| Age | Commit message (Collapse) | Author |
|---|
|
(#8482)
cmake's find_package() will only find the configs when they are in a
slang/ subdirectory.
Co-authored-by: Mukund Keshava <mkeshava@nvidia.com>
|
|
This PR resolves a configuration conflict where the
SLANG_ENABLE_SPIRV_TOOLS_MIMALLOC option was defined twice with
different defaults, causing user settings to be overridden.
Changes:
- Removed duplicate definition in external/CMakeLists.txt that only
applied to WIN32
- Define default in main CMakeLists.txt
- ON by default for Windows
- OFF by default for other platforms (still need to resolve crashes so
set it to `OFF` for now)
Preserved user configuration - explicit
-DSLANG_ENABLE_SPIRV_TOOLS_MIMALLOC=value settings are now properly
respected
Related to https://github.com/shader-slang/slang/issues/8158
---------
Co-authored-by: slangbot <ellieh+slangbot@nvidia.com>
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
|
* Added optional mimalloc integration for SPIRV-Tools to improve
compilation performance
* Cloning `mimalloc` repo to external/ on demand during cmake config
* Enabled by default on Windows, configurable via
`SLANG_ENABLE_SPIRV_TOOLS_MIMALLOC` CMake option, to enable static
mimalloc build in spirv-tools.
* There are some crashes in Linux and Mac with static mimalloc enabled,
likely due to the system malloc and mimalloc mixed usage. This might be
expected as mimalloc in spirv-tools is not extensively tested according
to
https://github.com/KhronosGroup/SPIRV-Tools?tab=readme-ov-file#dependency-on-mimalloc.
So by default only Windows has this enabled.
Close: https://github.com/shader-slang/slang/issues/8158
---------
Co-authored-by: slangbot <ellieh+slangbot@nvidia.com>
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
|
This PR marks the `slang-rhi` a required dependecy for `platform` and
`gfx-unit-test`, and only build them when `SLANG_ENABLE_SLANG_RHI=ON`.
This should allow the slang still to be built without those tests
components when `SLANG_ENABLE_SLANG_RHI=OFF`.
---------
Co-authored-by: slangbot <ellieh+slangbot@nvidia.com>
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
|
Allows opt-in for sourcing the following dependencies from the system,
instead of using the vendored ones:
- miniz
- lz4
- vulkan-headers
- spirv-tools
- glslang
(some of these already had options that weren't working, as either it
expected them to be static libraries or it was expecting to be embedded
in another CMakeList that should provide the package, instead of finding
the package itself)
This is based on a patch we currently maintain inside nixpkgs, but as it
frequently conflicts with new slang releases, it would be nice to see
get
this upstream.
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
Co-authored-by: Sam Estep <sam@samestep.com>
|
|
find_package is locally scoped, so like other find_packages we should do
it at the top level
Closes https://github.com/shader-slang/slang/issues/7643
|
|
* print tool versions for formatter
* formatting
* bump cmake formatter version
|
|
Closes https://github.com/shader-slang/slang/issues/7309
|
|
* Reset minimum cmake version to 3.22
* update slang-rhi
|
|
* Revert "Disable OptiX tests by default. (#1331)"
This reverts commit e45f8c1f49855cebe90b6722324ec24146ff5a3d.
* Enable optix submodule to build
Add support for default entry points in compilation
Implemented logic to check for defined entry points in the module
when no explicit entry points are provided. If found, these entry points
are added to the `specializedEntryPoints` list, with the assumption that
no specialization is needed for them at this time.
* Disable optix if cuda is not enabled
* Add submodule OptixSDK path in search
* Distinguish user-explicit vs auto-detected SLANG_ENABLE_OPTIX
When SLANG_ENABLE_OPTIX is explicitly set by user and CUDA is not available,
show SEND_ERROR to maintain strict validation. When OptiX is auto-detected
(e.g., local submodule present) but CUDA unavailable, gracefully disable
with STATUS message to allow builds to continue.
This addresses review feedback to keep error for explicit requests while
handling auto-detection gracefully.
* Apply CMake formatting to SLANG_ENABLE_OPTIX validation logic
* revert: slang-rhi changes
as those are merged independently as in PR # slang-rhi#400
|
|
Fixes: #7453
|
|
fixes: #7417
Define was called `SLANG_ENABLE_IR_BREAK_ALLOC,`. This is incorrect, supposed to be `SLANG_ENABLE_IR_BREAK_ALLOC`.
|
|
* Remove unused compile definitions
No need for full external path; slang is already linking with the
SPIRV-Headers cmake target which adds the proper include flags
|
|
* Disable Link-Time-Optimization by default
LTO was requested for the release package a while ago.
