kc3-lang/angle/src/libANGLE/renderer/vulkan/vk_cache_utils.h

src/libANGLE/renderer/vulkan/vk_cache_utils.h

Log

Author Commit Date CI Message
Shahbaz Youssefi adde4265 2022-10-19T23:33:48 Vulkan: Separate pipeline cache query and insertion In preparation for VK_EXT_graphics_pipeline_library usage, the query and insertion functions of the graphics pipeline cache are separated. This will allow the implementation using VK_EXT_graphics_pipeline_library to query the monolithic pipeline cache, and if a pipeline is not found, create it through the pipeline library caches. Bug: angleproject:7369 Change-Id: Iebf7669ae3ea95e180646198c4861cc59d67e580 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3963854 Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Sungyong Choi c6390143 2022-10-12T09:57:55 Vulkan: Make compatible with GCC Resolves below warnings occurred with GCC build. 1) deperecated-copy Overriding an assignment operator without a copy constructor caused the deprecated-copy warnings. 2) unused-function 3) parenthesis Warnings occurred due to missing parenthesis around some logical expressions, add them to quiet the warnings. 4) unused variable 5) 'maybe-unused' attribute ignored Introduces 'ANGLE_MAYBE_UNUSED_PRIVATE_FIELD' macro to avoid 'attribute ignored' warning which is only occurred with GCC because GCC doesn't warn about 'unused non static data member' whereas Clang has Wno-unused-private-field. Signed-off-by: Sungyong Choi <sywow.choi@samsung.com> Bug: angleproject:7764 Change-Id: I8e7410a5ed8cb9b8f8b3202073d779fea63d6b75 Reviewed-by: Jeff Vigil <j.vigil@samsung.com> Reviewed-by: Mohan Maiya <m.maiya@samsung.com> Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3963830 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 7b4b56f0 2022-10-19T00:05:22 Vulkan: Missing output mask in GraphicsPipelineDesc Currently, there's some program state used in creating pipelines alongside what's in GraphicsPipelineDesc. This works because the pipeline cache lives in the program executable. With VK_EXT_graphics_pipeline_library however, we could create vertex input and fragment output partial pipelines that are independent from and are shared between multiple programs. To support this, any program state that's necessary for pipeline creation should be part of the GraphicsPipelineDesc structure. This change places the state affecting fragment output in GraphicsPipelineDesc. Bug: angleproject:7369 Change-Id: I2e48fc9da220475e1b2ed376fc947ce13489610e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3963652 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Shahbaz Youssefi f0e3d8f9 2022-10-18T13:51:31 Vulkan: Shader component type in GraphicsPipelineDesc Currently, there's some program state used in creating pipelines alongside what's in GraphicsPipelineDesc. This works because the pipeline cache lives in the program executable. With VK_EXT_graphics_pipeline_library however, we could create vertex input and fragment output partial pipelines that are independent from and are shared between multiple programs. To support this, any program state that's necessary for pipeline creation should be part of the GraphicsPipelineDesc structure. This change places the state affecting vertex input in GraphicsPipelineDesc. A follow up change will do the same for state affecting fragment output. Bug: angleproject:7369 Change-Id: Iccf691a1597d786efa1625f7b1c22f906201f2e5 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3964751 Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Shahbaz Youssefi b521be4c 2022-10-14T23:09:26 Vulkan: Decouple shader-set from pipeline caches In preparation for VK_EXT_graphics_pipeline_library, where different pipeline caches (for the complete pipelines and the shaders subset partial pipelines) may create pipelines from the same shader set, the pipeline caches are pulled out of ShaderProgramHelper. In an upcoming change, ProgramExecutableVk will have more than one GraphicsPipelineCache instance, both creating pipelines through the same ShaderProgramHelper. The pipeline creation methods in ShaderProgramHelper are now const. This means a thread would be able to create pipelines using an object of this class while the main thread creates other pipelines using the same object, but in a different pipeline cache. Bug: angleproject:7369 Change-Id: Ib8a76dedf1105ba9dfcad9e972157c92ba18e349 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3956944 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 09b079bb 2022-10-14T14:28:10 Vulkan: Move pipeline subset out of pipeline desc This way, the same pipeline desc can be used to query multiple caches with different subsets. This design requires the pipeline cache to be templated, while the previous design required the desc to be copied (so the subset bits would be different). Bug: angleproject:7369 Change-Id: Id3fd92e6f9b059069952ff092a5c867e79287604 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3956940 Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Shahbaz Youssefi eaa71709 2022-10-14T11:47:07 Vulkan: s/ContextVk/Context in pipeline creation With VK_EXT_graphics_pipeline_library, pipeline creation may happen on a thread. This change prepares the interface such that only a vk::Context is needed, instead of ContextVk. Bug: angleproject:7369 Change-Id: Ib7e9e7e140e27a4af71bffee069c88e3d7f47dfa Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3956935 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 8e644365 2022-10-11T11:47:55 Vulkan: Rearrange state specification for pipeline libraries This change splits pipeline state setup based on whether the complete pipeline is being created or only a subset of it. However, no partial pipelines are yet created, that will be done in a following CL. Bug: angleproject:7369 Change-Id: Ife34c4bc39641ab8f06e7c78502f95a6bc16d15d Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3949917 Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Shahbaz Youssefi 94b45181 2022-10-12T14:54:14 Vulkan: Rearrange graphics pipeline desc bits In preparation for use of VK_EXT_graphics_pipeline_library, the pipeline state is split in three contiguous regions: - Vertex input - Pre-rasterization and fragment stages ("Shaders" for short) - Fragment output There is some state that affects both Shaders and Fragment output, which is split and placed in between the two. This will allow the hash and compare functions to eventually access each of those pipeline subsets as a contiguous piece of memory. Bug: angleproject:7369 Change-Id: Iedc4cf15ed6c7fed6ba93039889fbf5dd191e041 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3949914 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 135022e4 2022-10-11T00:03:11 Vulkan: Create robust pipelines based on context state Previously, pipelines were made robust based on whether any context in the share group has so far been made robust. This means that pipelines created on non-robust contexts would still be compiled as robust. Inefficiency aside, this was buggy because robustness was not part of the pipeline cache key, so if a pipeline was created as non-robust first, then recreated in a robust context, it would reuse the non-robust variant. With VK_EXT_pipeline_protected_access, a similar situation arises for context protected-ness. However, it is incorrect in that case to create pipelines as protected unnecessarily. This change makes pipeline robustness a part of the pipeline cache key, in preparation for protectedness to be added similarly. Compute programs may now generate multiple pipelines as a result too. Bug: angleproject:7629 Change-Id: Ie95f10eff878f8c8b221c1018da44385c7aad15e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3943534 Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com>
