Skip to content

Commit 776e9d6

Browse files
ThireusThireus
authored
Merge pull request #1 from thevishalagarwal/github-workflow/update-cuda-12.8
GitHub workflow/update cuda 12.8
2 parents 8726392 + d971e0a commit 776e9d6

File tree

4 files changed

+51
-36
lines changed

4 files changed

+51
-36
lines changed

‎.github/actions/windows-setup-cuda/action.yml‎

Lines changed: 26 additions & 26 deletions
Original file line numberDiff line numberDiff line change
@@ -36,32 +36,32 @@ runs:
3636
echo "CUDA_PATH=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7" | Out-File -FilePath $env:GITHUB_ENV -Append -Encoding utf8
3737
echo "CUDA_PATH_V11_7=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.7" | Out-File -FilePath $env:GITHUB_ENV -Append -Encoding utf8
3838
39-
- name: Install Cuda Toolkit 12.4
40-
if: ${{ inputs.cuda_version == '12.4' }}
39+
- name: Install Cuda Toolkit 12.8
40+
if: ${{ inputs.cuda_version == '12.8' }}
4141
shell: pwsh
4242
run: |
43-
mkdir -p "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4"
43+
mkdir -p "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8"
4444
choco install unzip -y
45-
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_cudart/windows-x86_64/cuda_cudart-windows-x86_64-12.4.127-archive.zip"
46-
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvcc/windows-x86_64/cuda_nvcc-windows-x86_64-12.4.131-archive.zip"
47-
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvrtc/windows-x86_64/cuda_nvrtc-windows-x86_64-12.4.127-archive.zip"
48-
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/libcublas/windows-x86_64/libcublas-windows-x86_64-12.4.5.8-archive.zip"
49-
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvtx/windows-x86_64/cuda_nvtx-windows-x86_64-12.4.127-archive.zip"
50-
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_profiler_api/windows-x86_64/cuda_profiler_api-windows-x86_64-12.4.127-archive.zip"
51-
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/visual_studio_integration/windows-x86_64/visual_studio_integration-windows-x86_64-12.4.127-archive.zip"
52-
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvprof/windows-x86_64/cuda_nvprof-windows-x86_64-12.4.127-archive.zip"
53-
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_cccl/windows-x86_64/cuda_cccl-windows-x86_64-12.4.127-archive.zip"
54-
unzip '*.zip' -d "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4"
55-
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\cuda_cudart-windows-x86_64-12.4.127-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y
56-
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\cuda_nvcc-windows-x86_64-12.4.131-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y
57-
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\cuda_nvrtc-windows-x86_64-12.4.127-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y
58-
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\libcublas-windows-x86_64-12.4.5.8-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y
59-
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\cuda_nvtx-windows-x86_64-12.4.127-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y
60-
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\cuda_profiler_api-windows-x86_64-12.4.127-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y
61-
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\visual_studio_integration-windows-x86_64-12.4.127-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y
62-
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\cuda_nvprof-windows-x86_64-12.4.127-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y
63-
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\cuda_cccl-windows-x86_64-12.4.127-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" /E /I /H /Y
64-
echo "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
65-
echo "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4\libnvvp" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
66-
echo "CUDA_PATH=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" | Out-File -FilePath $env:GITHUB_ENV -Append -Encoding utf8
67-
echo "CUDA_PATH_V12_4=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.4" | Out-File -FilePath $env:GITHUB_ENV -Append -Encoding utf8
45+
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_cudart/windows-x86_64/cuda_cudart-windows-x86_64-12.8.90-archive.zip"
46+
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvcc/windows-x86_64/cuda_nvcc-windows-x86_64-12.8.93-archive.zip"
47+
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvrtc/windows-x86_64/cuda_nvrtc-windows-x86_64-12.8.93-archive.zip"
48+
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/libcublas/windows-x86_64/libcublas-windows-x86_64-12.8.4.1-archive.zip"
49+
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvtx/windows-x86_64/cuda_nvtx-windows-x86_64-12.8.90-archive.zip"
50+
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_profiler_api/windows-x86_64/cuda_profiler_api-windows-x86_64-12.8.90-archive.zip"
51+
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/visual_studio_integration/windows-x86_64/visual_studio_integration-windows-x86_64-12.8.90-archive.zip"
52+
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_nvprof/windows-x86_64/cuda_nvprof-windows-x86_64-12.8.90-archive.zip"
53+
curl -O "https://developer.download.nvidia.com/compute/cuda/redist/cuda_cccl/windows-x86_64/cuda_cccl-windows-x86_64-12.8.90-archive.zip"
54+
unzip '*.zip' -d "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8"
55+
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\cuda_cudart-windows-x86_64-12.8.90-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8" /E /I /H /Y
56+
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\cuda_nvcc-windows-x86_64-12.8.93-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8" /E /I /H /Y
57+
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\cuda_nvrtc-windows-x86_64-12.8.93-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8" /E /I /H /Y
58+
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\libcublas-windows-x86_64-12.8.4.1-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8" /E /I /H /Y
59+
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\cuda_nvtx-windows-x86_64-12.8.90-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8" /E /I /H /Y
60+
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\cuda_profiler_api-windows-x86_64-12.8.90-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8" /E /I /H /Y
61+
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\visual_studio_integration-windows-x86_64-12.8.90-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8" /E /I /H /Y
62+
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\cuda_nvprof-windows-x86_64-12.8.90-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8" /E /I /H /Y
63+
xcopy "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\cuda_cccl-windows-x86_64-12.8.90-archive\*" "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8" /E /I /H /Y
64+
echo "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
65+
echo "C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8\libnvvp" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
66+
echo "CUDA_PATH=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8" | Out-File -FilePath $env:GITHUB_ENV -Append -Encoding utf8
67+
echo "CUDA_PATH_V12_8=C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.8" | Out-File -FilePath $env:GITHUB_ENV -Append -Encoding utf8