When we added it, LTO was enabled by default although it was needed only for
the release packages.
Later we found that the Release build cannot be incrementally recompiled
when LTO is enabled. It sometimes works fine, but it required full recompilation
when it doesn't work. We added a new CMake option, `SLANG_ENABLE_RELEASE_LTO`,
to disable it for developers. But many Slang developers don't know the
option exists.
I was going to update the document, CONTRIBUTING.md, but I thought it
will be better to change the default behavior.
* Fix a compiler warning treated as an error on linux
A padding variable was uninitialized, which is fine, but the compiler
was complaining about it.
* Fix other gcc error for uninitialized variable
* Fix more compile warning treated as error
* Fix compiler warning from gcc 11
It appears that this is a valid warning that the `delete this` is done
on an offset 8 when the class uses multiple inheritance.
The compiler warning is following:
```
In file included from /home/runner/work/slang/slang/source/core/slang-memory-file-system.h:5,
from /home/runner/work/slang/slang/tools/slang-unit-test/unit-test-module-ptr.cpp:3:
In destructor ‘virtual Slang::ComBaseObject::~ComBaseObject()’,
inlined from ‘uint32_t Slang::ComBaseObject::_releaseImpl()’ at /home/runner/work/slang/slang/source/core/slang-com-object.h:49:16,
inlined from ‘virtual uint32_t Slang::MemoryFileSystem::release()’ at /home/runner/work/slang/slang/source/core/slang-memory-file-system.h:34:5:
/home/runner/work/slang/slang/source/core/slang-com-object.h:33:31: error: ‘void operator delete(void*, std::size_t)’ called on pointer ‘<unknown>’ with nonzero offset 8 [-Werror=free-nonheap-object]
33 | virtual ~ComBaseObject() {}
| ^
In destructor ‘virtual Slang::ComBaseObject::~ComBaseObject()’,
inlined from ‘uint32_t Slang::ComBaseObject::_releaseImpl()’ at /home/runner/work/slang/slang/source/core/slang-com-object.h:49:16,
inlined from ‘virtual uint32_t Slang::MemoryFileSystem::release()’ at /home/runner/work/slang/slang/source/core/slang-memory-file-system.h:34:5,
inlined from ‘Slang::ComPtr<T>::~ComPtr() [with T = ISlangMutableFileSystem]’ at /home/runner/work/slang/slang/include/slang-com-ptr.h:113:34,
inlined from ‘void _modulePtr_impl(UnitTestContext*)’ at /home/runner/work/slang/slang/tools/slang-unit-test/unit-test-module-ptr.cpp:92:1:
/home/runner/work/slang/slang/source/core/slang-com-object.h:33:31: error: ‘void operator delete(void*, std::size_t)’ called on pointer ‘<unknown>’ with nonzero offset 8 [-Werror=free-nonheap-object]
33 | virtual ~ComBaseObject() {}
| ^
/home/runner/work/slang/slang/tools/slang-unit-test/unit-test-module-ptr.cpp: In function ‘void _modulePtr_impl(UnitTestContext*)’:
/home/runner/work/slang/slang/tools/slang-unit-test/unit-test-module-ptr.cpp:36:69: note: returned from ‘void* operator new(std::size_t)’
36 | ComPtr<ISlangMutableFileSystem>(new Slang::MemoryFileSystem());
| ^
```
The problem is on the fact that `ComBaseObject` is not the first in the
multiple inheritance:
```
class MemoryFileSystem : public ISlangMutableFileSystem, public ComBaseObject
{
public:
// ISlangUnknown
SLANG_COM_BASE_IUNKNOWN_ALL
```
It should be:
```
class MemoryFileSystem : public ComBaseObject, public ISlangMutableFileSystem
```
The chain of ComObject release is little complicated and it is easy to
make a mistake. Here is summary with details,
1. `release()` is declared as a pure-virtual in ISlangUnknown, which is
one of the base classes of `ISlangMutableFileSystem`.
```
struct ISlangUnknown
{
virtual SLANG_NO_THROW uint32_t SLANG_MCALL release() = 0;
```
2. `release()` is implemented with the macro
`SLANG_COM_BASE_IUNKNOWN_RELEASE`.
