| Age | Commit message (Collapse) | Author |
|---|
|
An earlier refactoring pass over the compiler codebase split the
type that had been called `CompileRequest` into three distinct
pieces:
* `FrontEndCompileRequest` which was supposed to own state and
options related to running the compiler front end and producing
IR + reflection (e.g., what translation units and source
files/strings are included).
* `BackEndCompileRequest` which was supposed to own state and options
related to running the compiler back end to translate the IR
for a `ComponentType` (program) into output code. (Note that the
`BackEndCompileRequest` was conceived of as orthogonal to the
`TargetRequest`s, which store per-target and target-specific
options.)
* `EndToEndCompileRequest` which was an umbrella object that owns
separate front-end and back-end requests, plus any state that is
only relevant when doing a true end-to-end compile (such as the
kinds of compiles initiated with `slangc`). As originally conceived,
the only state that this type was supposed to own was stuff related
to "pass-through" compilation, as well as state related to writing
of generated code to output files.
That refactoring work was very useful at the time, because it allowed
us to "scrub" the back end compilation steps to remove all
dependencies on front-end and AST state (this was important for our
goals of enabling linking and codegen from serialized Slang IR).
At this point, however, it is clear that the hierarchy that was built
up serves very little purpose:
* The `BackEndCompileRequest` type is only used in two places:
* As part of an `EndToEndCompileRequest`, where the settings on
the `BackEndCompileRequest` can be configured, but only through
the `EndToEndCompileRequest`
* As part of on-demand code generation through the `IComponentType`
APIs. In this case, the settings stored on the
`BackEndCompileRequest` are not accessible to the application
at all, and will always use their default values, so that
instantiating a "request" object doesn't really make any sense.
* The `FrontEndCompileRequest` type has a similar situation:
* Front-end compilation as part of an `EndToEndCompileRequest`
supports user configuration of `FrontEndCompileRequest` settings,
but only through the `EndToEndCompileRequest`
* Front-end compilation triggered by an `import` or a `loadModule()`
call does not support user configuration of settings at all. It
will always derive all relevant settings from thsoe on the
session ("linkage").
In addition, subsequent changes have been made to the compiler that
show a bit of a "code smell" and/or forward-looking worries for this
decomposition:
* In some cases we've had to add the same setting to multiple types
in the breakdown (front-end, back-end, end-to-end, linkage, target,
etc.) which makes it harder for us to validate that all the possible
mixtures of state work correctly.
* Related to the above, in some cases we have manual logic that copies
state from one of the objects in the breakdown to another, in order
to ensure that the user's intention is actually followed.
* As a forward-looking concern, it seems that developers have sometimes
added new configuration options and state to places that don't really
make sense according to the rationale of the original decomposition
(e.g., we probably don't want to have a lot of state that is only
available via end-to-end requests, given that the API structure is
meant to push users *away* from end-to-end compiles).
As a result of all of the above, I've been planning a large refactor
with the following big-picture goals:
* Eliminate `BackEndCompileRequest`
* Move all relevant state/options from the back-end request to
the end-to-end request, since that is the only place they could
be set anyway.
* Introduce a transient "context" type to be used for the duration
of code generation that serves the main functions that back-end
requests really served in the codebase
* Make `EndToEndCompileRequest` be a subclass of
`FrontEndCompileRequest`
* Consider addding a transient "context" type for front-end
compiles that can be used in `import`-like cases rather than
needing a full front-end request object. If this works, then
eliminate `FrontEndCompileRequest` and be back to world with
just a single `CompileRequest` type
* Move *all* compiler configuration options to a distinct type (named
something like `CompilerConfig` or `CompilerOptions` or whatever)
which stores setting as key-value pairs, and has a notion of
"inheritance" such that one configuration can extend or build on top
of another. Make all the relevant types use this catch-all structure
instead of redundantly storing flags in many places.
This change deals with the first of those bullets: removeal of
`BackEndCompileRequest`. The addition of the `CodeGenContext` type is
perhaps an unncessary additional step, but making that change helps
clean up a bunch of the code related to per-target code generation,
so I think it is the right choice.
