* Support pipelining codegen and template specialization
* Support ReduceCopy pipelining for AllReduce, ReduceScatter, and Reduce (currently enabled for bfloat16)
* Remove need for FUNC_INDEX_TOTAL
* Add pipeline field to device function key construction logic
* Avoid unneeded codegen for LL/LL64 kernels
* Modify conditions and add pipeline dtypes env
* Optimize selection for both gfx942 and gfx950
* Increase pipeline bitfield width
* Use __forceinline__ for all device functions
* Realign reduceCopy with original form
* Add opt-out option to enable perf debugs
* Remove force-reduce-pipelining option from README
* Update CHANGELOG.md
---------
Co-authored-by: Jeffrey Novotny <jnovotny@amd.com>
* add working-directory to ver_check step in rocprofiler-compute-packaging.yml
* Remove compute mi-rhel9 workflow badge since workflow is no longer in develop
* Update actions to v5 in rocprofiler-compute-docs
* Add working directory to steps in rocprofiler-compute-docs.yml
* Revert back to v4 pages
* Remove rocprofiler-compute-docs.yml workflow
* Remove docs workflow badge from rocprofiler-compute in README.md
* Remove rocprofiler-compute-packaging.yml, update README.md badges
Analysis data dump
* Add `--output-format` and `--output-name` option to analyze mode
* Remove `--output` and `-save-dfs` option to analyze mode
* Add documentation on `rocpd` output format and analysis database file
* Create sqlite3 database using object relation mapping (ORM) provided
by sqlalchemy library
* Fix metrics config to remove metrics marked as `null`, fix `Unit` header, add
missing `title`
* Add test cases to ensure analysis data dump work
* Move amd-smi to use caching mechanism
* Add VCN and JPEG activity to rocpd
* Switch cpu_freq to use caching mechanism
* Different approach with xcp activity & applied suggestions from code review
* Applied suggestions from code review
* Fix shadowing
* Applied suggestions from code review
* Added gpu-board and base-board temperatures to amd-smi metric
* Updated Changelog and adjusted the metric base-board/gpu-board output
* Adjusted output of metric to hide base/gpu-board when not relevant
---------
Signed-off-by: gabrpham_amdeng <Gabriel.Pham@amd.com>
* Added gpu-board and base-board temperatures to amd-smi metric
* Updated Changelog and adjusted the metric base-board/gpu-board output
* Adjusted output of metric to hide base/gpu-board when not relevant
---------
Signed-off-by: gabrpham_amdeng <Gabriel.Pham@amd.com>
[ROCm/amdsmi commit: b13fc16d60]
* Revert "SWDEV-547589 - Add hipDeviceMallocUncached to hipMemCreate (#815)"
This reverts commit 5ce7103555.
* Revert "SWDEV-547589 - comment for flag hipDeviceMallocUncached in hipMemcreate (#339)"
This reverts commit 04dac5eae3.
* SWDEV-551942 - implement hipMemAllocationTypeUncached in hipMemCreate
* Support gfx950 in topo_expl
* Fix dependencies and fetch fmt from sources
* Remove third_party folder in make clean
* Add empty target when fmt is found
* Add MI350 example
* Update README.md
---------
Co-authored-by: isaki001 <ioannissakiotis@gmail.com>
When stochastic sampling is not active, the trap handler is incorrectly
branching to .check_exceptions, bypassing the software trap ID checks
and inturn not advancing the PC. Fixed the issue to always check software
traps regardless of PC sampling state.
Co-authored-by: Shweta Khatri <shweta.khatri@amd.com>
* Using semaphore to sync with all peer processes in finalization stage
[rocprofv3] Implement synchronization using POSIX semaphore in finalization
* clang format code
* clang 11 format code
* Add process sync option for rocprofv3
* Default value of process sync is false
* Update source/lib/rocprofiler-sdk-tool/tool.cpp
Apply suggestion by Copilot
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* update according to comments
* add new line to helper.hpp
---------
Co-authored-by: Huanran Wang <huanrwan@amd.com>
Co-authored-by: Huanran Wang <huanran.wang@amd.com>
Co-authored-by: Madsen, Jonathan <Jonathan.Madsen@amd.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* add direct allgather algorithm
* minor fix
* add debug print for memory allocation tracker
* add message size threshold for direct allgather
* scatter transfers across ranks
* update changelog
* minor fix
* Update CHANGELOG.md
Co-authored-by: Jeffrey Novotny <jnovotny@amd.com>
* enable direct AG when pxn is ON on MI300X or MI350
---------
Co-authored-by: Jeffrey Novotny <jnovotny@amd.com>
This commit reverts the following related commits which cause
test failures:
6d15779b3e rocr/driver: add PC sampling support to driver interface
56cb9390ff rocr/driver: add PC sampling support to driver interface
76bf829f09 rocr/driver: add ASAN header page management to Driver class
a47c060d6a rocr/driver: add ASAN header page management to Driver class
02d7eaf3b7 rocr: add memory sharing call to Driver interface
9312468655 rocr: add memory sharing call to Driver interface
Assigning a null terminator at
the end of the string wrote
past the end of the allocated
buffer. This patch corrects that.
Signed-off-by: Ashutosh Mishra <ashutosh.mishra@amd.com>
Changed `TestBedChild` protocol to send the result code before the return value to avoid hanging if the call fails. Switched `TestBedChild::GetUniqueId` to use this.
David Galiffi
Julia Jiang
Kian Cossettini
systems-assistant[bot]
Milan Radosavljevic
Mustafa Abduljabbar
Jason Bonnell
vedithal-amd
Satyanvesh Dittakavi
Pham, Gabriel
Nusrat Islam
Jimbo
Jeffrey Novotny
MachineTom
xuchen-amd
shwetakhatri-amd
SaleelK
cfreeamd
Ashutosh Mishra
ywang103-amd
corey-derochie-amd
Nilesh M Negi
Taylor Ding
adapryor
Maisam Arif
jamessiddeley-amd
Danylo Lytovchenko
David Yat Sin
Daniel Su
Oosman Saeed