```
SLANG_NO_THROW uint32_t SLANG_MCALL release() SLANG_OVERRIDE \
{ \
return _releaseImpl(); \
}
inline uint32_t ComBaseObject::_releaseImpl()
{
// Check there is a ref count to avoid underflow
SLANG_ASSERT(m_refCount != 0);
const uint32_t count = --m_refCount;
if (count == 0)
{
delete this;
}
return count;
}
```
3. The instance of `MemoryFileSystem` is handled by ComPtr. And
`ComPtr::~ComPtr()` calls the `release()`.
```
ComPtr<ISlangMutableFileSystem> memoryFileSystem =
ComPtr<ISlangMutableFileSystem>(new Slang::MemoryFileSystem());
SLANG_FORCE_INLINE ~ComPtr()
{
if (m_ptr)
((Ptr)m_ptr)->release();
}
```
4. When `delete this` is called, because ComBaseObject is not the first
in the multiple inheritance, `this` is 8 byte off from the actual
instance address.
A fix for this is to change the order of the inheritance and make
ComBaseObject to be the first in the order.
|
|
* Update build to allow setting external lua path
Update the build to allow setting user-specific path for the external
module lua.
* T
* Fix an include path
|
|
* Update build to allow setting more external paths
Update the build to allow setting user-specific paths for the external
modules: glm, imgui, slang-rhi, and tinyobjloader.
|
|
* Add Slang Byte Code generation and interpreter.
* Fix compile issues.
* format code
* More compile fix.
* Fix clang issue.
* Fix more clang issues.
* Another clang fix.
* Fix clang issues.
* Fix another clang issue.
* Fix wasm build.
* Update building.md
* Fix test-server.
* Fix compile error.
* Fix bug.
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
|
* Use GITHUB_TOKEN for fetching prebuilt
This PR extends Commit c6b702c to use GITHUB_TOKEN if set for fetching
prebuilt binaries.
This allows webgpu-dawn and slang-tint to be downloaded for certain IPs
where the github API rate is limited.
Fixes #6689
* Don't ignore download failure if github token is provided
* Update readme for getting github access token
* format code
* combine cmake_parse_arguments calls
* format code
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
Co-authored-by: Jay Kwak <82421531+jkwak-work@users.noreply.github.com>
Co-authored-by: Ellie Hermaszewska <ellieh@nvidia.com>
|
|
|
|
* Ignore failure to fetch webgpu_dawn and slang-tint
Fixes #6683
* format code
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
|
Add option to exclude webgpu-dawn and slang-tint from the build,
for environments that need to build Slang without internet access.
|
|
* Update build to allow setting external paths
Update the build to allow setting user-specific paths for the external modules.
This allows building Slang without also fetching the external modules, assuming
they are already present elsewhere locally.
|
|
* SP004: implement initialize list translation to ctor
- We synthesize a member-wise constructor for each struct follow
the rules described in SP004.
- Add logic to translate the initialize list to constructor invoke
- Add cuda-host decoration for the synthesized constructor
- Remove the default constructor when we have a valid member init constructor
- Disable -zero-initialize option, will re-implement it in followup (#6109).
- Fix the overload lookup issue
When creating invoke expression for ctor, we need to call
ResolveInvoke() to find us the best candidates, however
the existing lookup logic could find us the base constructor
for child struct, we should eliminate this case by providing
the LookupOptions::IgnoreInheritance to lookup, this requires
us to create a subcontext on SemanticsVisitor to indicate that
we only want to use this option on looking the constructor.
- Do not implicit initialize a struct that doesn't have explicit default
constructor.
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
|
* Distribute slang-glsl-module.bin in release packages.
* Fix.
* fix2.
* Build glsl modlue into a dll.
|
|
* Add SLANG_ENABLE_RELEASE_LTO cmake option
* Fix cmake static build
* Disable install SlangTargets to avoid static build failing
|
|
|
|
* Remove unnecessary warnings on windows
* Correctly set debug flags on gcc
* Emit debug info for Release builds
* Perform LTO for relwithdebinfo builds
* Release from release builds not relwithdebinfo
|
|
* Split debug info for all targets
Work towards https://github.com/shader-slang/slang/issues/5724
* release separate debug info
Closes https://github.com/shader-slang/slang/issues/5724
* Add split debug info support for MacOS
* Add SLANG_ENABLE_SPLIT_DEBUG_INFO option
* Sign and package debug info on MacOS
* Set --build-id where available
* Correct debug info installing
* Keep cpack macos signing workaround
* Neaten cmake
* Disable sccache if building split debug info on Windows
* Only repack necessary files on MacOS releases
|
|
Closes https://github.com/shader-slang/slang/issues/5764
Also mention other installed targets in cmake config
|
|
|
|
* Modify package config
* Apply formatting.
