- Symlink creation is corrected only for deb packages
- It is follow up package of http://git.amd.com:8080/c/hsa/ec/hsa-runtime/+/334403
- configure_file() is called to update the scripts with proper cmake variable values
Signed-off-by: Pruthvi Madugundu <pruthvi.madugundu@amd.com>
Change-Id: I0e833ead265166411e83593fd57265a9ab356904
CPack now incorrectly adds two copies of directory symlinks when
building Debian packages. This causes dpkg to see a file conflict
and fail installing.
The correct long term solution is to remove the symlink and use a
flat directory structure. This patch adds the symlink in the post
install script as a workaround until we can switch to flat layout.
Change-Id: I879b6cbc2661c19df3db639cb42fba0972fddb93
HDP will now be used for coarse grain kernarg so needs to be
reported without consideration of fine grain vram over pcie.
Change-Id: I648167299faa583876a3d8685c3b3c4d8d31ebf9
Setting to 1 prevents the scratch handler from reducing peak occupancy.
Scratch allocations that would normally reduce peak occupancy will
instead fail.
Diagnostic for TF and PyTorch.
Change-Id: I2d7ea47077eb5cf708251c8aa3fd183ad4261be0
scratch_used_large_ was uninitialized leading to the observed hang.
DynamicScratchHandler would wait for a large scratch release despite no
large scratch having yet been allocated. Fixes .
The patch also removes a potential race between AddScratchNotifier and
ReleaseQueueScratch. The race condition does not exist today since both
scratch alloc and release run on the same thread. The changes will
prevent this potential race from manifesting if the async event handler
is ever updated to use multiple threads.
Also enhances scratch occupancy reduction reporting. Reporting now
prints the initial request size as well as the allocated size and the
effect on occupancy this has. Occupancy is computed in terms of the
requesting dispatch grid size so may be >100%.
Change-Id: I0fc5ee01467ff4c29bdd25d545177c97862c3bd9
Ensures that all CPU agents will have a pool handle to allocate
system memory. These pools will have no numa binding since the
node their owning Agent represents has no installed memory.
Change-Id: I9f72b455d633646839753c6719ff7f6a4c41f7c4
- This new path is required when libhsaruntime.so is referred
from the top level ROCm lib directory.
- Once ROCm stack lib/lib64 structure is flatten, RUNPATH in all
the libraries needs to be updated.
Change-Id: I369131ce93e14958ec57a54701671f2bfd8d522a
Signed-off-by: Pruthvi Madugundu <pruthvi.madugundu@amd.com>
Attribute optimize(0) doesn't appear to be helpful helpful. This
prevents optimization in the function but not at call sites to the
function. The function may still be inlined since it has no side
effect (in some cases that we currently don't support).
Having a side effect prevents a call site optimization that allows
removal of a noinline function call with no side effect. Call site
optimization should only happen (in GCC at least) when using whole
program optimization so this may be stronger than we strictly need.
Also added _amdgpu_r_debug to the exported symbol list (global) and
switched to the standard macro for an exported symbol (HSA_API).
Without being in the global list the debugger will not find this
symbol if the binary has been stripped.
Change-Id: Ieb00175ccc55fda4491deee44711cd55b3f24aeb
Adding patch number based on ROCM build/release to have unique
file name for libraries across multiple versions of ROCM.
Signed-off-by: Pruthvi Madugundu <pruthvi.madugundu@amd.com>
Change-Id: I58d665b0e7d577b5bd7a6000d1202a0242672727
Lack of cache controls only allow operating SDMA at
agent scope. All copy APIs are defined at system scope so may
result in data errors.
Change-Id: I9cd10007defddcbf8feb14a2e3daa1ba17c0489f
Queues should transition to ref counting for all queues eventually.
That cleanup will be part of shared queue pooling support.
Change-Id: I217ff5d573156678b9559da6fb81baa8cd31c617
Temporary workaround for 2.10 release. RCCL, compiler, or firmware
must be corrected and this code reverted before another ASIC release.
Change-Id: I27851353289b93df9acb72d28b8c6ccb9f7f7d7a
Debugger path is taken for (trap_id >= 3) and single step exceptions.
Other traps/exceptions behave as before.
Change-Id: I276c0eb69953709968353a57717ee017d22348a2
Strip should only apply to the output target library. Symlinks
with .so endings which will be relocated during install will cause
strip to fail, aborting the build.
Change-Id: Ieb598c2cec5277d9d14c8afa88b91ca2c7f4412d
Using branch point for count since last change since we don't
have questions answered on tags yet.
Removed unused CMake files.
Restructured CMake to use the cache rather than only commandline
and be ccmake & cmake-gui friendly.
Dependency search paths are added for the Repo tree layout.
Search paths still needed for install paths.
Simplified packaging. hsa-ext-rocr-dev package and contents now
build from the package CMake rather than being 3 separate projects.
Not applying new version number or new install paths!
Change-Id: Ibea50dc8a6ab091e91857f78833f5379a4511547
- Use new buffer resource descriptor layout
- Handle wave32 scratch allocation error from CP
- Make wavefront size a property of scratch allocation requests
- Repurpose wave64-specific amd_queue_t.scratch_workitem_byte_size field
- Clear index_stride field in V# on gfx10, calculated per-dispatch by CP
Change-Id: If2acdf6430772abd4d6a8c792fc8c11260764dda
doorbell_queue_map should always be allocated or we will need to
add branches around all accesses.
Change-Id: I994c0eaf4be62c1a4a37bd06894272dba1fc1da6
sdma end ts must be 256 bit aligned in oss 3.0 and prior. Using
the ts pool requires copying into the signal and is a significant
performance penalty for small copies.
SharedSignal is 128 bytes due to alignment so can host the end ts.
Move sdma end ts into SharedSignal and remove ts pool and ts copy.
Change-Id: I7899bda36ebc9adcaad1d3a3d2b7a489857cc9e8
Impacts GPU_ONLY signal type latency when waiting for small operations.
Using this type improves total SDMA small copy performance by ~40% if
the signal is allowed to spin freely.
Change-Id: I27aa128c63a1bacb3f51fb08f166e4e1d6fef651
Remove agent lookup in time stamp translation for IPC signals. The copy
agent handle is not shared so does not need to be checked for cross
process use. Cross process copy-timestamp read is illegal and continues
to deliver garbage.
Store the copy agent properly when doing CPU-CPU copies.
Change-Id: Ib4008f66ff866922047749dd556c84a32021c1fd
ucode versions are per asic so not valid for feature enablement outside
of bringup/dev. Feature is older than the latest ioctl change that
the thunk depends on so use of this patch with kernel packages that
don't contain the feature is not possible in a supported environment.
Change-Id: I36b14176a7d642017ef1518aeade454b0f3dc749
Ramesh Errabolu
Pruthvi Madugundu
Sean Keely
Saleel Kudchadker
Freddy Paul
Laurent Morichetti
Srinivasan Subramanian
Qingchuan Shi
Chris Freehill
Jay Cornwall
Christian Sigg