Since rcp implementations of non-default rounded versions are not correct or supported in OCML, guard them using the same macro OCML_BASIC_ROUNDED_OPERATIONS. Also update the docs and tests.
[ROCm/hip commit: 0cfaa52d15]
In ROCm-Device-Libs, they have dropped the non-default-rounded versions of add, sub, mul, div, sqrt and fma. Therefore, ocml has removed the rte, rtp, rtn, and rtz counterparts. This will remove the same math APIs in HIP for _ru, _rd, _rn, and _rz.
[ROCm/hip commit: 2b108a2cfd]
from a header into a source file such that
there's only an unique copy in an executable
and prevent wasting static memory on the host
Change-Id: Id5b62766f77809c8d7b47892cb7149c490dcbdb9
[ROCm/hip commit: 0ff408a56c]
+ Fix typo with missing comma in counterNames array
+ Change std::string argument to const std::string& argument in all functions
[ROCm/hip commit: eff5d3fc1b]
+ for all CUDA versions up to 10.0
+ add missing types
+ fix typos
+ sync with HIP
+ sync with CUDA Driver API
+ update CUDA_Runtime_API_functions_supported_by_HIP.md
+ update CUDA Driver API map accordingly
+ update CUDA_Driver_API_functions_supported_by_HIP.md
+ formatting, annotating
[ROCm/hip commit: c5bfa09b04]
Clang 7.0.0 release has a known bug https://bugs.llvm.org/show_bug.cgi?id=36384, thus 7.0.0 is unstable release for Linux and Windows.
Last stable and tested LLVM+Clang release is 6.0.1.
[ROCm/hip commit: 368e9d5100]
Move remaining function definitions from device_util.cpp to hip_runtime.h header. We can now remove the static library completely as device_util.cpp was the last file part of target hip::hip_device .
[ROCm/hip commit: 817b27d530]
[Reason]
Avoid cmake error occurred when mixing signatures with AddLLVM.cmake
[Error]
CMake Error at CMakeLists.txt:26 (target_link_libraries): The keyword signature for target_link_libraries has already been used with the target "hipify-clang".
All uses of target_link_libraries with a target must be either all-keyword or all-plain.
Error occurred only on Linux and LLVM 7.0.0
[ROCm/hip commit: 2c367bc8af]
[Reasons]
+ Simplify the support
+ Optimize compilation and code generation time
[Additionally]
+ Update minimum required version of cmake
+ Fix MSVC compile options force update
[ROCm/hip commit: 5d1255272b]
When compiling with Early Finalization enabled in HCC,
the resulting .kernel section of the host object now may
contain more than one device code bundles. This is to
teach the HIP runtime to correctly extract all the
bundles from the .kernel section.
[ROCm/hip commit: 30ce25e627]
+ for all CUDA versions
+ add missing types
+ fix typos
+ sync with HIP
+ update CUDA_Driver_API_functions_supported_by_HIP.md
+ formatting, annotating
[ROCm/hip commit: 520e795a8d]
Evgeny
Maneesh Gupta
Yaxun Sam Liu
Aaron Enye Shi
Siu Chi Chan
Evgeny Mankov
Jeff Daily
Rahul Garg
Evgeny Mankov
Anton Gorenko
Alex Voicu
Qianfeng Zhang