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
* Framework for texture views testing working; Small tweaks to texture testing utils; Changed D3D12 readTextureResource to account for non-1 depth
* Basic framework for texture views tests working; Test file needs a rename at some point
* 1D, 2D, and 3D textures working for ShaderResource, UnorderedAccess, and RenderTarget tests; Fixed some small issues with handling the depth field of 3D textures in Vulkan causing incorrectly cleared textures
|
|
* Fixed the mapping of the *InstanceID() and *InstanceIndex() functions to GLSL.
* Fixed and somewhat improved the vkray/closesthit test.
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Added sample-grad-clamp-lod sample.
* Fix handling multiple reopenings of namespaces.
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Meta-2 test works.
* Add new generic test for static const variable in a function in a generic.
* Generic function with static const variable doesn't work.
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Added sample-grad-clamp-lod sample.
|
|
Improved the type printing function to include the generic substitutions and parent types.
Added a test for it, mismatching-types.slang
|
|
|
|
from texture copying tests into a new util file to make them accessible to all gfx unit tests (#2169)
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Fix for = {} initialization with a field that is generic type parameter.
* Handling for if a non type is passed to a generic parameter which requires a type.
* Small comment improvements.
Fix some tab issues.
* This fixes the matrix.slang issue. Move the matrix.slang test into bugs as generic-default-matrix.slang
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Split doc extractor such that can be used in C++ extractor.
* Compiles. Update the stdlib docs.
* Fix issue on release builds.
* Add support for extracting documentation to C++ extractor.
* Dump out markup.
Make enum value backing type take tokens.
* Node::Type -> Node::Kind
* More improvements around Node::Type -> Node::Kind
* Support for parsing callable types.
* Fix issue params for callable, and default value for variable.
* Add support for static.
* Improve handling parsing of contained types.
* Small improvements around template consumption.
* Improve dumping with markup/static.
* Small improvements around reflection.
* Add more flexible handling of markers.
Allow reflection without markers.
* Handling external "C"
unsigned/signed
|
|
* Working on finding a better way to do result comparisons in order to accommodate testing different formats and texture types, copying small to large texture currently half broken
* All tests except copying into a buffer with an offset working with new code for results comparison, support for most formats WIP
* Replaced SubresourceStuff with ValidationTextureData; Added implementation for ValidationTextureFormat::initializeTexel() and changed generateTextureData() to call this
* All implemented formats work on D3D12, a few broken for Vulkan
* Fixed numerous locations in render-vk.cpp to set aspect masks based on either the format or the provided TextureAspect; All tests except copy to buffer with offset working on new code structure
* All texture to texture copying tests work for all non-compressed formats; Test code cleaned up to minimize the amount of stuff needing to be passed between functions
* All tests working with all texture types except TextureCube and all non-compressed formats; Temporarily removing test for copying to a buffer at an offset, will likely return in a separate test file
* Small cleanup changes
* build fixes
|
|
|
|
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Split doc extractor such that can be used in C++ extractor.
* Compiles. Update the stdlib docs.
* Fix issue on release builds.
* Add support for extracting documentation to C++ extractor.
* Dump out markup.
Make enum value backing type take tokens.
* Node::Type -> Node::Kind
* More improvements around Node::Type -> Node::Kind
* Support for parsing callable types.
* Fix issue params for callable, and default value for variable.
* Add support for static.
* Improve handling parsing of contained types.
* Small improvements around template consumption.
* Improve dumping with markup/static.
* Small improvements around reflection.
* Add more flexible handling of markers.
Allow reflection without markers.
|
|
Use `ShortList` instead of `Array` for `m_waitHandles`.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
queue submit. (#2158)
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Split doc extractor such that can be used in C++ extractor.
* Compiles. Update the stdlib docs.
* Fix issue on release builds.
* Add support for extracting documentation to C++ extractor.
* Dump out markup.
Make enum value backing type take tokens.
* Node::Type -> Node::Kind
* More improvements around Node::Type -> Node::Kind
|
|
* Vulkan: deferred shader compilation and pipeline creation.
* Fix 32bit build.
* gfx: restructure the code in render-d3d12.cpp
* Move `Submitter`.
* Fix.
* merge with master.
* Revert dictionary change in previous PR.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* Vulkan: deferred shader compilation and pipeline creation.
* Fix 32bit build.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* Added function to IDevice that retrieves the row alignment for the particular API; Added rowDstStride argument to copyTextureToBuffer and changed D3D12 footprint row pitch to check that the user-supplied stride is correctly aligned before assigning to the footprint's row pitch
* Changed alignment from Uint to size_t
Co-authored-by: jsmall-nvidia <jsmall@nvidia.com>
|
|
* Pass through the downstream compiler error messages if they are not recognized.
* Added a help message that is printed on -h, -help, --help.
Added -version as an alias for -v.
* Fixed the bug in -lang option processing.
|
|
* Use SlangResult value. Make legacy SLANG_ERROR_ macros use SlangResult values.
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Fix some typos around command line docs.
|
|
|
|
|
|
`ImageSubscript` for GLSL (#2146)
|
|
* Added isOccluded() and setFullScreenMode() to ISwapchain and provided implementations for D3D12 (all others currently return false)
* Removed isWindowOccluded variable and directly return the result of m_swapchain3->Present
* Restart CI
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Switch on generateDebugInfo on glslang i s any debug level is set.
* Take copy of SpvOptions.
Co-authored-by: Theresa Foley <10618364+tangent-vector@users.noreply.github.com>
|
|
* gfx: d3d12 performance optimizations.
* Fix.
* Fix unit test bug.
* Add gfx interface for directly allocating GPU descriptor tables.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* Optimize d3d12 mutable shader object implementation.
* Disable mismatched clear value warning message from d3d sdk.
* Fix.
* Fix.
* gfx: Avoid redundant d3d12 QueryInterface call.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
* Added implementation for dispatchRays() and ShaderTableImpl, currently missing extensions and createShaderTable()
* Code written, working on finding and fixing bugs
* SBT issues fixed; Added implementation for endEncoding() to ensure the bound pipeline is properly reset; Ray tracing pipeline example successfully runs without any validation errors
* Fixed some incorrectly merged lines
* Fixed spacing
* Fixed alignment for member variables in VulkanApi
* Restart CI
* Removed accidental comment kept from merge; Changed getName() call to getNameOverride()
|
|
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* Various gfx fixes.
* Fix test case.
* Fix crash.
* Trigger build
* Trigger build 2
* Fix vulkan unit tests.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Small fixes.
Added compiler crash with generic defined in a function.
Added enum-flags test that works (by limiting backing type to int), and using __EnumType constraint.
* Add comment about crash.
* Disable crashing test.
* Fixes to make compile on OSX.
* Add github build for OSX.
* Make premake generator a utility.
* Fix osx compilation issue.
* More fixes for OSX build.
* OSX fix due to ambiguity around size_t and integer types.
* Disable xlib on build on osx.
* Use 'prebuildcommands' to make prebuild make utility projects do something.
* Small fixes for premake so utility works on linux/osx.
* Another hack to try and make generators run when 'utility'
* Fix typo in macos.yml.
* Revert premake to old style, and disable stdlib embedding on OSX.
* OSX testing.
* Fix pipe handling for OSX.
* Enable testing on OSX.
* Small fix because uname -p is not x64 on darwin.
* Use glslang from slang-binaries for OSX.
* Add slang-llvm on OSX.
* Upgrade slang-llvm.
* Update slang-llvm for OSX JIT fix.
* slang llvm 26
* Upgrade slang-llvm.
* Use slangUtil to get shared library filename.
* Use fixed slang-binaries.
* Another small fix.
* Rename linux build for clarity.
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Fix linux build issue with renderer-shared.h
* Move to .cpp file.
* Move some templates to stop definition order issue on linux.
