- 15 Jun, 2015 2 commits
-
-
Austin Kinross authored
Change-Id: I9b93cc6ee7218a92dda80dbc3f70aec6fef69ded Reviewed-on: https://chromium-review.googlesource.com/276598Tested-by:
Austin Kinross <aukinros@microsoft.com> Reviewed-by:
Geoff Lang <geofflang@chromium.org>
-
Cooper Partin authored
BUG=angleproject:1005 Change-Id: I8f2a6c0a5a6157303e436b96cc4c4ff1c29cfe18 Reviewed-on: https://chromium-review.googlesource.com/271600Reviewed-by:
Jamie Madill <jmadill@chromium.org> Reviewed-by:
-
- 12 Jun, 2015 8 commits
-
-
Jamie Madill authored
We can also defer loading InputLayouts and other resources in Blit11 if we don't need them immediately. BUG=angleproject:1014 Change-Id: I859583e4696e5864dfd8d436143ed9ddbc2fd60d Reviewed-on: https://chromium-review.googlesource.com/275779Reviewed-by:
Brandon Jones <bajones@chromium.org> Tested-by:
Kenneth Russell <kbr@chromium.org>
-
Jamie Madill authored
Introduce a 'DeferredShader' utility class. Other classes can use DeferredShader to delay loading shader resources until we render with them, which saves on loading time as well as memory, if we aren't using particular shaders. BUG=angleproject:1014 Change-Id: Ic6767c3c422a7fedbf23cce5d0c9d822aaf2e652 Reviewed-on: https://chromium-review.googlesource.com/275778Tested-by:
-
Jamie Madill authored
Saves us some time on startup. Also assume every multisample format can support 1x multisampling, and that if we support a higher format we support all lower sample counts (eg - if we support 16x we support 8x and 4x, etc) BUG=angleproject:1014 Change-Id: I62143e5db561b717edd1a0c96f5e3acd4de3d154 Reviewed-on: https://chromium-review.googlesource.com/275777Reviewed-by:
-
Austin Kinross authored
BUG=angleproject:1002 The last patch added new tests which failed on some Intel Win7 machines. This was due to a driver bug. Windows Updates have been applied to the failing machine, and the tests now pass. Changes between second and current (third) failed 16-bit patch: - Reenable usage of 16-bit DXGI textures on all Intel cards - Disable 16bpp tests on Intel D3D11 until the bots are updated - Updated Win7 test machine. Tests now pass Changes between first and second failed 16-bit patch: - Disabled usage of 16-bit DXGI textures on all Intel cards - Added DXGI->GL conversion entries for 16-bit formats, which fixes FBO readback issues. The ES CTS and dEQP tests which failed with the last patch now pass. Change-Id: Id8ea89bfbc450beea54afdd398956bc97ecf3c01 Reviewed-on: https://chromium-review.googlesource.com/275082Reviewed-by:
-
Jamie Madill authored
This will save us some queries on D3D11 startup. BUG=angleproject:1014 Change-Id: I5aaeec43d3b9bc73d011de17e2abac902ac3b7b9 Reviewed-on: https://chromium-review.googlesource.com/275776Tested-by:
-
Jamie Madill authored
This will save us some time on D3D11 startup. BUG=angleproject:1014 Change-Id: Ic00b3dfd9f6230d2d54e6e9bb74be55b923697fe Reviewed-on: https://chromium-review.googlesource.com/275775Tested-by:
-
Jamie Madill authored
This saves us some startup time in the D3D11 Renderer. BUG=angleproject:1014 Change-Id: I5d2607b58a86ee0a8932032a0f8a31b21d39b525 Reviewed-on: https://chromium-review.googlesource.com/275774Tested-by:
-
Jamie Madill authored
With the list of failures, this gives a 100% passing GLES2 test suite. BUG=angleproject:998 Change-Id: I20985c729104d7a114352d156c238bca3ea1698c Reviewed-on: https://chromium-review.googlesource.com/274423Reviewed-by:
-
- 11 Jun, 2015 4 commits
-
-
Jamie Madill authored
We can use the ANGLE platform to capture more specific information than total number of iterations. We can also capture time spent initializing DLLs, calling D3D11CreateDevice and allocating some default resources. BUG=angleproject:1014 Change-Id: If0f4965e7aac22eb09dfa0d2300ff449e7481f6e Reviewed-on: https://chromium-review.googlesource.com/276631Reviewed-by:
-
Jamie Madill authored
This was broken by the angle test platforms refactor patch, https://chromium-review.googlesource.com/#/c/276460/. It was leading to the wrong tests be executed for the perf tests, and also had bugs with supporting a default platform. BUG=None TEST=angle_perftests Change-Id: Ib99342d4189fd98bf93dce088d15afcecc4a7e37 Reviewed-on: https://chromium-review.googlesource.com/277036Tested-by:
Zhenyao Mo <zmo@chromium.org>
-
Olli Etuaho authored
TEST=angle_unittests Change-Id: I49f17444bf0fe6568ec4d3445421c27a562adf1b Reviewed-on: https://chromium-review.googlesource.com/277040Tested-by:
Olli Etuaho <oetuaho@nvidia.com> Reviewed-by:
-
Corentin Wallez authored
Replace usage of std::exception by std::runtime_error BUG=angleproject:892 Change-Id: I214fe6a93944dcfb39b0633f46caf3b1187bd5d3 Reviewed-on: https://chromium-review.googlesource.com/275815Reviewed-by:
Corentin Wallez <cwallez@chromium.org>
-
- 10 Jun, 2015 17 commits
-
-
Jamie Madill authored
Parent patch fixed. BUG=angleproject:1014 This reverts commit 79ae150e. Change-Id: I6431d73e5a766d1937732ddb7964751d46465674 Reviewed-on: https://chromium-review.googlesource.com/276776Reviewed-by:
-
Jamie Madill authored
*re-land with fix for test build* These tables can save us on startup time. Instead of querying the tables, we can keep certain assumptions about texture formats because certain formats are guaranteed to be supported in various feature levels. Testing with angle_perftests indicates this patch series will save about 50% of the time we spend in Renderer11::initializeDevice. BUG=angleproject:1014 Change-Id: I5477eea3e8433d36119cd099fb0c08eea168489a Reviewed-on: https://chromium-review.googlesource.com/276758Reviewed-by:
-
Jamie Madill authored
Hidden conflict with Geoff's test refactor patch. BUG=angleproject:1014 This reverts commit 396ee48b. Change-Id: I8182acb350b61b31f4040a20c86652e23706905a Reviewed-on: https://chromium-review.googlesource.com/276774Reviewed-by:
-
Jamie Madill authored
Hidden conflict with Geoff's test refactor patch. BUG=angleproject:1014 This reverts commit a3028ab9. Change-Id: Ic202e377032e9c74f62450ba0ffcba85a1fc0cc7 Reviewed-on: https://chromium-review.googlesource.com/276773Reviewed-by:
-
Jamie Madill authored
This saves us from querying a lot of the formats on startup. BUG=angleproject:1014 Change-Id: I499625b9a131cf10791d1580d7a7b477a161045a Reviewed-on: https://chromium-review.googlesource.com/275773Reviewed-by:
-
Jamie Madill authored
These tables can save us on startup time. Instead of querying the tables, we can keep certain assumptions about texture formats because certain formats are guaranteed to be supported in various feature levels. Testing with angle_perftests indicates this patch series will save about 50% of the time we spend in Renderer11::initializeDevice. BUG=angleproject:1014 Change-Id: I84dbeaba1d1c635ad60c8413705cc41af8a51f06 Reviewed-on: https://chromium-review.googlesource.com/275772Reviewed-by:
-
Jamie Madill authored
This can be useful for our testing code, to filter based on system configurations, such as GPU type, OS, or Debug/Release. BUG=angleproject:998 Change-Id: I88b7fe1a7a1446262cc084e62cdb335776fa8912 Reviewed-on: https://chromium-review.googlesource.com/274422Reviewed-by:
-
Geoff Lang authored
Limit the supported ES version based on these caps. BUG=angleproject:879 Change-Id: If3a097576465b33af0015fcb28dab10e88fdc4d9 Reviewed-on: https://chromium-review.googlesource.com/273160Reviewed-by:
-
Geoff Lang authored
For each config, determine if a context can be created at test instantiation time. This allows skipping of ES3 tests when the hardware does not support ES3. Updated the perf_tests to use the EGLPlatformParameters struct so that they can be filtered in the same way. Change-Id: If664604b057cec4005eb4b63bebd83cd4964b7b2 Reviewed-on: https://chromium-review.googlesource.com/276460Reviewed-by:
-
Geoff Lang authored
BUG=angleproject:947 Change-Id: I8cbcd99e8c2cb7c461b0c624596631a4dec6dcc8 Reviewed-on: https://chromium-review.googlesource.com/276668Reviewed-by:
-
Qingqing Deng authored
BUG=angleproject:947 TEST=angle_unittests, angle_end2end_tests Change-Id: Id1afd7630d9720b6701225b7abf1ce4a3bf528b6 Reviewed-on: https://chromium-review.googlesource.com/260232Reviewed-by:
