diff options
| author | Simon Kallweit <64953474+skallweitNV@users.noreply.github.com> | 2025-05-07 20:14:02 +0200 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-05-07 11:14:02 -0700 |
| commit | eb1198fcc7f0dacb1167d119aad7ddfc47f84c78 (patch) | |
| tree | 3c33299763bdc4f93ee69798ef831ed76939d4e7 | |
| parent | ccdb2e39da37753961f3694d0f90e676bf859006 (diff) | |
[CUDA] Fix surface write intrinsics (#7004)
* fix cuda surface write intrinsics
* format code (#7023)
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
---------
Co-authored-by: Mukund Keshava <mkeshava@nvidia.com>
Co-authored-by: slangbot <ellieh+slangbot@nvidia.com>
Co-authored-by: slangbot <186143334+slangbot@users.noreply.github.com>
| -rw-r--r-- | prelude/slang-cuda-prelude.h | 14 |
1 files changed, 8 insertions, 6 deletions
diff --git a/prelude/slang-cuda-prelude.h b/prelude/slang-cuda-prelude.h index 91ff98a17..7c68384b3 100644 --- a/prelude/slang-cuda-prelude.h +++ b/prelude/slang-cuda-prelude.h @@ -1346,7 +1346,7 @@ SLANG_FORCE_INLINE SLANG_CUDA_CALL void surf1Dwrite_convert<float2>( { const float vx = v.x, vy = v.y; asm volatile( - "{sust.p.1d.b32." SLANG_PTX_BOUNDARY_MODE " [%0, {%1}], {%2,%3};}\n\t" ::"l"(surfObj), + "{sust.p.1d.v2.b32." SLANG_PTX_BOUNDARY_MODE " [%0, {%1}], {%2,%3};}\n\t" ::"l"(surfObj), "r"(x), "f"(vx), "f"(vy)); @@ -1362,7 +1362,7 @@ SLANG_FORCE_INLINE SLANG_CUDA_CALL void surf2Dwrite_convert<float2>( { const float vx = v.x, vy = v.y; asm volatile( - "{sust.p.2d.b32." SLANG_PTX_BOUNDARY_MODE " [%0, {%1,%2}], {%3,%4};}\n\t" ::"l"(surfObj), + "{sust.p.2d.v2.b32." SLANG_PTX_BOUNDARY_MODE " [%0, {%1,%2}], {%3,%4};}\n\t" ::"l"(surfObj), "r"(x), "r"(y), "f"(vx), @@ -1380,7 +1380,8 @@ SLANG_FORCE_INLINE SLANG_CUDA_CALL void surf3Dwrite_convert<float2>( { const float vx = v.x, vy = v.y; asm volatile( - "{sust.p.2d.b32." SLANG_PTX_BOUNDARY_MODE " [%0, {%1,%2,%3}], {%4,%5};}\n\t" ::"l"(surfObj), + "{sust.p.2d.v2.b32." SLANG_PTX_BOUNDARY_MODE + " [%0, {%1,%2,%3}], {%4,%5};}\n\t" ::"l"(surfObj), "r"(x), "r"(y), "r"(z), @@ -1398,7 +1399,8 @@ SLANG_FORCE_INLINE SLANG_CUDA_CALL void surf1Dwrite_convert<float4>( { const float vx = v.x, vy = v.y, vz = v.z, vw = v.w; asm volatile( - "{sust.p.1d.b32." SLANG_PTX_BOUNDARY_MODE " [%0, {%1}], {%2,%3,%4,%5};}\n\t" ::"l"(surfObj), + "{sust.p.1d.v4.b32." SLANG_PTX_BOUNDARY_MODE + " [%0, {%1}], {%2,%3,%4,%5};}\n\t" ::"l"(surfObj), "r"(x), "f"(vx), "f"(vy), @@ -1416,7 +1418,7 @@ SLANG_FORCE_INLINE SLANG_CUDA_CALL void surf2Dwrite_convert<float4>( { const float vx = v.x, vy = v.y, vz = v.z, vw = v.w; asm volatile( - "{sust.p.2d.b32." SLANG_PTX_BOUNDARY_MODE + "{sust.p.2d.v4.b32." SLANG_PTX_BOUNDARY_MODE " [%0, {%1,%2}], {%3,%4,%5,%6};}\n\t" ::"l"(surfObj), "r"(x), "r"(y), @@ -1437,7 +1439,7 @@ SLANG_FORCE_INLINE SLANG_CUDA_CALL void surf3Dwrite_convert<float4>( { const float vx = v.x, vy = v.y, vz = v.z, vw = v.w; asm volatile( - "{sust.p.2d.b32." SLANG_PTX_BOUNDARY_MODE + "{sust.p.2d.v4.b32." SLANG_PTX_BOUNDARY_MODE " [%0, {%1,%2,%3}], {%4,%5,%6,%7};}\n\t" ::"l"(surfObj), "r"(x), "r"(y), |