‎.github/workflows/build.yml‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -844,7 +844,7 @@ jobs:
844844

845845
strategy:
846846
matrix:
847-
cuda: ['12.4', '11.7']
847+
cuda: ['12.8', '11.7']
848848

849849
steps:
850850
- name: Clone

‎.github/workflows/release.yml‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -377,7 +377,7 @@ jobs:
377377

378378
strategy:
379379
matrix:
380-
cuda: ['12.4', '11.7']
380+
cuda: ['12.8', '11.7']
381381

382382
steps:
383383
- name: Clone

‎ggml/src/ggml-cuda/CMakeLists.txt‎

Lines changed: 23 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -15,6 +15,7 @@ if (CUDAToolkit_FOUND)
1515
# 80 == Ampere, asynchronous data loading, faster tensor core instructions
1616
# 86 == RTX 3000, needs CUDA v11.1
1717
# 89 == RTX 4000, needs CUDA v11.8
18+
# 120 == RTX 5000, needs CUDA v12.8
1819
#
1920
# XX-virtual == compile CUDA code as PTX, do JIT compilation to binary code on first run
2021
# XX-real == compile CUDA code as device code for this specific architecture
@@ -23,19 +24,33 @@ if (CUDAToolkit_FOUND)
2324
# The default behavior for a non-native is to build virtual architectures as needed to cover all features needed
2425
# for best performance and to also build real architectures for the most commonly used GPUs.
2526
if (GGML_NATIVE AND CUDAToolkit_VERSION VERSION_GREATER_EQUAL "11.6" AND CMAKE_VERSION VERSION_GREATER_EQUAL "3.24")
27+
# Use the GPUs available on this system
2628
set(CMAKE_CUDA_ARCHITECTURES "native")
27-
elseif(GGML_CUDA_F16 OR GGML_CUDA_DMMV_F16)
28-
if (CUDAToolkit_VERSION VERSION_GREATER_EQUAL "11.8")
29-
set(CMAKE_CUDA_ARCHITECTURES "60-virtual;61-virtual;70-virtual;75-virtual;80-virtual;86-real;89-real")
29+
else()
30+
set(ARCH_LIST "")
31+
32+
# Base architectures - depending on feature flags
33+
if (GGML_CUDA_F16 OR GGML_CUDA_DMMV_F16)
34+
# FP16 support (Pascal and newer)
35+
list(APPEND ARCH_LIST "60-virtual")
3036
else()
31-
set(CMAKE_CUDA_ARCHITECTURES "60-virtual;61-virtual;70-virtual;75-virtual;80-virtual;86-real")
37+
# Maxwell and newer
38+
list(APPEND ARCH_LIST "50-virtual")
3239
endif()
33-
else()
40+
41+
# Always included after base architecture assuming CUDA toolkit version is 11.1 or higher
42+
list(APPEND ARCH_LIST "61-virtual" "70-virtual" "75-virtual" "80-virtual" "86-real")
43+
44+
# Version-dependent architectures for newer GPUs
3445
if (CUDAToolkit_VERSION VERSION_GREATER_EQUAL "11.8")
35-
set(CMAKE_CUDA_ARCHITECTURES "50-virtual;61-virtual;70-virtual;75-virtual;80-virtual;86-real;89-real")
36-
else()
37-
set(CMAKE_CUDA_ARCHITECTURES "50-virtual;61-virtual;70-virtual;75-virtual;80-virtual;86-real")
46+
list(APPEND ARCH_LIST "89-real")
3847
endif()
48+
49+
if (CUDAToolkit_VERSION VERSION_GREATER_EQUAL "12.8")
50+
list(APPEND ARCH_LIST "120-real")
51+
endif()
52+
53+
set(CMAKE_CUDA_ARCHITECTURES ${ARCH_LIST})
3954
endif()
4055
endif()
4156
message(STATUS "Using CUDA architectures: ${CMAKE_CUDA_ARCHITECTURES}")

0 commit comments

Comments
 (0)