Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in
Toggle navigation
Menu
Open sidebar
OpenDAS
vllm_cscc
Commits
94144e72
Unverified
Commit
94144e72
authored
Sep 10, 2024
by
Tyler Michael Smith
Committed by
GitHub
Sep 10, 2024
Browse files
[CI/Build][Kernel] Update CUTLASS to 3.5.1 tag (#8043)
parent
1d5e397a
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
13 additions
and
2 deletions
+13
-2
CMakeLists.txt
CMakeLists.txt
+13
-2
No files found.
CMakeLists.txt
View file @
94144e72
...
@@ -195,9 +195,13 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
...
@@ -195,9 +195,13 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
FetchContent_Declare
(
FetchContent_Declare
(
cutlass
cutlass
GIT_REPOSITORY https://github.com/nvidia/cutlass.git
GIT_REPOSITORY https://github.com/nvidia/cutlass.git
# CUTLASS 3.5.1
GIT_TAG v3.5.1
GIT_TAG 06b21349bcf6ddf6a1686a47a137ad1446579db9
GIT_PROGRESS TRUE
GIT_PROGRESS TRUE
# Speed up CUTLASS download by retrieving only the specified GIT_TAG instead of the history.
# Important: If GIT_SHALLOW is enabled then GIT_TAG works only with branch names and tags.
# So if the GIT_TAG above is updated to a commit hash, GIT_SHALLOW must be set to FALSE
GIT_SHALLOW TRUE
)
)
FetchContent_MakeAvailable
(
cutlass
)
FetchContent_MakeAvailable
(
cutlass
)
...
@@ -231,6 +235,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
...
@@ -231,6 +235,7 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
"-gencode arch=compute_90a,code=sm_90a"
)
"-gencode arch=compute_90a,code=sm_90a"
)
endif
()
endif
()
#
#
# Machete kernels
# Machete kernels
...
@@ -289,6 +294,12 @@ define_gpu_extension_target(
...
@@ -289,6 +294,12 @@ define_gpu_extension_target(
USE_SABI 3
USE_SABI 3
WITH_SOABI
)
WITH_SOABI
)
# If CUTLASS is compiled on NVCC >= 12.5, it by default uses
# cudaGetDriverEntryPointByVersion as a wrapper to avoid directly calling the
# driver API. This causes problems when linking with earlier versions of CUDA.
# Setting this variable sidesteps the issue by calling the driver directly.
target_compile_definitions
(
_C PRIVATE CUTLASS_ENABLE_DIRECT_CUDA_DRIVER_CALL=1
)
#
#
# _moe_C extension
# _moe_C extension
#
#
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment