leezu commented on a change in pull request #17808: [WIP] Windows dev
environment configuration, update install instructions from source in the docs
URL: https://github.com/apache/incubator-mxnet/pull/17808#discussion_r401038549
##########
File path: ci/build_windows.py
##########
@@ -33,16 +33,43 @@
import zipfile
from distutils.dir_util import copy_tree
from enum import Enum
-from subprocess import check_call
+from subprocess import check_call, call
+import urllib.request
from util import *
+
+# Fix for broken PATH with newline inserted presumably by VS studio
installation of SQL server or
+# other component which makes visual studio stop working.
+os.environ['PATH']=os.environ.get('PATH').replace('\n','')
+
KNOWN_VCVARS = {
+ # https://gitlab.kitware.com/cmake/cmake/issues/18920
'VS 2015': r'C:\Program Files (x86)\Microsoft Visual Studio
14.0\VC\bin\x86_amd64\vcvarsx86_amd64.bat',
- 'VS 2017': r'C:\Program Files (x86)\Microsoft Visual
Studio\2017\Community\VC\Auxiliary\Build\vcvarsx86_amd64.bat'
+ 'VS 2017': r'C:\Program Files (x86)\Microsoft Visual
Studio\2017\Community\VC\Auxiliary\Build\vcvars64.bat',
+ 'VS 2019': r'C:\Program Files (x86)\Microsoft Visual
Studio\2019\Community\VC\Auxiliary\Build\vcvars64.bat'
}
+def get_cuda_arch(default='"5.2"'):
+ # cuda arch auto detection in cmake doesn't work on windows ci
Review comment:
The autodetection works fine on Windows. But we are compiling on a CPU
instance, so of course you can't detect the gpu arch automatically. Can you
remove this function and specify the MXNET_CUDA_ARCH argument as is done
before? This function adds unnecessary complexity.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
With regards,
Apache Git Services