Amirali Abdolrashidi c19ec948 2022-08-23T10:43:59 Vulkan: Implement imageless framebuffers * Added the attachment image and create info objects to be used for imageless framebuffers created in getFramebuffer(). * New helper class for framebuffers in RenderPassCommandBufferHelper: MaybeImagelessFramebuffer, which includes a framebuffer object, if the framebuffer is imageless, and the image views. This is to make sure that the args for render pass begin info will be correctly set up according to the status of the used framebuffer. * Refactored the collection of attachments in getFramebuffer() into a new function, getAttachmentsAndImagesFromRenderTargets(). It also returns their corresponding ImageHelper* objects used to create the framebuffer (from their image properties). * New struct: RenderTargetInfo; which keeps track of render targets and whether resolve image should be used for the render pass in the form of the enum class RenderTargetImage. * Added a new arg to getFramebuffer(): resolveRenderTargetIn; to use when there is a valid resolveImageViewIn. * Without using the framebuffer cache, we would require to handle the framebuffer destruction by adding it to the garbage instead of releasing it. For example, FramebufferVk::destroy() now adds mCurrentFramebuffer to the garbage. * Added new framebuffer unit tests. * Added tests where two textures with different attributes are bound to the same framebuffer before drawing, one after another. * Added test where a blit occurs from a multisample texture into a non-zero level of a resolve texture, each bound to a separate FBO. * Added a new perf test to compare performance for enabled imageless framebuffers vs disabled. (Credit: cclao) Bug: angleproject:7553 Change-Id: Iacdbd73aaa01cbb0e37abf01ae4892bdfdd4b12f Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3827644 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com> Reviewed-by: Charlie Lao <cclao@google.com>
Shahbaz Youssefi ba3b4515 2022-09-15T01:06:51 Vulkan: Implement GL_ANGLE_logic_op Enabled if the logicOp device feature is available. According to gpuinfo, it's pretty much universal except for ARM. Bug: angleproject:7654 Change-Id: I4808b519fdd6273b2f8c1bb17f59517eb65bfe8d Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3898317 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 952d6e8f 2022-08-21T21:51:03 Vulkan: Workaround ARM driver bug with dynamic vertex strides According to ARM, vkCmdBindVertexBuffers2EXT is broken when more than one attribute binding is involved. In this change, the vertex strides dynamic state is avoided altogether on ARM. Bug: fuchsia:107106 Change-Id: Ie41fc07037083214d2dc99e4ed062485384911c5 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3844640 Reviewed-by: Cody Northrop <cnorthrop@google.com> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 493b5aff 2022-08-09T14:57:24 Vulkan: Workaround ARM bug with stencil write mask Bug: angleproject:7556 Change-Id: I0aa17c178071cc15d8ee15f700b0c4932819c72a Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3821367 Reviewed-by: Ian Elliott <ianelliott@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Cody Northrop <cnorthrop@google.com>
Amirali Abdolrashidi 7fe0694c 2022-07-14T16:42:30 Vulkan: Use push constants for driver uniforms * Updated the driver uniforms so they would be defined as push constants in SPIR-V. Their data would be updated via pushConstants() when handling the driver uniform dirty bits. * Updated TOutputVulkanGLSL to be able to generate the push constants as required. * When handling the driver uniform dirty bits, we no longer allocate a buffer. * Removed the driver uniform descriptor set from the pipeline layout. * Removed the binding-related functions and flags for the driver uniforms. * In invalidateGraphicsDescriptorSet(), DIRTY_BIT_DESCRIPTOR_SETS is used instead of DIRTY_BIT_DRIVER_UNIFORMS_BINDING. (Same for invalidateComputeDescriptorSet()) * DIRTY_BIT_DRIVER_UNIFORMS_BINDING is replaced with DIRTY_BIT_DRIVER_UNIFORMS in other places. * Removed mDriverUniform and DriverUniformsDescriptorSet from ContextVk. * Added mSupportedVulkanShaderStageMask to RendererVk, which is used in creating the pipeline layout and updating the push constants. * Added a TODO note for driverUniformsDescriptorSetIndex. Bug: angleproject:6858 Change-Id: I91037d378528962a816b12ff3f21249ee17b7652 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3782570 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Charlie Lao 5db09e9b 2022-07-21T10:45:12 Vulkan: Try allocate from existing DescriptorPool before create new When we allocate a new descriptorSet, right now we only look at current bound pool or the last pool, or allocate a new pool. We never look at the other pools for the possibility of allocation. This is likely due to we never free descriptorSet in the past. With the recent CLs, we now release invalid descriptorSets when texture or buffer gets deleted or re-specified. This opens up opportunity to allocate from other existing pools before allocating a new pool. This CL changes the allocation logic to add the pass to iterate over existing pools for allocation before allocating a new pool, thus reducing the number of descriptorSetPools. This CL also consolidates DynamicDescriptorPool::getOrAllocateDescriptorSet()'s actual descriptorSet allocation logic with DynamicDescriptorPool::allocateDescriptorSet(). This CL also moves mEmptyDescriptorSets from ProgramExecutableVk to DynamicDescriptorPool so that we will only have one empty descriporSet per pool instead of per program. Bug: b/235523746 Change-Id: I012346acce17f785ee6683ec55fdf21be00ea1a0 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3780847 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Charlie Lao <cclao@google.com> Auto-Submit: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Charlie Lao 00e48c13 2022-07-20T17:35:01 Vulkan: Destroy DescriptorPoolHelper when its unused. DynamicDescriptorPool keeps an array of DescriptorPoolHelpers. Right now we only grow this array of pools, we never shrink the pool count. This is partly due to we never release descriptorSet. But in the past few CLs, we now release invalid descriptorSet when texture/buffer gets deleted or re-specified. This means we could now have a pool with no valid descriptorSets. This CL adds the ability to actually release the pool when all of its descriptorSets has been released, thus reduce the pool count when a lot of textures have been deleted. Bug: b/235523746 Change-Id: I2d5047269154cc8ece8305408f08f2ad7c9dd8a6 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3780845 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Charlie Lao <cclao@google.com>
Lingfeng Yang bd82a5e4 2022-07-07T17:03:27 Vulkan: __samplerExternal2DY2YEXT-aware ProgramExecutableVk This CL incorporates __samplerExternal2DY2YEXT into the Vulkan objects involved in ProgramExecutableVk. Specifically, we need: 1. TextureVk to create and return a VkSampler imbued with identity conversion model, so we can create an immutable sampler from it. 2. Access to the identity conversion desc so that it can be traced in mImmutableSamplerIndexMap. Bug: b/223456677 Change-Id: Ia302226a8aa4a97d4a42f79c79ce47af3abec37c Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3753294 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Lingfeng Yang <lfy@google.com>
Lingfeng Yang 6c418f8c 2022-07-07T14:28:06 Vulkan: __samplerExternal2DY2YEXT-aware TextureVk This CL adds the ability for TextureVk to return an ImageView that is created with a VkSamplerYcbcrConversion object that uses an identity conversion model. This allows direct sampling of YUV values without RGB conversion, which is needed for __samplerExternal2DY2YEXT. Bug: b/223456677 Change-Id: Ie1d4e12375b7808a1f060747bc2d74baeda3fdea Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3751889 Commit-Queue: Lingfeng Yang <lfy@google.com> Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Charlie Lao 53d40aed 2022-07-15T15:03:25 Vulkan: Destroy descriptorSet cache when BufferHelper destroyed For atomic counter buffers or other cases, dynamic descriptor is not been used. Right now when such buffer is destroyed, the cache is still lingers around. With this CL, when a new cache entry has been created, we record the cache entry in the BufferHelper. When BufferHelper is destroyed, we also immediately destroy the cache entry since the cache will no longer reused. Bug: b/237686097 Change-Id: I26eee96318fbc003e65318c0b8263dc61092f350 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3764044 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Charlie Lao c7459a46 2022-07-15T09:55:03 Vulkan: Destroy descriptorSet cache when BufferBlock destroyed When a new cache entry has been created, we record the cache entry in the BufferBlock. When BufferBlock is destroyed, we also immediately destroy the cache entry since the cache will no longer reused. This CL also removes DescriptorCacheResult from various APIs since it is now redundant with newSharedCacheKey argument. Bug: b/237686097 Change-Id: I14fa8906fdbe7d9226c8e8ecddef2beb05fbaa5c Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3756694 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Charlie Lao <cclao@google.com> Auto-Submit: Charlie Lao <cclao@google.com>
Charlie Lao 01092c48 2022-07-12T10:11:22 Vulkan: Destroy descriptorSet cache when shader image is destroyed Similar to texture descriptor set, this applies to images used as shader resource. When a texture is used in a shader resource descriptorSet, we record it. When texture is destroyed, we also destroy that shader resource descriptorSet cache. Bug: b/237686097 Change-Id: I475982fcec45535cc285a4aebca922d01efc7ed2 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3758884 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Auto-Submit: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Charlie Lao <cclao@google.com>
Charlie Lao e25137c0 2022-06-29T09:57:50 Vulkan: Move DescriptorSet cache to DynamicDescriptorPool DynamicDescriptorPool has an array of DescriptorPoolHelper. Right now mDescriptorSetCache is stored in DescriptorPoolHelper. This means when you search cache for the match, you have to iterate over the array of DescriptorPoolHelper to decide to make conclusion if there is a match or not. Each hash map search means compute hash and do key comparison. I think this was done this way because of we used to not able to evict cache when we decide to reuse DescriptorSetPoolHelper object (when we call pool->init). But with recent changes of SharedDescriptorSetCacheKey, we are able to evict specific cache entries. This CL moves mDescriptorSetCache to DynamicDescriptorPool so that cache look up can be cheaper. This also made CacheStats simpler as well. With Gfxbench gl_driver2_off and cpu clock locked, this CL improves score by 2.16% (from 4019 before CL to 4105 after CL). See bug for detailed data. Bug: b/237686097 Change-Id: Ia6fa7a6b725974e0150bc21cdf0140d9198c8332 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3735736 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Yuxin Hu 0d3ecf46 2022-06-24T16:37:17 Vulkan: Multisample Framebuffer Fetch Implement Multisample Framebuffer Fetch. This should fix the deqp failure dEQP.GLES31/functional_blend_equation_advanced_msaa_colorburn Bug: angleproject:7351 Bug: angleproject:3586 Bug: angleproject:6195 Bug: b/234173199 Change-Id: Idd7559dcba3d91e36d8f253f1554fb931a7a6775 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3724165 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Charlie Lao <cclao@google.com>
Mohan Maiya 81a69da6 2022-06-30T09:56:54 Vulkan: min/mag filters follow chroma filter value The Vulkan spec states that for those formats lacking support for VK_FORMAT_FEATURE_SAMPLED_IMAGE_YCBCR_CONVERSION_SEPARATE_RECONSTRUCTION_FILTER_BIT the min and mag filter must be equal to the sampler YCbCr conversion's chroma filter. Having the min/mag filters follow the chroma filter leads to simpler and more efficient code. Also update getPreferredFilterForYUV function to return existing filter value when preferLinearFilterForYUV feature is disabled. Bug: angleproject:7382 Bug: angleproject:7392 Tests: Texture2DTestES3YUV.TexStorage2DYuvFilterModes* Tests: ImageTestES3.SourceYUVAHBTargetExternalYUVSampleLinearFiltering/ES3_Vulkan Change-Id: I550ef8feede1dc6c3a0d8e32f790113e90ef7a4b Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3739582 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: mohan maiya <m.maiya@samsung.com>
Charlie Lao 723cc880 2022-06-10T17:55:54 Reland "Reland "Vulkan: Destroy DescriptorSet cache when it becomes invalid"" This is a reland of commit 551a26aeedbfd971d6199c8eddb433a4f4ff871c Original change's description: > Reland "Vulkan: Destroy DescriptorSet cache when it becomes invalid" > > This is a reland of commit 0779ccbcd427dcb00e53afa6385fb4e8e2377993 with > the fix for angleproject:7466. When DescriptorPoolhelper gets > release/destroyed, we ensure all sharedCacheKeys are destroyed. > > Original change's description: > > Vulkan: Destroy DescriptorSet cache when it becomes invalid > > > > When a new texture descriptorSet is allocated, we store one reference of > > the cache key in ProgramExecutableVk and all TextureVks that it > > associated with. When any of the TextureVk is destroyed or its view > > destroyed, we immediately erase the descriptorSet from the cache and > > track GPU progress and free the descriptorSet when it's GPU completed. > > That way we delete the dead descriptorSet that will for sure never been > > reused ASAP so that its space is avialable for reuse. > > > > Bug: b/235523746 > > Change-Id: Ib1b9662a254eea5a3f410dc6d5d89fca6727a647 > > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3700226 > > Reviewed-by: Lingfeng Yang <lfy@google.com> > > Commit-Queue: Charlie Lao <cclao@google.com> > > Reviewed-by: Ian Elliott <ianelliott@google.com> > > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> > > Bug: b/235523746 > Bug: angleproject:7466 > Change-Id: I4413bec27ea0ca830010e2ca15036c2e667141c0 > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3726964 > Reviewed-by: Ian Elliott <ianelliott@google.com> > Commit-Queue: Charlie Lao <cclao@google.com> > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Bug: b/235523746 Bug: angleproject:7466 Change-Id: I6b88b884841c5dbc625ee7e0c52c45af09dec199 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3741027 Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Ian Elliott <ianelliott@google.com>
Ian Elliott 4b745c2b 2022-07-01T01:25:36 Revert "Reland "Vulkan: Destroy DescriptorSet cache when it becomes invalid"" This reverts commit 551a26aeedbfd971d6199c8eddb433a4f4ff871c. Reason for revert: Blink test failures at: https://ci.chromium.org/ui/p/chromium/builders/ci/WebKit%20Linux%20MSAN/15546/overview Original change's description: > Reland "Vulkan: Destroy DescriptorSet cache when it becomes invalid" > > This is a reland of commit 0779ccbcd427dcb00e53afa6385fb4e8e2377993 with > the fix for angleproject:7466. When DescriptorPoolhelper gets > release/destroyed, we ensure all sharedCacheKeys are destroyed. > > Original change's description: > > Vulkan: Destroy DescriptorSet cache when it becomes invalid > > > > When a new texture descriptorSet is allocated, we store one reference of > > the cache key in ProgramExecutableVk and all TextureVks that it > > associated with. When any of the TextureVk is destroyed or its view > > destroyed, we immediately erase the descriptorSet from the cache and > > track GPU progress and free the descriptorSet when it's GPU completed. > > That way we delete the dead descriptorSet that will for sure never been > > reused ASAP so that its space is avialable for reuse. > > > > Bug: b/235523746 > > Change-Id: Ib1b9662a254eea5a3f410dc6d5d89fca6727a647 > > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3700226 > > Reviewed-by: Lingfeng Yang <lfy@google.com> > > Commit-Queue: Charlie Lao <cclao@google.com> > > Reviewed-by: Ian Elliott <ianelliott@google.com> > > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> > > Bug: b/235523746 > Bug: angleproject:7466 > Change-Id: I4413bec27ea0ca830010e2ca15036c2e667141c0 > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3726964 > Reviewed-by: Ian Elliott <ianelliott@google.com> > Commit-Queue: Charlie Lao <cclao@google.com> > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Bug: b/235523746 Bug: angleproject:7466 Change-Id: Icdde2752c462b7ebbb51d46fd35ce749b5caf377 No-Presubmit: true No-Tree-Checks: true No-Try: true Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3739585 Reviewed-by: Ian Elliott <ianelliott@google.com> Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com> Auto-Submit: Ian Elliott <ianelliott@google.com> Commit-Queue: Ian Elliott <ianelliott@google.com>
Charlie Lao 551a26ae 2022-06-10T17:55:54 Reland "Vulkan: Destroy DescriptorSet cache when it becomes invalid" This is a reland of commit 0779ccbcd427dcb00e53afa6385fb4e8e2377993 with the fix for angleproject:7466. When DescriptorPoolhelper gets release/destroyed, we ensure all sharedCacheKeys are destroyed. Original change's description: > Vulkan: Destroy DescriptorSet cache when it becomes invalid > > When a new texture descriptorSet is allocated, we store one reference of > the cache key in ProgramExecutableVk and all TextureVks that it > associated with. When any of the TextureVk is destroyed or its view > destroyed, we immediately erase the descriptorSet from the cache and > track GPU progress and free the descriptorSet when it's GPU completed. > That way we delete the dead descriptorSet that will for sure never been > reused ASAP so that its space is avialable for reuse. > > Bug: b/235523746 > Change-Id: Ib1b9662a254eea5a3f410dc6d5d89fca6727a647 > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3700226 > Reviewed-by: Lingfeng Yang <lfy@google.com> > Commit-Queue: Charlie Lao <cclao@google.com> > Reviewed-by: Ian Elliott <ianelliott@google.com> > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Bug: b/235523746 Bug: angleproject:7466 Change-Id: I4413bec27ea0ca830010e2ca15036c2e667141c0 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3726964 Reviewed-by: Ian Elliott <ianelliott@google.com> Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Jose Dapena Paz 1dc3385a 2022-06-27T18:13:33 libstdc++: move FramebufferHelper to cache utils vk_helper.h includes vk_cache_utils.h. FramebufferHelper was declared in vk_helper.h. And FramebufferCache, in vk_cache_utils.h, requires FramebufferHelper for its payload hashmap. This was solved with a forward declaration. But, with std::pair definition in libstdc++ this is not enough. Bug: chromium:957519 Change-Id: I32e3e247c3c8bef8b52806aa2601e4012e6aa79e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3727671 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Yuly Novikov f8690429 2022-06-27T13:50:44 Revert "Vulkan: Destroy DescriptorSet cache when it becomes invalid" This reverts commit 0779ccbcd427dcb00e53afa6385fb4e8e2377993. Reason for revert: crashes in blink_web_tests on linux-rel https://ci.chromium.org/ui/p/chromium/builders/try/linux-rel/1051045/overview Original change's description: > Vulkan: Destroy DescriptorSet cache when it becomes invalid > > When a new texture descriptorSet is allocated, we store one reference of > the cache key in ProgramExecutableVk and all TextureVks that it > associated with. When any of the TextureVk is destroyed or its view > destroyed, we immediately erase the descriptorSet from the cache and > track GPU progress and free the descriptorSet when it's GPU completed. > That way we delete the dead descriptorSet that will for sure never been > reused ASAP so that its space is avialable for reuse. > > Bug: b/235523746 > Change-Id: Ib1b9662a254eea5a3f410dc6d5d89fca6727a647 > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3700226 > Reviewed-by: Lingfeng Yang <lfy@google.com> > Commit-Queue: Charlie Lao <cclao@google.com> > Reviewed-by: Ian Elliott <ianelliott@google.com> > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Bug: b/235523746, angleproject:7466 Change-Id: I7e5067de2f2add08af1f9804cc2e952238b2e942 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3726097 Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
Shahbaz Youssefi e50351cb 2022-06-10T22:28:58 Vulkan: Don't close render pass on framebuffer fetch For applications that use framebuffer fetch in the same RP as non-fetch programs, we can save some extra RenderPasses by always creating our RP objects with input attachments enabled. This works almost identically except for needing to use the images in a "GENERAL" layout instead of "COLOR_ATTACHMENT_OPTIMAL". According to partners it is possible to achieve performance parity even with GENERAL layout. To remove any potential negative impacts of using the GENERAL layout, the context enters this always-framebuffer-fetch mode only and as soon as a framebuffer fetch program is created. Applications that don't use framebuffer fetch are thus unaffected. This eliminates 20 render passes in the Genshin Impact trace (out of about 58). On a Pixel 6 the resulting benchmark score speeds up by ~25%. For Real Racing 3, the speed up is ~30%. Based on change by jmadill@chromium.org Bug: angleproject:7375 Change-Id: Ib6c73e95d06229f8545d502b388ee2a55a582323 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3697308 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Eddie Hatfield 91976352 2022-06-21T15:41:02 Use C++17 attributes instead of custom macros Bug: angleproject:6747 Change-Id: Iad6c7cd8a18d028e01da49b647c5d01af11e0522 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3718999 Reviewed-by: Cody Northrop <cnorthrop@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Charlie Lao 0779ccbc 2022-06-10T17:55:54 Vulkan: Destroy DescriptorSet cache when it becomes invalid When a new texture descriptorSet is allocated, we store one reference of the cache key in ProgramExecutableVk and all TextureVks that it associated with. When any of the TextureVk is destroyed or its view destroyed, we immediately erase the descriptorSet from the cache and track GPU progress and free the descriptorSet when it's GPU completed. That way we delete the dead descriptorSet that will for sure never been reused ASAP so that its space is avialable for reuse. Bug: b/235523746 Change-Id: Ib1b9662a254eea5a3f410dc6d5d89fca6727a647 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3700226 Reviewed-by: Lingfeng Yang <lfy@google.com> Commit-Queue: Charlie Lao <cclao@google.com> Reviewed-by: Ian Elliott <ianelliott@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Charlie Lao 72e457fe 2022-06-03T15:36:01 Vulkan: Promptly destroy cached framebuffer when it becomes invalid When Texture gets respecified, the VkFramebuffer cache created out of it becomes invalid and will never possibly get used. Before this CL, we never clear such invalid framebuffer objects from the cache. This CL keeps a reference to the cache key in each attachment and will immediately destroy the cached VkFramebuffer object when one of the attachment has become invalid. Bug: b/234769934 Change-Id: Ib01f6dffe9211084b1ada340081daf905e3f1bef Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3682164 Reviewed-by: Ian Elliott <ianelliott@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Charlie Lao <cclao@google.com>
Shahbaz Youssefi 97a6e581 2022-05-30T16:50:26 Vulkan: Useful implementation of program binaries ANGLE already serializes the pipeline state for the sake of OES_get_program_binary. This serialization had limited usefulness however, since the Vulkan driver hasn't actually created any pipelines yet (which is a costly part of program creation). Simultaneously, ANGLE deferred Vulkan pipeline creation to draw time, which causes hitching. In this change, a handful of Vulkan pipelines are precreated at link time; those at least that are sure to create different blobs in the pipeline cache (different spec consts or SPIR-V generation). These pipelines are created in the program executable's cache. The cache is then merged into the shared renderer cache (for potential blob reuse by other programs). With this, two goals are achieved: - Most pipelines created at draw time hit the pipeline cache, avoiding costly compilation. - When the program binary is retrieved, the contents of the program executable's pipeline cache is also returned. On reload, the cache is recovered, resulting in faster startup. Bug: angleproject:5881 Change-Id: I46c5451a7d0b16dffd40e44015e094640886880b Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3671977 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 0bc70e96 2022-06-10T22:27:04 Vulkan: Rename getShareGroupVk to getShareGroup + other miscellaneous clean up Bug: angleproject:7375 Change-Id: I25690860478a2fd181a67ce2b6cb4d7aac7dbaa7 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3700197 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi a9516865 2022-06-01T22:48:04 Vulkan: Output cache look up feedback in pipeline graph Bug: angleproject:6565 Change-Id: I12bb9ab5756860de9ba26d6b4a9429a78b65df39 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3686029 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 53ec886e 2022-05-30T16:49:11 Vulkan: Externally synchronize the pipeline cache In preparation for a future change that requires this as it may perform pipeline cache merges during creation of pipelines. Bug: angleproject:5881 Change-Id: Ic7921b781aa773ae23b60a0bb6fa2111b1fc401e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3679479 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi e04314fd 2022-05-30T01:09:31 Vulkan: Include state in pipeline graph dump Each program generates a group of nodes. The group's description is the common state among all nodes. Each node contains the diff with the shared state. Arrows between nodes indicate the GraphicsPipelineTransitionBits that have caused the transition. State that is 0 is not output for brevity. Bug: angleproject:6565 Change-Id: I7b6a95efcee63f1d7d368d8e062c5c1d221fb8b9 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3673829 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi 23f213d1 2022-05-26T23:43:44 Vulkan: Make drawable size uniform As a specialization constant, it's impossible to predict the value the drawable size may take, which in turn makes it impossible to warm up the Vulkan pipeline cache at link time. Bug: angleproject:7366 Change-Id: Ia3d1860a4fcb8e3078fdcb8d02a2e0cd173ea028 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3671976 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 98c2e169 2022-05-20T16:17:49 Vulkan: Reduce pre-rotation spec const to bool The specialization constant now only dictates whether x and y should be swapped. The complete 8 possible states of rotation and y-flip are achieved by using this swap in combination with a driver uniform for x and y flip. Swapping is still a specialization constant to avoid degrading performance of dFdx/dFdy which otherwise would need both to be evaluated instead of one. On platforms which don't support pre-rotation, the specialization constant will never change and driver uniforms entirely govern y-flip. On platforms that do support pre-rotation, only two variations of the pipeline are needed. Bug: angleproject:7366 Change-Id: I73f84e89fa9349d2098fa5b21573aee57d93a30c Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3663151 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Charlie Lao <cclao@google.com>
Jamie Madill 7b880204 2022-05-25T10:09:15 Vulkan: Remove invalid ASSERT in DescriptorSetDesc. This invalid ASSERT somehow was undetected pre-commit when landing the descriptor set cache CLs. It was code from before a refactor that broke the ASSERT's invariant condition. Bug: angleproject:7354 Change-Id: I4e84c78911e4c6aad826b1acabcd87216884509c Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3666250 Commit-Queue: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Auto-Submit: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 11e9c835 2022-05-22T23:52:35 Vulkan: Dump graphics pipeline cache graph Bug: angleproject:6565 Change-Id: Ica8fdc40ece3e1c66fae8a4f8f70a3d33374fdb8 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3656639 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi e4a517c9 2022-05-17T23:58:23 Vulkan: Rearrange GraphicsPipelineDesc to reduce footprint Bits for VK_EXT_extended_dynamic_state2 are moved to the end of the desc. Some bits are also rearranged and some paddings removed for a total of 4 bytes of reduction in desc size. Bug: angleproject:7328 Change-Id: I48f20408a8a2cb7bce0d8e2d0d2ccd8a121f7894 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3652747 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Shahbaz Youssefi bb3afdf4 2022-05-17T16:08:06 Drop support for 64xMSAA+ Hardware that supports anything more than 32xMSAA is rare (practically only Nvidia). That high number of samples is hardly useful either way. This change reduces the number of words for the sample mask to 1, reducing the amount of state needed to track it. Bug: angleproject:7328 Change-Id: Iea9add1cbeef494ff9bb383b10c82b839d1e53a4 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3652738 Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi 7c83a363 2022-05-16T22:22:44 Vulkan: Dynamic state for rasterizer discard enable Interaction with primitives generated query are tested by those tests added in https://chromium-review.googlesource.com/c/angle/angle/+/2976181 Bug: angleproject:5906 Change-Id: I0ab9f54995504be770a93bf13337a5ffe20bf7eb Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3651582 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 2faabfe5 2022-05-16T16:19:03 Vulkan: Optimize pipeline desc w.r.t dynamic state State corresponding to VK_EXT_extended_dynamic_state is moved to the back of the description and is excluded from the hash. Bug: angleproject:5906 Change-Id: I3b6efb2674e955eed6e12b6c7096a7fea63748a9 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3651581 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 891a4682 2022-05-12T22:32:07 Vulkan: Dynamic state for stencil test/op Bug: angleproject:5906 Change-Id: I436c6c238a15e54919a74d2a697d1f114bb3eaf9 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3647207 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi 15ecf068 2022-05-12T10:33:00 Vulkan: Dynamic state for depth bounds This state is actually unused. EXT_depth_bounds_test (which can set this state) is a desktop GL extension. If ever implemented, it should be implemented as a dynamic state. Bug: angleproject:5906 Change-Id: If613008952bb80cd98858f33eac7cdf31267ca29 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3646428 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi 4e5da2d2 2022-05-11T13:58:07 Vulkan: Dynamic state for cull mode Bug: angleproject:5906 Change-Id: I3d7b888e7999d4892ff71d636ea16a2edcf7a27f Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3642800 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Jamie Madill 21ad9b3c 2022-04-07T09:57:26 Vulkan: Add generic descriptors for DS cache. With the new design, the descriptor set cache keys include all identifying information needed to reconstruct the update descriptor sets calls except the specific resource handles. The places for the resource handles are held by serials intead. When we miss the cache, we no longer need a second step to then construct the update calls, and can build the update calls directly from the key structures in combination with a list of resource handles. Bug: angleproject:6776 Change-Id: If1660a557585a75e9aa2560d6a38c56b62f555c8 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3484981 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Jamie Madill d8d396db 2022-04-07T09:57:25 Vulkan: Add shared descriptor set caches. This allows programs with the same sets of descriptors to share descriptor sets. Currently there is no cache eviction. This CL adds a new "Meta" class to manage the descriptor set caches. Each shared descriptor pool is unique to a descriptor set layout. The descriptor set cache is moved into the pool class. Now every instance of a descriptor pool in ANGLE has easy access to a descriptor set cache as well. Bug: angleproject:6776 Change-Id: I06982e0349f5a87e4578e769fa356ce8e7ab49f0 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3424660 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 4ffab3bf 2022-05-10T16:17:54 Vulkan: Dynamic state for stencil reference Bug: angleproject:5906 Change-Id: I1aaf54208b173ca58ff1afd2900eca7ee78726cf Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3638990 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 1d8227da 2022-05-10T16:00:39 Vulkan: Dynamic state for stencil write mask Bug: angleproject:5906 Change-Id: I74adf56ec0b7b251ab3c5204b68b062d5fbc91eb Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3638989 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi c3def6fa 2022-05-10T14:28:04 Vulkan: Dynamic state for stencil compare mask Bug: angleproject:5906 Change-Id: Ie581c6e9fe15c90c7a6d0c7c246dd5b1b30b6507 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3638988 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Shahbaz Youssefi dcaa18b9 2022-05-10T12:27:15 Vulkan: Dynamic state for blend color Bug: angleproject:5906 Change-Id: If450e0d84410069126027142414586181fd5f0de Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3638986 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 536d6f57 2022-05-10T11:44:58 Vulkan: Dynamic state for depth bias Bug: angleproject:5906 Change-Id: I8fd7e3262fddf3aec855afdd3e4c1b9801040da1 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3638983 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 780f1a41 2022-05-10T00:30:35 Vulkan: Dynamic state for line width Bug: angleproject:5906 Change-Id: Iacf9c14d9d255c8048c71c725173e4764bcfe166 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3634733 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Shahbaz Youssefi e73121b1 2022-05-10T22:47:20 Vulkan: Fix VulkanPipelineCachePerfTest Bug: angleproject:5906 Change-Id: Ide86708df10679309cc2aca0df088e8595d2a8c1 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3641142 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Jamie Madill e0d00563 2022-05-03T17:02:06 Vulkan: Simplify external format Chroma code. The prior code used a const_cast and some other twiddling when we really just needed to add a separate small setting function. Also encapsulates the YcbcrConversionDesc string into a class with private data. The end goal is to refactor image view init and caching. The const_cast and the SamplerState were getting in the way. Bug: angleproject:7269 Change-Id: Ie2fb4ba848c46adc63618f313f7a68a1df5cef8f Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3625492 Reviewed-by: Trevor Black <vantablack@google.com> Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Jamie Madill 77cd0b5a 2022-04-14T13:32:07 Re-land: "Vulkan: Cache ImageView serials on texture changes." Re-land fixes BindTexImage serial caching. This significantly reduces overhead when changing textures before draw calls in the Vulkan back-end. Bug: angleproject:6776 Change-Id: I2cc03cb8a70e8c181f7658ab45df780e412bfc57 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3623860 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Auto-Submit: Jamie Madill <jmadill@chromium.org> Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Jamie Madill e1fdef90 2022-04-22T14:55:11 Revert "Vulkan: Cache ImageView serials on texture changes." This reverts commit 7d229d047f742e93ca194272da9b93959e49d75e. Reason for revert: Crash on Mac blink-web-tests: https://ci.chromium.org/ui/p/chromium/builders/try/mac-rel/979339/overview Original change's description: > Vulkan: Cache ImageView serials on texture changes. > > This significantly reduces overhead when changing textures before > draw calls in the Vulkan back-end. > > Bug: angleproject:6776 > Change-Id: I24e2938c0122c3b23db4a90ec85b7d8f459cc063 > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3583360 > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> > Reviewed-by: Yuxin Hu <yuxinhu@google.com> > Commit-Queue: Jamie Madill <jmadill@chromium.org> Bug: angleproject:6776 Change-Id: Icdf4f4a3628690fd9e074c6f400614890848be7f Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3600910 Auto-Submit: Jamie Madill <jmadill@chromium.org> Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Jamie Madill 7d229d04 2022-04-14T13:32:07 Vulkan: Cache ImageView serials on texture changes. This significantly reduces overhead when changing textures before draw calls in the Vulkan back-end. Bug: angleproject:6776 Change-Id: I24e2938c0122c3b23db4a90ec85b7d8f459cc063 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3583360 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Jamie Madill b2a1f0d2 2022-04-14T07:58:32 Track total vs per-frame descriptor set counters. This will give more consistent measurements for descriptor set caches and descriptor set allocations. Bug: angleproject:6776 Change-Id: I584b8807ad19f8393ae54cc1d88b319c8f7f9f39 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3584636 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Jamie Madill 4b381f41 2022-04-08T09:20:45 Vulkan: Fix descriptorSet perf counter values. Some counters were getting reset in multiple places, which could result in queries returning zero counts. Fix this by consolidating per-frame counter resets. Also updates how we compute cache hit/miss counters. This results in correct and consistent counts for cache accesses. Also includes a fix to not update the overlay when there are no enabled widgets. Also does away with some of the object- specific perf counters that were made to track descriptor set allocations. Bug: angleproject:6776 Change-Id: I769c715986defc50f0cfd0d997c338d34174e9f0 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3573389 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi 041c4c6d 2022-03-25T16:30:03 Vulkan: Track color attachment usage like D/S in render pass That is in preparation for optimizing mid-render-pass clears, which requires an answer to the following query: "has this color image been read from / written to so far in the render pass?" With this change, a future CL will also be able to optimize color attachment invalidates, which currently break the render pass unconditionally, the same way depth/stencil is optimized. Bug: angleproject:5048 Change-Id: I3d3ee40d8444e6861c06340d5d52b17f5ee895b4 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3542989 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Lingfeng Yang <lfy@google.com> Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Mohan Maiya f9ade9fb 2022-02-18T15:08:54 Vulkan: Add feature to create pipeline during glLinkProgram Creating the pipeline will trigger the compilation of shaders to byte code thus warming up Vulkan shader caches. Typically most apps call into glLinkPrograms during app loadtime and the goal is to improve cache hit rate and reduce CPU workload during game play. Bug: angleproject:7046 Test: ProgramBinary*CreatePipelineDuringLink* Change-Id: I71351d45a9aa84e220ca38503735e94cff1dcf98 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3478354 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: mohan maiya <m.maiya@samsung.com>
Jamie Madill ef17f38a 2022-03-01T10:15:26 Vulkan: Add overlay widget for cache key size. Bug: angleproject:6776 Change-Id: I35ab18bc5919129b2decf58d541499f771140e47 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3472754 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Jamie Madill c874943b 2022-02-18T15:29:14 Rename FastUnordered types to FlatUnordered. These names are consistent with the common parlance. Flat indicates the values are packed in memory, while fast indicates that lookup is as fast as possible. Bug: angleproject:6776 Change-Id: I7b56af26d7fdbf5956872be5033c3aa3f6b1b8c7 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3484978 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Commit-Queue: Jamie Madill <jmadill@chromium.org> Auto-Submit: Jamie Madill <jmadill@chromium.org>
Jamie Madill 8ade4c2f 2022-02-16T13:14:54 Vulkan: Add overlay and stats for descriptor set caches. This adds more overlay widgets for different metrics of descriptor sets, including for uniforms and for all sets. Bug: angleproject:6776 Change-Id: If1a053bd711454227e7ccbdb01202a712a5f9787 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3469228 Reviewed-by: Charlie Lao <cclao@google.com> Auto-Submit: Jamie Madill <jmadill@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Steven Noonan 760253ce 2022-01-27T07:24:27 Vulkan: replace uniforms FastIntegerMap with FastUnorderedMap This is in preparation for removing FastIntegerSet and FastIntegerMap. Bug: angleproject:6954 Change-Id: Ifa0a5c5ca39c705a9f774558bb6cad72ed8f0bc1 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3420952 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Yuxin Hu e5045587 2022-02-11T22:29:55 Debug Overlay Demo Add a new debug overlay to display texture descriptor set cache size Bug: angleproject:6976 Change-Id: I8bb2d88d4550947f1c6cbd732f936e3e060b1b04 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3457356 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Yuxin Hu <yuxinhu@google.com>
Shahbaz Youssefi d285d6a7 2021-11-23T23:12:00 Vulkan: Use VK_EXT_depth_clip_control Bug: angleproject:6679 Change-Id: Iea3b7a2d8db630dcea19836732de8f9e0a7a568f Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3298356 Reviewed-by: mohan maiya <m.maiya@samsung.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 038adcae 2022-02-08T16:46:40 Vulkan: Render the overlay in the graphics pipeline Bug: angleproject:6976 Change-Id: I388d429f0726b4d6a1c4ecd446ead93579a14a1b Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3448643 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Jamie Madill 70c510eb 2022-02-03T07:54:29 Vulkan: Use common storage for descriptor set descs. This lets us avoid using templates when implementing descriptor set caches. As an added bonus, the key size now scales with the number of active items in the cache. In the case of using a smaller set of active textures, our keys now use sizeof(uint32_t) * (kFastBufferWordLimit) + sizeof(void *) bytes of storage (132/136) instead of sizeof(uint32_t) * 3 * 96 bytes (1152) per set. This reduction should improve cache coherency. Bug: angleproject:6776 Change-Id: I904fc5527ff033af927e25abe6d13d98142e3f14 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3436767 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Jamie Madill 959d4be5 2022-02-03T08:03:14 Vulkan: Remove buffer get method from descriptors desc. This will make a future refactor to use a common desc class easier to implement. We have the necessary buffer already in the call context so we can instead pass this buffer down to where it gets used. Bug: angleproject:6776 Change-Id: Ia9782cba9f91c9604b1e937252bb293a346d025c Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3436766 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi 665ddccd 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>