* Move to header because it's templated.
* Fix warning about GeometryInstanceFlags::Enum flags : 8; being too small by making the enum backed by 8 bits.
* Check if vkDestroySampler function is available.
|
|
* gfx: Various Vulkan fixes.
* Fix test case.
* Fix vulkan crash.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Small fixes.
Added compiler crash with generic defined in a function.
Added enum-flags test that works (by limiting backing type to int), and using __EnumType constraint.
* Add comment about crash.
* Disable crashing test.
* Fixes to make compile on OSX.
* Add github build for OSX.
* Make premake generator a utility.
* Fix osx compilation issue.
* More fixes for OSX build.
* OSX fix due to ambiguity around size_t and integer types.
* Disable xlib on build on osx.
* Use 'prebuildcommands' to make prebuild make utility projects do something.
* Small fixes for premake so utility works on linux/osx.
* Another hack to try and make generators run when 'utility'
* Fix typo in macos.yml.
* Revert premake to old style, and disable stdlib embedding on OSX.
* OSX testing.
* Fix pipe handling for OSX.
* Enable testing on OSX.
* Small fix because uname -p is not x64 on darwin.
* Use glslang from slang-binaries for OSX.
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Small fixes.
Added compiler crash with generic defined in a function.
Added enum-flags test that works (by limiting backing type to int), and using __EnumType constraint.
* Add comment about crash.
* Disable crashing test.
* Fixes to make compile on OSX.
* Add github build for OSX.
* Make premake generator a utility.
* Fix osx compilation issue.
* More fixes for OSX build.
* OSX fix due to ambiguity around size_t and integer types.
* Disable xlib on build on osx.
* Use 'prebuildcommands' to make prebuild make utility projects do something.
* Small fixes for premake so utility works on linux/osx.
* Another hack to try and make generators run when 'utility'
* Fix typo in macos.yml.
* Revert premake to old style, and disable stdlib embedding on OSX.
* OSX testing.
* Fix pipe handling for OSX.
* Enable testing on OSX.
* Small fix because uname -p is not x64 on darwin.
|
|
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* Various fixes to gfx.
* Fix.
* Fixes.
* Fix.
* gfx: support root parameter via user-defined attribute.
* Fix.
* Fix.
* Skip d3d12 tests on win x86.
* Fixes.
* gfx: support shader record overwrite.
* Fix QueyPool implementation.
* Rename to `getBindingRangeLeafVariable`
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* Various fixes to gfx.
* Fix.
* Fixes.
* Fix.
* gfx: support root parameter via user-defined attribute.
* Fix.
* Fix.
* Skip d3d12 tests on win x86.
* Fixes.
Co-authored-by: Yong He <yhe@nvidia.com>
Co-authored-by: jsmall-nvidia <jsmall@nvidia.com>
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Small fixes.
Added compiler crash with generic defined in a function.
Added enum-flags test that works (by limiting backing type to int), and using __EnumType constraint.
* Add comment about crash.
* Disable crashing test.
* Fixes to make compile on OSX.
* Add github build for OSX.
* Make premake generator a utility.
* Fix osx compilation issue.
* More fixes for OSX build.
* OSX fix due to ambiguity around size_t and integer types.
* Disable xlib on build on osx.
* Use 'prebuildcommands' to make prebuild make utility projects do something.
* Small fixes for premake so utility works on linux/osx.
* Another hack to try and make generators run when 'utility'
* Fix typo in macos.yml.
* Revert premake to old style, and disable stdlib embedding on OSX.
|
|
* Various fixes to gfx.
* Fix.
* Fixes.
* Fix.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* Checkpoint
* Fixed problem with global variables in supported resource states test, test is functional for both Vulkan and D3D12
* Removed a comment
* Added a loop over all existing formats
* Rerun tests
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Small fixes.
Added compiler crash with generic defined in a function.
Added enum-flags test that works (by limiting backing type to int), and using __EnumType constraint.
* Add comment about crash.
* Disable crashing test.
|
