Skip to content

kernel-builder: support capabilities 10.3 and 12.1 #783

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
danieldk merged 1 commit into from
Aug 25, 2026
Merged

kernel-builder: support capabilities 10.3 and 12.1 #783

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 23 additions & 0 deletions examples/kernels/flake.nix
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,6 +15,7 @@
inherit (kernel-builder.inputs.nixpkgs) lib;

cudaVersion = "cu126";
cuda13Version = "cu132";
rocmVersion = "rocm71";
xpuVersion = "xpu20253";
torchVersion = "212";
Expand All @@ -39,6 +40,7 @@
path = ./cpp20-symbols;
drv = sys: out: out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-cpu-${sys}"};
}
# This test should check the capabilities of the oldest supported CUDA.
{
name = "relu-kernel";
path = ./relu;
Expand All @@ -55,6 +57,27 @@
"9.0"
];
}
# This test should check the capabilities of the latest supported CUDA.

@sayakpaul sayakpaul Aug 25, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

So cool!

{
name = "relu-kernel-cu13";
path = ./relu;
drv =
sys: out:
out.packages.${sys}.redistributable.${"torch${torchVersion}-cxx11-${cuda13Version}-${sys}"};
checkCudaCapabilities = [
"7.5"
"8.0"
"8.6"
"8.7"
"8.9"
"9.0"
"10.0"
"10.3"
"11.0"
"12.0"
"12.1"
];
}
{
# Check arch intersection. 5.0 is dropped because it is not supported.
name = "relu-archs-subset";
Expand Down
5 changes: 4 additions & 1 deletion kernel-builder/src/pyproject/templates/torch/preamble.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -98,7 +98,10 @@ option(BUILD_ALL_SUPPORTED_ARCHS "Build all supported architectures" off)

if(DEFINED CMAKE_CUDA_COMPILER_VERSION AND
CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 13.0)
set(CUDA_DEFAULT_KERNEL_ARCHS "7.5;8.0;8.6;8.7;8.9;9.0;10.0;11.0;12.0;12.1+PTX")
set(CUDA_DEFAULT_KERNEL_ARCHS "7.5;8.0;8.6;8.7;8.9;9.0;10.0;10.3;11.0;12.0;12.1+PTX")

@sayakpaul sayakpaul Aug 25, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🔥

elseif(DEFINED CMAKE_CUDA_COMPILER_VERSION AND
CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 12.9)
set(CUDA_DEFAULT_KERNEL_ARCHS "7.0;7.2;7.5;8.0;8.6;8.7;8.9;9.0;10.0;10.1;10.3;12.0;12.1+PTX")
elseif(DEFINED CMAKE_CUDA_COMPILER_VERSION AND
CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 12.8)
set(CUDA_DEFAULT_KERNEL_ARCHS "7.0;7.2;7.5;8.0;8.6;8.7;8.9;9.0;10.0;10.1;12.0+PTX")
Expand Down
5 changes: 4 additions & 1 deletion kernel-builder/src/pyproject/templates/tvm_ffi/preamble.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -77,7 +77,10 @@ if(GPU_LANG STREQUAL "CUDA")

if(DEFINED CMAKE_CUDA_COMPILER_VERSION AND
CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 13.0)
set(CUDA_DEFAULT_KERNEL_ARCHS "7.5;8.0;8.6;8.7;8.9;9.0;10.0;11.0;12.0;12.1+PTX")
set(CUDA_DEFAULT_KERNEL_ARCHS "7.5;8.0;8.6;8.7;8.9;9.0;10.0;10.3;11.0;12.0;12.1+PTX")
elseif(DEFINED CMAKE_CUDA_COMPILER_VERSION AND
CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 12.9)
set(CUDA_DEFAULT_KERNEL_ARCHS "7.0;7.2;7.5;8.0;8.6;8.7;8.9;9.0;10.0;10.1;10.3;12.0;12.1+PTX")
elseif(DEFINED CMAKE_CUDA_COMPILER_VERSION AND
CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 12.8)
set(CUDA_DEFAULT_KERNEL_ARCHS "7.0;7.2;7.5;8.0;8.6;8.7;8.9;9.0;10.0;10.1;12.0+PTX")
Expand Down
Loading