-
Olli Etuaho authored
Add a traverser that checks precision qualifiers of folded constants and hoists them to separate precision qualified variables if needed. Fixes sdk/tests/conformance/glsl/bugs/constant-precision-qualifier.html TEST=WebGL conformance tests, angle_unittests BUG=angleproject:817 Change-Id: I1639595e0e49470736be93274f0af07ee732e1fe Reviewed-on: https://chromium-review.googlesource.com/275095Reviewed-by:
-
Geoff Lang authored
BUG=angleproject:817 Change-Id: Iaa971c71c0f5bcec46fffb5d2e70f108e6ca6b02 Reviewed-on: https://chromium-review.googlesource.com/276607Reviewed-by:
-
Olli Etuaho authored
Refactor translating a ESSL shader string into a target language so that compiler initialization and cleanup code can be reused between test classes. BUG=angleproject:817 TEST=angle_unittests Change-Id: Idb229dceb9e17b13ed6ad2a68ab55ed5c968780e Reviewed-on: https://chromium-review.googlesource.com/275814Tested-by:
-
Austin Kinross authored
Change-Id: I25236240eec135b265a25eefdbbd28f069633eee Reviewed-on: https://chromium-review.googlesource.com/276308Tested-by:
-
Olli Etuaho authored
This broke Chromium build on Linux. This reverts commit c6a61c27. Change-Id: I4df380eaed3ecc19b85aa99a397968ceb5ce3af8 Reviewed-on: https://chromium-review.googlesource.com/276293Reviewed-by:
-
Corentin Wallez authored
BUG=angleproject:892 Change-Id: Iceae30f5bc12546b2486e8aac0346eab1f09ded3 Reviewed-on: https://chromium-review.googlesource.com/276200Reviewed-by:
-
- 09 Jun, 2015 7 commits
-
-
Austin Kinross authored
Change-Id: I52bcf0cfb1aa123e085a35730fdefb006b617c3c Reviewed-on: https://chromium-review.googlesource.com/269232Tested-by:
-
Jamie Madill authored
This should complete the breakdown of the time we spent in eglInitialize. BUG=angleproject:1014 BUG=436191 Change-Id: I3b44f10c56322924f9391a56973df5f439fa83bb Reviewed-on: https://chromium-review.googlesource.com/276268Tested-by:
-
Jamie Madill authored
Introduce a "null" GL renderer for this specific benchmark, which does not do any work on drawArrays and drawElements. We could use the same kind of trick for buffer and texture updates, if we need it. This gives us a good baseline for comparing GL and D3D perf. BUG=angleproject:1040 Change-Id: I4bf7c75df01932de45ddd0a4e42e8fc82f15e37e Reviewed-on: https://chromium-review.googlesource.com/276192Tested-by:
-
Jamie Madill authored
Use value-parameterized tests, loading a full test list from case files. BUG=angleproject:998 Change-Id: Iff65a2d722e9d105e22649b17a18cc74dccea2b1 Reviewed-on: https://chromium-review.googlesource.com/274421Reviewed-by:
-
Jamie Madill authored
A prototype based on work from phaulos@. This should eventually be integrated into the dEQP repo. BUG=angleproject:998 Change-Id: I9be137fb54423bc180623f172dde63a22311b791 Reviewed-on: https://chromium-review.googlesource.com/274420Reviewed-by:
-
Geoff Lang authored
BUG=angleproject:890 Change-Id: I66fb7b97ea3f09d40ba25a769c8bc7fe04527805 Reviewed-on: https://chromium-review.googlesource.com/275307Reviewed-by:
-
Olli Etuaho authored
Setting the type for folded aggregate nodes should work in a similar way as other constant folding. Common functionality between the different folding functions is refactored into a single function. TEST=dEQP-GLES3.functional.shaders.constant_expressions.* BUG=angleproject:817 Change-Id: Ie0be561f4a30e52e52d570ff0b2bdb426f6e4f7a Reviewed-on: https://chromium-review.googlesource.com/275186Reviewed-by:
-
- 08 Jun, 2015 2 commits
-
-
Geoff Lang authored
Seeing failures on some Intel HD4000 cards as well as AMD. BUG=angleproject:1038 Change-Id: I45dfcdf62af69cd3295918d963def85f2eb50ebe Reviewed-on: https://chromium-review.googlesource.com/274943Reviewed-by:
-
Jamie Madill authored
Use the dEQP null platform for speed and to avoid popping up a new window. Store the case lists as .gz files to save space in the repo. BUG=angleproject:998 Change-Id: Ib9b10b8a83edc510cc98f120847b1cf93a3d2f78 Reviewed-on: https://chromium-review.googlesource.com/273975Reviewed-by:
-
