Open
Description
Describe the issue
Hi. Has anyone try compiling Intel Extension for pytorch locally? https://github.com/intel/intel-extension-for-pytorch. I've followed instructions, setup Visual Studio 2022 and OneAPI. The main issue I have is that the linking process doesn't find the basic Windows libraries (they are on my system) and I do start the build batch with the 'lib' environment variable having the location of those. For example the kernel32.lib etc. It seems that the information of the 'lib' env gets lost along the cmake process of the build. Documentation states that the Intel compiler/linker does not need specifc -L directives when 'lib' contains them.