DeepSeek/DeepGEMM
1
0
Fork 0
mirror of https://github.com/deepseek-ai/DeepGEMM synced 2025-06-26 23:15:49 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Add DG_NVCC_OVERRIDE_CPP_STANDARD

Browse Source
This commit is contained in:
Chenggang Zhao
2025-04-03 15:53:29 +08:00
parent 3a5539b7db
commit d14962f072
2 changed files with 3 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
deep_gemm/jit/compiler.py
View File

@@ -96,7 +96,8 @@ def put(path, data, is_binary=False):
def build(name: str, arg_defs: tuple, code: str) -> Runtime:
# Compiler flags
nvcc_flags = ['-std=c++20', '-shared', '-O3', '--expt-relaxed-constexpr', '--expt-extended-lambda',
cpp_standard = int(os.getenv('DG_NVCC_OVERRIDE_CPP_STANDARD', 20))
nvcc_flags = [f'-std=c++{cpp_standard}', '-shared', '-O3', '--expt-relaxed-constexpr', '--expt-extended-lambda',
'-gencode=arch=compute_90a,code=sm_90a',
'--ptxas-options=--register-usage-level=10' + (',--verbose' if 'DG_PTXAS_VERBOSE' in os.environ else ''),
# Suppress some unnecessary warnings, such as unused variables for certain `constexpr` branch cases