Jamie Madill 168e0ae5 2022-01-27T16:50:00 Vulkan: Refactor descriptor pool code. This uses vk::Context to init DynamicDescriptorPool, and removes some accesses to ProgramExecutableVk member variables so the methods can more easily be moved into another class. The changes should have no effective behaviour difference. Note that the immutable sampler descriptor count is entirely set by the format of the immutable samplers. However we can't determine this count by looking at the descriptor set layout desc with the current implementation. Bug: angleproject:6776 Change-Id: I0a323d19fa3720b5afd8300401ed9c30b105c0ef Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3424657 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Lingfeng Yang 6f1c39b5 2021-12-16T16:10:49 Vulkan: Restore color attachment mask-related blend states When we end up resetting blend state such as when changing to a depth-only FBO and back, we can end up adding attachment states. They need to be updated with the current GL blend state. Bug: b/210543392 Bug: angleproject:6828 Change-Id: I3419acb8081eb7d662576a75382bb8684cc0d29f Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3346320 Reviewed-by: Cody Northrop <cnorthrop@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Lingfeng Yang <lfy@google.com>
Mohan Maiya 36fcf80b 2021-12-11T20:55:53 Vulkan: Consolidate SamplerYcbcrConversionCache Now that SamplerDesc and ImageHelper have the fully qualified description of a VkSamplerYcbcrConversion, we can lazy allocate the conversion handle along with streamlining the interface to the SamplerYcbcrConversionCache class. Bug: angleproject:6732 Test: Texture2DTestES3.*Yuv*Vulkan Change-Id: I7deaa0220031bc4c500d88021459017bf937218c Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3333629 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: mohan maiya <m.maiya@samsung.com>
Mohan Maiya 44d0ef69 2021-11-25T14:37:38 Vulkan: Enhance SamplerDesc for YUV formats In order to uniquely identify a VkSamplerYcbcrConversion, having just the format is insufficient. We need to account for conversion model and color component range. Refactor SamplerDesc by encapsulating YUV related fields into its own YcbcrConversionDesc. Bug: angleproject:6732 Test: Texture2DTestES3.*Yuv*Vulkan Change-Id: I23114c50646232dccde988a52b0649778ad72a90 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3301899 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Mohan Maiya <m.maiya@samsung.com>
Jamie Madill 4a2446c6 2021-10-19T11:55:17 Vulkan: Implement robust shader outputs. In this CL we change ANGLE to pass the mask of missing shader outputs down to pipeline creation. We then use the color mask bits to block SwiftShader writing to unused outputs. This fixes the undefined behaviour present in Genshin Impact. Note that the other GLES implementations we tested don't seem to modify outputs even if they're unused. It was easier to mask out the color attachments in initialize rather than set up the pipeline desc to mask out the attachments. This was because we manipulate the color mask in a fairly complex way before we initialize the pipeline desc. Bug: angleproject:6566 Change-Id: Ie659fcd511cd286fa573fd25e3e6a0b9e123ebd6 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3232435 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com>
Jamie Madill 191c236a 2021-10-14T17:01:45 Vulkan: Remove "current queue serial". Queries, semaphores, and pipelines instead use the normal vk::Resource design to track their lifetimes. Removes the current serial APIs from all classes. Current serials are still tracked internally in the command queue classes. Bug: b/169788986 Change-Id: Idcd2c2a93bc8225c6f3f7c247eb8fcfb76be1030 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3223644 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi e637e4c9 2021-10-18T13:54:00 Vulkan: Optimize updating blend state in pipeline desc Updating blend funcs and equations always updated all 8 slots. Now that's only done for the attachments that are present. Bug: angleproject:6298 Change-Id: I58fa7e4dfa27d05fef54cc9d56c7b2aa5ef43dd8 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3202550 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com>
Jamie Madill 1dc2b702 2021-10-13T16:02:37 Vulkan: Pass shader module map to init pipelines. Instead of pulling out the shader modules from the shader map when we start the call chain, pull them out right in the init call. This saves a bunch of boilerplate code. Refactoring change only. Bug: angleproject:6566 Change-Id: Ib8d79bd7284d7ddb83522270f3d4df9086ab7300 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3221134 Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Cheryl Wei b95cdec3 2021-10-12T11:26:15 Vulkan: Switch to `std::unordered_map` to cache renderpass ANGLE use `flat_hash_map` by default,while it will not retain pointer stability which causes random null pointer issue when running MH31 with asynccommandqueue.This change switch to using `std::unordered_map` to cache renderpass. Bug: angleproject:6436 Bug: angleproject:6569 Change-Id: Ia2829a41251eea39ce2a79a9a42b6fe9dff6b120 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3219010 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 371542cd 2021-08-23T23:10:09 Vulkan: Use VK_EXT_load_store_op_none With read-only depth/stencil attachments, ANGLE utilizes storeOp=NONE to optimize memory bandwidth (by avoiding write back of tile memory at the end of the render pass). Simultaneoulsy, this avoids a synchronization hazard with the next write to that depth/stencil image. If a framebuffer contains a depth/stencil attachment but it's unused, ANGLE utilizes loadOp=NONE/storeOp=NONE to effectively remove any memory bandwidth wasted on the attachment. Bug: angleproject:5371 Change-Id: I76cbadbf1194041532ac4b690ffe087298f2de51 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3114232 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com>
Charlie Lao 8ea87a67 2021-08-17T18:46:36 Vulkan: Avoid texture format fallback when possible Some texture formats are not renderable on some hardware. For example, R4G4B4A4 are not renderable on nvidia and not blendable on ARM. R5G5B5A1 are also not blendable on nvidia. Right now when we generate format table, we are being most conservative, picking an actual format that is always renderable and blendable. This means when R4G4B4A4 is used on one of these GPUs, we are always falling back to R8G8B8A8 regardless if the texture is actually being used as color attachment or not. This CL adds a actualRenderableImageFormatID field in vk::Format. Initially we will pick actualImageFormatID which only ensures texture sample capability. If later on the texture is being attached to FBO, then we will switch to actualRenderableImageFormatID and do data copy if necessary. This way we save memory and reduce texture bandwidth for most usage of these textures. For renderBuffer and surfaces and EGLImages, we always pick the renderable textures. Bug: b/196456356 Change-Id: I02eec3365c2a317b0d1bad6dbdc3e741114c5bba Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3104514 Reviewed-by: Ian Elliott <ianelliott@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Charlie Lao <cclao@google.com>
Mohan Maiya 5c8bf081 2021-06-08T13:12:24 Vulkan: Add support for YUV internal format extension 1. Add a function to upload YUV image data to textures 2. Modify stageSubresourceUpdate method to account for YUV images 3. Create VkSamplerYcbcrConversion when initializing ImageHelper 4. Update hasImmutableSampler to account for native YUV format support 5. Skip initializeNonZeroMemory for YUV formats Bug: angleproject:5773 Test: Texture2DTestES3.TexStorage2D*Yuv*Vulkan* Change-Id: I270f04bbf903cf2bf19f100eb95f32953d491c39 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2947767 Commit-Queue: Mohan Maiya <m.maiya@samsung.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Cheryl Wei 45965e72 2021-06-17T10:21:26 Vulkan: Translate border color's channel by image view format ANGLE implementes some formats as other formats,such as ALPHA8 to R8, this caused some tests failed due to missing border color's channel translation,this change add a new textureBorderLoadFunction to translate channels of border color by image view format. Bug: angleproject:6046 Change-Id: I94ce719b4db3724ffd3dc862b51a412b5d9f3cce Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2972328 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Shahbaz Youssefi 5b314268 2021-06-15T17:37:45 Vulkan: Support OVR_multiview and OVR_multiview2 Multiview is supported in Vulkan simply by specifying the number of views in the render pass, and creating the appropriate image views. A number of changes to the way image views and render targets are stored are made to support those that don't cover the entire range of layers. One particular detail that is not implemented in this change is the use of queries in combination with multiview. Vulkan specifies that N queries are actually produced (N being the number of views) which must be summed by the application, but this is not currently done. Bug: angleproject:6048 Change-Id: I1d4a9894c232d3a93d7a97c9fa0eedc334e57469 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2967625 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com>
Shahbaz Youssefi 9bad7a19 2021-06-15T23:39:52 Vulkan: Unpack RenderPassDesc Upcoming multiview support needs to add 2~3 more bits to this struct, but this struct is already fully packed. As the combination with emulated multisampled render to texture is not planned, some bits used for MSRTT can be aliased with multiview. However, that makes the packing of this struct even more unwieldy. Since only tens of render passes are expected per frame, increasing the size of the render pass cache key should not have a dramatic effect. Bug: angleproject:6048 Change-Id: I5c9a0d155f23a37e2787b38525d3ce721b54ec1e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2966218 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Cody Northrop 851fbedb 2021-06-10T21:27:01 Vulkan: Switch viewport and scissor to dynamic state Heavily based on http://crrev/c/1316888 Some apps are creating a large number of viewport combinations and are running out of graphics memory. This CL drops their graphics pipeline use from tens of thousands to tens. Performance testing shows little impact to application traces. Bug: b/190026813 Bug: angleproject:3143 Change-Id: Ib7415be1128f8fedae4a7ca72e067b2815201223 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2954925 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Mohan Maiya dd312cc6 2021-05-21T08:49:48 Reland "Vulkan: Add support for EXT_texture_border_clamp" This is a reland of 4b92e089c01e94730fe26c4c743d76d6c935088d Initializes mPadding to 0 in SamplerDesc::update Original change's description: > Vulkan: Add support for EXT_texture_border_clamp > Add support for GL_EXT_texture_border_clamp. This is implemented by > using VK_EXT_custom_border_color. > > Bug: angleproject:3577 > Test: dEQP-GLES31.functional.texture.border_clamp* > Change-Id: Ie9fa1eb5dd03b997b5ae182787641a53080a9e51 > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2830192 > Reviewed-by: Jamie Madill <jmadill@chromium.org> > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> > Commit-Queue: Mohan Maiya <m.maiya@samsung.com> Bug: angleproject:3577 Change-Id: I8684242c4bce6e1a006dbe926defaa495fcc2282 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2911571 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Mohan Maiya <m.maiya@samsung.com>
Yuly Novikov 44fabb7b 2021-05-21T00:51:06 Revert "Vulkan: Add support for EXT_texture_border_clamp" This reverts commit 4b92e089c01e94730fe26c4c743d76d6c935088d. Reason for revert: breaks "Linux MSan Tests" bot https://ci.chromium.org/ui/p/chromium/builders/ci/Linux%20MSan%20Tests/28489/overview Original change's description: > Vulkan: Add support for EXT_texture_border_clamp > > Add support for GL_EXT_texture_border_clamp. This is implemented by > using VK_EXT_custom_border_color. > > Bug: angleproject:3577 > Test: dEQP-GLES31.functional.texture.border_clamp* > Change-Id: Ie9fa1eb5dd03b997b5ae182787641a53080a9e51 > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2830192 > Reviewed-by: Jamie Madill <jmadill@chromium.org> > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> > Commit-Queue: Mohan Maiya <m.maiya@samsung.com> Bug: angleproject:3577, angleproject:5996 Change-Id: I6b62f1a4a43fcfe2d0df2c73de24db83606ce430 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2911533 Reviewed-by: Yuly Novikov <ynovikov@chromium.org> Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
Mohan Maiya 4b92e089 2021-05-12T09:20:24 Vulkan: Add support for EXT_texture_border_clamp Add support for GL_EXT_texture_border_clamp. This is implemented by using VK_EXT_custom_border_color. Bug: angleproject:3577 Test: dEQP-GLES31.functional.texture.border_clamp* Change-Id: Ie9fa1eb5dd03b997b5ae182787641a53080a9e51 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2830192 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Mohan Maiya <m.maiya@samsung.com>
Shahbaz Youssefi 9e8fea5b 2021-05-13T07:04:05 Vulkan: Fix desc set cache bug with xfb offset Prior to this change, the transform feedback buffer offset was not stored in the descriptor set key, so if the offset changed, stale descriptor sets could be used. Bug: angleproject:5963 Change-Id: I3dec4ab9fa82092a65e9a75bdd19c5f2cf49521c Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2894513 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Jamie Madill ccc0fbaa 2021-03-24T17:56:38 Vulkan: Related fixes for buffer descriptor set cache. Includes some stats counter gathering and a few related refactors and cleanups. Also includes a new overlay widget. Bug: angleproject:5736 Change-Id: Ida8d2cd815c5b598c6a442dd9bbfdf51e9c05180 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2785431 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Tim Van Patten <timvp@google.com>
Jamie Madill 0c77f3ad 2021-03-10T15:58:00 Vulkan: Implement shader buffers descriptor cache. Implements a descriptor set cache for UBOs, SSBOs, and atomic counter buffers. Storage Images and framebuffer fetch input attachments are not yet included. Requires moving the buffer barrier handling into ContextVk, similarly to how we handle the barriers for Textures. The packed description key for the descriptors uses a "fast" vector with a basic minimum size. For most cases of a few buffers this will fit easily in stack memory, but for larger programs with many buffers we fit this into heap memory. The key has a large upper bound due to the high ES 3.2 requirements and the need to index several values such as the offset and binding size. We use dynamic offsets for uniform buffers when possible. This ensures applications like Manhattan 3.1 that use sets of common buffers with changing offsets hit the cache most of the time. Because of resource limits we pick at compilation time whether to use dynamic or static descriptor sets. Mostly this applies to tests that use a large number of uniform buffers. A future implementation could be smart and would recompile the program with heuristics to use a minimal number of dynamic indices. Reduces the number of descriptor set updates from ~300 -> ~30 per frame in Manhattan 3.1 and in Asphalt 9 from 900+ to as low as 0 per frame. Bug: angleproject:5736 Change-Id: I5c2a3881bec90d301dab15cc86c8a70e60674ad7 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2757515 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Jamie Madill efc53d0f 2021-04-03T21:30:15 Vulkan: Refactor to descriptor offset array. This is in preparation for using dynamic offsets for other sets than the default descriptors. Bug: angleproject:5736 Change-Id: I546e688ad57a35b380f8d7f303ffc0c62773ac18 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2804813 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Tim Van Patten <timvp@google.com>
kmx.io     mail     discord kmxgit v0.4.0