* Make sure build works for Emscripten
* Add documentation on install target.
---------
Co-authored-by: obhi-d <obi.de.online@gmail.com>
|
|
It seems that release workflows were unexpectedly broken when STDLIB
was renamed to CORE_MODULE.
Changed the option default because I think it just makes more sense.
Removed explicit calls to enable it. Fixed docs.
|
|
* format cmake files
* format code
---------
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
|
|
The directory name of `source/slang-stdlib` is changed to `source/slang-core-module`.
We will use the term "core module" instead of "standard library" from now on.
|
|
This commit renames the files and projects to prefer "core-module" over
"stdlib".
The directory name `source/slang-stdlib` needs to be renamed too, and
there will be another commit for it soon.
|
|
* Split examples cmake desc
* declutter top level CMakeLists.txt
* fail if building tests without gfx
* Move llvm fetching to another cmake file
* Further split CMakeLists.txt
* Neaten llvm fetching
* Remove last premake remnant
* correct cross builds
* Neaten
* Neaten project organization in vs
|
|
* Make slang-llvm fetching and failure more robust
Improve error reporting when things go wrong.
Fall back by default to a non-llvm build
Closes https://github.com/shader-slang/slang/issues/5247
Tested sensible behavior with:
- `FETCH_BINARY_IF_POSSIBLE`, valid tag
- No errors or warnings, successful build with llvm
- `FETCH_BINARY_IF_POSSIBLE`, no valid tag
- Warning message, successful llvm build with slang-llvm from latest release
- `FETCH_BINARY_IF_POSSIBLE`, no valid tag, bad `SLANG_SLANG_LLVM_BINARY_URL` specified
- Warning message, successful no-llvm build
- `FETCH_BINARY_IF_POSSIBLE`, no valid tag, unable to fetch release information
- Warning message, successful no-llvm build
- `FETCH_BINARY`, valid tag
- No errors or warnings, successful build with llvm
- `FETCH_BINARY`, no valid tag
- Warning message, successful llvm build with slang-llvm from latest release
- `FETCH_BINARY`, no valid tag, bad `SLANG_SLANG_LLVM_BINARY_URL` specified
- Error, explaining that we couldn't fetch it
- `FETCH_BINARY`, no valid tag, unable to fetch release info
- Error, explaining that we couldn't fetch it
* Allow downloading from a local file
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
D3D12Core.dll had been copied to a wrong directory and slang has been using D3D12Core.dll from the system directory, C:\windows\system32.
D3D12Core.dll has to be copied from external/slang-binaries/bin/windows-x64 to build/Release/bin/D3D12 not to build/Release/bin.
The same is true for the debug build and it had to be copied to build/Debug/bin/D3D12 not build/Debug/bin.
It hasn't been a problem for Release build, because the debug-layer is not enabled for Release build and it didn't cause the version mismatching problem with D3D12SDKLayers.dll. The Release build was loaded from either build/Release/bin or from C:\windows\system32, and it didn't matter which one was used.
The Debug build, however, got into a problem where D3D12Core.dll was loaded from the system directory whereas D3D12SDKLayers.dll was loaded from build/Debug/bin and it failed to load D3D12.dll entirely. This caused D3D12 to be "Not supported" for "Windows/Debug" configuration. Note that our CI explicitly excludes DX12 tests for the "Windows/Debug" configuration with a command-line argument "-api all-dx12", and DX12 tests were going to be ignored anyway.
The actual problem was observed when WGPU is implemented. WGPU started printing explicit errors for the load failure of D3D12.dll.
See more detailed explanation:
https://devblogs.microsoft.com/directx/gettingstarted-dx12agility/#d3d12sdkpath-should-not-be-the-same-directory-as-the-application-exe
Closes #5305
Closes #5276
|
|
|
|
* Lower minimum CMake version to 3.22
Reverts https://github.com/shader-slang/slang/pull/4193
* Update build instructions to mention older CMake versions
|
|
* Don't use __assume for SLANG_ASSERT + build fixes.
