Currently std::complex and some other std functions require uses to
include hip_runtime.h before any other headers to work, which is not
reliable.
changes are made in clang to fix this issue:
https://reviews.llvm.org/D81176
which requires hipcc and HIP headers to make corresponding changes.
This patch will make sure the clang change will not break
HIP/ROCclr during this transition.
After the transition is done, we can remove explicitly setting
include path for HIP-Clang and HIP header in hipcc and hip config
cmake files and rely on clang driver to set it automatically.
Change-Id: I5d226861c2560ffa6c5ab17343a43cc378048061
[ROCm/clr commit: 0a513d8a02]
1.Added hipModuleLaunchKernel multithreaded multi GPU scenario.
2.removed hipCtxCreate API from earlier test as it is deprecated.
SWDEV-238517 for enhancing hip unit tests
Change-Id: Id102d80887b6ff61a59938dbeb9fa2a26a3275b2
[ROCm/clr commit: 0920bac577]
this change follows CUDA convention where lower number is greater priority
Change-Id: I72596a36449e818cbd8c175bf8519c51f46b1610
[ROCm/clr commit: 83b8e1fbac]
Workaround hipStream deadlock issue as the same lock was used twice SWDEV-236746
Change-Id: Icc60104ce6edf4cfd2a3a889bab78a6caadd50b7
[ROCm/clr commit: f3ee29cdb2]
Support gfx908 as part of the default AMDGPU_TARGETS. MIGraphX requires this change.
Change-Id: I692f87f27829778e04f59c9ca655c6e8cbc00abc
[ROCm/clr commit: 29c7c9b1c2]
Similar to HCC, link with compiler-rt to support __fp16 and _Float16 type conversions in ONNX models. This should resolve SWDEV-238491.
Change-Id: Iad8dcff568831719f501f562a04023326ae8036c
[ROCm/clr commit: d93134e727]
The hipOccupancyMaxPotentialBlockSize API is meant to return the
number of threads for the highest-occupancy workgroup, and the number
of those workgroups. It was previously calculating the number of
maximum-sized workgroups that would fit on a single CU. This is
a mixture of the API we wanted (to calculate max potential block size)
and the MaxBlocksPerMultiprocessor function.
This patch fixes it up so that the internal occupancy calculation
function works for two uses: the traditional function that calculates
the maximum blocks per multiprocessor when a user passes in a fixed
block size (used for hipMaxBlocksPerMultiprocessor style functions)
and a function that calculates the size of a block that would lead
to maximum occupancy, and how many blocks of that size would be
needed to fill the whole GPU (for hipOccupancyMaxPotentialBlockSize
style functions).
This also updates the occupancy calculation function to prepare for
gfx10, which does not have SGPR-based occupancy limits.
Change-Id: Ie007b3f9d5ebc4e166b50a3a051498af35650f35
[ROCm/clr commit: 90453b68d3]
Git may not be available, and this may not be a git checkout, as would
happen in a release tarball. Doesn't really attempt to get a nicer
version formatting if some of the git subcommands fail.
Change-Id: Ib568cd1310983a43f2664ded72528d7e41f554c0
[ROCm/clr commit: eb31e45123]
SWDEV-237377 - This fixes time calculation where the event may
be recorded on Null stream and work submitted on other streams
Change-Id: Ie36310dea5cee2fed4a514ed01f04db4b47e571c
[ROCm/clr commit: facb05495f]
1. Updated FAQ with shft*sync not supported hip_faq.md
2. Corrected some of input parameter description in hcc_details/hip_runtime_api.h
3. Redirect shfl*() to shfl_*_sync() for nvcc path where CUDA > 9.0
Change-Id: I3d8184db5fcc622852c9bad96b706348e8dfc16c
[ROCm/clr commit: e288338e4a]
Rahul Garg
Vlad Sytchenko
kjayapra-amd
Siu Chi Chan
jujiang
Dittakavi Satyanvesh
Christophe Paquot
Yaxun (Sam) Liu
Saleel Kudchadker
Jason Tang
rohit pathania
Lakhan Singh Thakur
Aryan Salmanpour
Payam
Aaron En Ye Shi
Evgeny
Jatin
agodavar
Joseph Greathouse
Matt Arsenault
Michael LIAO
Sarbojit Sarkar
Ramesh Errabolu