lines
3.7 KiB

-------------------------------------------------------------------------
drawElements Quality Program Test Specification
-----------------------------------------------
Copyright 2014 The Android Open Source Project
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.
-------------------------------------------------------------------------
GL 3 conditional render
Tests:
+ dEQP-GL3.functional.conditional_render.*
Includes:
+ All valid rendering commands included/excluded by query
- DrawArrays
- DrawArraysInstanced
- MultiDrawArrays
- DrawElements
- DrawElementsInstanced
- DrawElementsInstancedBaseVertex
- DrawElementsBaseVertex
- DrawRangeElements
- DrawRangeElementsBaseVertex
- MultiDrawElements
- MultiDrawElementsBaseVertex
- Clear
- ClearBuffer*
+ All depth comparison modes
+ Framebuffer swap between query & render
- Excluding BY_REGION query modes
+ All modes
- QUERY_(BY_REGION_)(NO_)WAIT
+ Dependent queries
Excludes:
+ Immediate mode rendering
+ Primitives other than triangles
+ Negative tests
+ Tests for NO_WAIT and BY_REGION modes are very limited
Description:
Conditional rendering is tested with various depth function/primitive
overlap conditions. Each query test renders a scene with one conditional
render call and verifies the result against the same scene with the
previously GL controlled conditional render call controlled by the test.
Draw calls (excluding clears) are tested with the same method, with the
conditional call indicated by the test name. The occlusion query used in
these tests is trivially true/false.
Clears are tested by rendering some geometry for an occlusion query and then
conditionally clearing the framebuffer. The verification used depends on the
expected result of the occlusion query. If the query is expected to pass, the
framebuffer is verified to contain only the color it was cleared to.
Otherwise the contents of the framebuffer are verified to match its contents
before the conditional clear.
Depth/stencil clear tests unconditionally clear the color buffer after
rendering the occlusion geometry. After the conditional depth/stencil clear
a full framebuffer triangle is rendered with depth/stencil tests setup
to pass if the clear took place. The resulting framebuffer is read out and
compared to the color of the intermediate clear or the filling triangle as
appropriate for the expected occlusion query result.
All clear tests use non-default framebuffers, all others use the default.
Framebuffer swap cases use a trivial occlusion query but change the bound
(read & write) FBO between the occludion query and the conditional rendering.
The occlusion query is performed with the default framebuffer.
Wait mode test render geometry that is not dependent on the result of the
occlusion queries. These tests thus only test for false positive occlusion
query results.
Dependent queries render a number of primitives where each conditionally
rendered primitive (excluding the first) is also used for an occlusion query
that controls the rendering of the next primitive, which in turn is used for
the next occlusion query and so on. In some of these test one primitive in
this chain is placed in an occluded location, starting a cascade of failing
occlusion queries.