* Fix.
* build slang-wasm conditionally
* Fix.
* revert retry open file
* revert include.
* another attempt of silencing compiler warnings.
* revert assume change.
|
|
* Fix assert when compiling an entrypoint that calls another entrypoint.
* Fix test.
|
|
* Restrict stdlib embed macros to single source file
* Build slang-without-embedded-stdlib with the same target type as libslang
To avoid building everything twice
|
|
Support for exceptions is enabled, since Slang uses them for diagnostics.
The size optimization arguments ('-Os') resolves some internal emscripten error during the
slang-wasm.wasm linking step, which happens when enabling exceptions.
("parse exception: too many locals")
|
|
* Add options to prevent usage of own submodules
Signed-off-by: Jacki <jacki@thejackimonster.de>
* Allow using external unordered dense headers
Signed-off-by: Jacki <jacki@thejackimonster.de>
* Link system wide installed unordered dense
Signed-off-by: Jacki <jacki@thejackimonster.de>
* Allow external header usage for lz4 and spirv
Signed-off-by: Jacki <jacki@thejackimonster.de>
* Add more options to disable targets
Signed-off-by: Jacki <jacki@thejackimonster.de>
* Add option to provide explizit path for spirv headers and remove earlier options that break the build process
Signed-off-by: Jacki <jacki@thejackimonster.de>
* Rename options to use common prefix
Signed-off-by: Jacki <jacki@thejackimonster.de>
* Fix indentation for the cmake changes
Signed-off-by: Jacki <jacki@thejackimonster.de>
* Add advanced_option function for cmake
* Normalize includes between system and submodule dependencies
Fix any before-accidentally-working problems
* Add option for enabling/disabling slang-rhi
Signed-off-by: Jacki <jacki@thejackimonster.de>
* Pass correct include path for cpu tests
* Correct include path
---------
Signed-off-by: Jacki <jacki@thejackimonster.de>
Co-authored-by: Ellie Hermaszewska <ellieh@nvidia.com>
|
|
* Compile fixes for Wasm
The issues are all are due to 'long' types being 32 bits on WASM.
- class members redeclared errors
- << with StringBuilder and unsigned long is ambiguous
This helps to address issue #5115.
* Use the host executable suffix for generators
Since the generators are run at build-time, we should not use CMAKE_EXECUTABLE_SUFFIX,
which is the suffix for the target platform.
Instead, define CMAKE_HOST_EXECUTABLE_SUFFIX as appropriate, and use that suffix instead.
This helps to address issue #5115.
* Add support for Wasm as a platform
This helps to address issue #5115.
* Add emscripten build
This closes #5115.
|
|
* Use GITHUB_TOKEN if set
We have been having an error from Github saying that API rate limit
exceeded for IP. It hits the limit more often if your public IP is
shared by many other collegues.
{"message":"API rate limit exceeded for 216.228.112.22. (But here's the good news: Authenticated requests get a higher rate limit. Check out the documentation for more details.)","documentation_url":"https://docs.github.com/rest/overview/resources-in-the-rest-api#rate-limiting"}
The suggested solution from GitHub is to use gh.exe tool with "auth
login" arguments. It will store a token and allow us to use API more
than the limit set for IP without token.
However, our cmake build doesn't use gh.exe and API is called via REST.
This commit adds an extra header to the HTTP request with the
infomration of the github token if the value is.
Usage: cmake.exe --preset vs2019 -DSLANG_GITHUB_TOKEN=your_token_here
* Adding a warning message to use LANG_GITHUB_TOKEN
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
* add slang-rhi submodule
* refactor render-test to use slang-rhi and remove OpenGL support
* remove -vk -glsl tests
* remove gl test
* disable failing test
* allow recursive submodules in github actions
* update slang-rhi
* update slang-rhi
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
* Dont require llvm for building generators
* Fetch slang-llvm.so from correct release (#4847)
* Fetch slang-llvm.so from correct release
Closes https://github.com/shader-slang/slang/issues/4648
Should close https://github.com/shader-slang/slang/issues/4812
* Update docs
* Correct documentation on cmake option
* Neaten cmake script
* Fix fetching on windows
---------
Co-authored-by: Yong He <yonghe@outlook.com>
* Be a bit more careful dealing with release list fetching failure
* clarify error messages
---------
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
This reverts commit 579d59c761cb0460701e0e820eb066275fe357b3.
|
