Making it easier to work with shaders https://git.yummers.dev/slang.git/atom?h=master 2025-10-16T03:59:47+00:00 2025-10-16T03:59:47+00:00 Yong He yonghe@outlook.com 2025-10-16T03:59:47+00:00 urn:sha1:01510f2c922af8629c7a730ef92a31fa83bd9f49 This PR implements `Access.Immutable` to allow pointers to immutable data. The new type `ImmutablePtr<T>` is defined as an alias of `Ptr<T, Address.Immutable>`. By forming a immutable pointer, the programmer is conveying to the compiler that the data at the pointer address will never change during the execution of the current program. Therefore loads from immutable pointers can be deduplicated by the compiler, and will translate to `__ldg` when generating code for CUDA. The SPIRV backend is not changed in this PR, since the current SPIRV spec makes it very difficult to specify loads from immutable address without generating tons of wrappers and boilerplate type declarations. We would like to see the spec evolved a bit to around its support of `NonWritable` physical storage pointers or immutable loads before we attempt to express such immutability in SPIRV. For now we simply emit ordinary pointers and loads when generating spirv. --------- Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com> 2025-10-01T02:08:23+00:00 Yong He yonghe@outlook.com 2025-10-01T02:08:23+00:00 urn:sha1:e4611e2e30a3e5969d402f5ed7e72706a0e3b024 This allows us to specialize functions whose argument is a sub element of a constant buffer, instead of being only applicable to entire buffer element. Closes #8421. This change also implements a proper heuristic to determine when to specialize the calls and defer the buffer loads. This PR addresses a pathological case exposed in `slangpy\slangpy\benchmarks\test_benchmark_tensor.py`, which used to take 27ms to finish, and now takes 1.25ms. For example, given: ``` struct Bottom { float bigArray[1024]; [mutating] void setVal(int index, float value) { bigArray[index] = value; } } struct Root { Bottom top[2]; [mutating] void setTopVal(int x, int y, float value) { top[x].setVal(y, value); } } RWStructuredBuffer<Root> sb; [shader("compute")] [numthreads(1, 1, 1)] void compute_main(uint3 tid: SV_DispatchThreadID) { sb[0].setTopVal(1, 2, 100.0f); } ``` We are now able to specialize the call to `setTopVal` into: ``` void compute_main(uint3 tid: SV_DispatchThreadID) { setTopVal_specialized(0, 1, 2, 100.0f); } void setTopVal_specialized(int sbIdx, int x, int y, float value) { Bottom_setVal_specialized(sbIdx, x, y, value); } void Bottom_setVal_specialized(int sbIdx, int x, int y, float value) { sb[sbIdx].top[x].bigArray[y] = value; } ``` And get rid of all unnecessary loads. Achieving this requires a combination of function call specialization and buffer-load-defer pass. The buffer-load-defer pass has been completely rewritten to be more correct and avoid introducing redundant loads. This PR also adds tests to make sure pointers, bindless handles, and loads from structured buffer or constant buffers works as expected.