lzhengning
/
JittorMirror
mirror of https://github.com/Jittor/Jittor
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

fix multiple different gpu archs compile error

This commit is contained in:
Dun Liang 2020-09-13 14:42:52 +08:00
parent 35b22799fd
commit f3e99b96bc
3 changed files with 8 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
python/jittor/__init__.py
View File

@ -7,7 +7,7 @@
# This file is subject to the terms and conditions defined in # This file is subject to the terms and conditions defined in
# file 'LICENSE.txt', which is part of this source code package. # file 'LICENSE.txt', which is part of this source code package.
# *************************************************************** # ***************************************************************
__version__ = '1.1.7.14' __version__ = '1.1.7.15'
from . import lock from . import lock
with lock.lock_scope(): with lock.lock_scope():
from . import compiler from . import compiler

5
python/jittor/compile_extern.py
View File

@ -268,7 +268,10 @@ def install_nccl(root_folder):
tar.extractall(root_folder) tar.extractall(root_folder)
LOG.i("installing nccl...") LOG.i("installing nccl...")
arch_flag = f" -arch={','.join(map(lambda x:'sm_'+str(x),flags.cuda_archs))} " arch_flag = ""
if len(flags.cuda_archs):
arch_flag = f" -arch=compute_{min(flags.cuda_archs)} "
arch_flag += ''.join(map(lambda x:f' -code=sm_{x} ', flags.cuda_archs))
run_cmd(f"make -j8 src.build CUDA_HOME='{cuda_home}' NVCC_GENCODE='{arch_flag}' ", cwd=dirname) run_cmd(f"make -j8 src.build CUDA_HOME='{cuda_home}' NVCC_GENCODE='{arch_flag}' ", cwd=dirname)
return dirname return dirname

4
python/jittor/compiler.py
View File

@ -982,7 +982,9 @@ with jit_utils.import_scope(import_flags):
flags = core.flags() flags = core.flags()
if has_cuda: if has_cuda:
nvcc_flags += f" -arch={','.join(map(lambda x:'sm_'+str(x),flags.cuda_archs))} " if len(flags.cuda_archs):
nvcc_flags += f" -arch=compute_{min(flags.cuda_archs)} "
nvcc_flags += ''.join(map(lambda x:f' -code=sm_{x} ', flags.cuda_archs))
flags.cc_path = cc_path flags.cc_path = cc_path
flags.cc_type = cc_type flags.cc_type = cc_type