Issue: mismatch undefined symbols in different user env
- Binary expects modified return value std::string&
- Fails to match libhip_hcc.so: return value is std::string& but doesn't match modified C++ env
Fix: Change return value to char*, create new key std::string in header from char*
+ Set -D__LP64__ in case of 64-bit hipify-clang binary
[partial workaround for clang's bug https://bugs.llvm.org/show_bug.cgi?id=38811]
C:/GIT/LLVM/trunk/llvm-64-release-vs2017/dist/lib/clang/9.0.0\include\__clang_cuda_device_functions.h(1609,45): error GEF7559A7: no matching function for call to 'roundf'
__DEVICE__ long lroundf(float __a) { return roundf(__a); }
#if defined(__LP64__)
__DEVICE__ long lround(double __a) { return llround(__a); }
__DEVICE__ long lroundf(float __a) { return llroundf(__a); } // ok: llroundf should be used when 64-bit
#else
__DEVICE__ long lround(double __a) { return round(__a); }
__DEVICE__ long lroundf(float __a) { return roundf(__a); } // error
#endif
+ Print more system info while testing in the following form:
========================================
CUDA 9.0 - will be used for testing
LLVM 9.0.0svn - will be used for testing
AMD64 - Platform architecture
Windows 10 - Platform OS
64 - hipify-clang binary bitness
32 - python 3.7.2 binary bitness
========================================
* Initial attempt to switch over to internally linked state.
* Add missing CMake update.
* hipLaunchKernelGGLImpl must be inline as well. Ensure internal linkage.
* Ensure global retrieval uses internally linked state.
* Hide HC in the implementation. Minimise ADL woes.
* Strange software exists, and must be catered to.
* Use a less spammy mechanism for ensuring internal linkage / non-export.
* Remove leftover internal detail.
HIP_PLATFORM detection logic relied on finding a working KFD. If it was
found, the platform was set as hcc else as nvcc.
However this logic is flawed since it is possible for the development
system to only have the user mode bits to build HIP application code.
Hence the better logic is to rely on finding a suitable compiler.
The new logic is as follows:
- look for a working HCC. If found, platform is set as hcc.
- else look for a working NVCC. If found, platform is set as nvcc.
- else the platform defaults to hcc for now.
Change-Id: Ifcc42c29a19f722153d5c23c55f1a8765dceaf6b
Aaron Enye Shi
Maneesh Gupta
Yaxun Sam Liu
Evgeny Mankov
Steven Noonan
eshcherb
Alex Voicu
Rahul Garg
Wen-Heng (Jack) Chung
Wilkin Chau