Hash :
665ddccd
Author :
Date :
2021-12-09T23:06:39
Vulkan: Emulate dithering Dithering in OpenGL is vaguely defined, to the extent that no dithering is also a valid dithering algorithm. Dithering is enabled by default, but emulating it has a non-negligible cost. Similarly to some other GLES drivers, ANGLE enables dithering only on low-bit formats where visual banding is particularly common; namely RGBA4444, RGBA5551 and RGB565. Dithering is emulated in the fragment shader and is controlled by a spec constant. Every 2 bits of the spec constant correspond to one attachment, with the value indicating: - 00: No dithering - 01: Dither for RGBA4444 - 10: Dither for RGBA5551 - 11: Dither for RGB565 The translator appends code to the shader that, based on the format specified by the specialization constant, adds dithering to each color attachment output. A 2x2 Bayer matrix is used for dithering, indexed by gl_FragCoord.xy % 2. Bug: angleproject:6755 Change-Id: Ib45da5938e299b6626bff921119d63e7357dd353 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3374261 Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
ANGLE integrates dEQP (i.e. the OpenGL CTS) for conformance testing. It uses gtest to run tests, and provides the means for dEQP to use ANGLE.
Occasionally, ANGLE overrides certain dEQP files by copying them to this directory, adding the
_override suffix, and modifying them. deqp.gni is used to select these override files to be
built with dEQP instead of the original files.
This is primarily done to fix tests until they are fixed upstream.
For every set of dEQP tests, for example GLES3 tests on the Vulkan backend, an expectations file exists to let the test harness know which tests it should skip (as they are known to crash), or expect to see failed. Warnings are generated if a test unexpectedly passes, but an unexpected failure is an error. This let’s ANGLE ensure there are no regressions.
If multiple test expectations in a file match a specific test due to wildcards, the test harness picks the first match in the file as the overriding expectation.
While developing a feature, or testing on a new platform, the expectations files can be modified to reflect the reality of the situation. The expected format for every line in these files is:
{BUG#} {MODIFIERS} : {TEST_NAME} = {PASS,FAIL,FLAKY,TIMEOUT,SKIP}
MODIFIERS can be a combination of the below list, combined with a logical AND:
WIN XP VISTA WIN7 WIN8 WIN10
MAC LEOPARD SNOWLEOPARD LION MOUNTAINLION MAVERICKS YOSEMITE ELCAPITAN SIERRA HIGHSIERRA MOJAVE
LINUX CHROMEOS ANDROID
NVIDIA AMD INTEL
DEBUG RELEASE
D3D9 D3D11 OPENGL GLES VULKAN
NEXUS5X PIXEL2ORXL PIXEL4ORXL PIXEL6
QUADROP400
SWIFTSHADER
PREROTATION PREROTATION90 PREROTATION180 PREROTATION270
NOSAN ASAN TSAN UBSAN
TEST_NAME can be a specific test name, or set of test names using '*' as wildcard anywhere in
the name. Examples:
// Disabled everywhere as is too slow:
3445 : dEQP-GLES31.functional.ssbo.layout.random.all_shared_buffer.48 = SKIP
// Crashes on both D3D11 and OPENGL:
1442 OPENGL : dEQP-GLES31.functional.separate_shader.* = SKIP
1442 D3D11 : dEQP-GLES31.functional.separate_shader.* = SKIP
// Unsupported feature:
3726 VULKAN ANDROID : dEQP-GLES31.functional.synchronization.inter_call.without_memory_barrier.*atomic_counter* = FAIL
// Failing test in Nvidia's OpenGL implementation on windows:
1665 WIN NVIDIA OPENGL : dEQP-GLES31.functional.draw_indirect.negative.command_offset_not_in_buffer_unsigned32_wrap = FAIL
// Failing when emulated pre-rotation is enabled, no matter which angle:
1234 PREROTATION : dEQP-GLES3.*blit* = FAIL
// Failing when emulated pre-rotation is enabled with 270 degree angle:
1234 PREROTATION270 : dEQP-GLES3.*blit* = FAIL
// Flaky when run with thread-sanitizer (TSan)
6678 TSAN : dEQP-EGL.functional.sharing.gles2.multithread.random* = FLAKY