76333 Коммитов

Автор SHA1 Сообщение Дата
Dmitrii 8cb6278f32 CMAKE - Use gersemi instead of cmake-format (#772)
cmake-format has been unmaintained for 4+ years now. Please use gersemi
instead.

See RDC changes where we adopted gersemi:
- <https://github.com/ROCm/rdc/commit/680b7a8dd8699ab5264f10f5a3bc0b30410881be>
- <https://github.com/ROCm/rdc/commit/40545dcb49ebcfa0a6fe040a8862cf8b338ad6b5>

Signed-off-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>
2025-09-10 18:09:02 -05:00
Ioannis Assiouras 35629e433d SWDEV-546146 - Added support for hipMemLocationTypeHost in hipMemSetAccess (#682) 2025-09-10 23:06:20 +01:00
Joseph Macaranas dd1a2dbf8a Fix LICENSE path for opencl build (#939) 2025-09-10 17:54:22 -04:00
amd-hsivasun cf8173a892 [Ex CI] Enable rocm-smi-lib (#937) 2025-09-10 17:30:57 -04:00
systems-assistant[bot] b60c0ceddd [rocprofv3] Unconditionally collect stream and kernel rename data in rocprofv3 for rocpd (#171)
* Remove config checks for stream and kernel rename data collection

* Updated csv generation to check if kernel rename is on before calling get_kernel_name

* Update metadata to use kernel_rename bool argument

* Formatting + unconditionally store kernel name in rocpd

* Readded kernel rename parameter after rebase

* Fixed rebase conflicts

* Updated comment in line with github comments

* Added check in rocpd csv.cpp to output kernel name if region name is empty

* Add test for kernel rename

---------

Co-authored-by: Ian Trowbridge <Ian.Trowbridge@amd.com>
2025-09-10 16:03:15 -05:00
Pryor, Adam 5ebd7b8022 Fix groups failing inside container (#684)
* Fix groups failing inside container

---------

Signed-off-by: adapryor <Adam.pryor@amd.com>
2025-09-10 15:36:26 -05:00
Pryor, Adam 0a2231deb7 Fix groups failing inside container (#684)
* Fix groups failing inside container

---------

Signed-off-by: adapryor <Adam.pryor@amd.com>

[ROCm/amdsmi commit: 5ebd7b8022]
2025-09-10 15:36:26 -05:00
amd-hsivasun 156de36a92 [Ex CI] Enable rocminfo (#932) 2025-09-10 16:30:57 -04:00
systems-assistant[bot] bfdb3bc636 fix(python): fix comparison to None (#211)
from PEP8 (https://peps.python.org/pep-0008/#programming-recommendations):

> Comparisons to singletons like None should always be done with is or is not, never the equality operators.

Co-authored-by: Eisuke Kawashima <e-kwsm@users.noreply.github.com>
2025-09-10 14:50:32 -05:00
systems-assistant[bot] 39ea16e544 fix(E712): fix comparison to True/False (#212)
from PEP8 (https://peps.python.org/pep-0008/#programming-recommendations):

> Comparisons to singletons like None should always be done with is or is not, never the equality operators.

Co-authored-by: Eisuke Kawashima <e-kwsm@users.noreply.github.com>
2025-09-10 14:50:23 -05:00
systems-assistant[bot] f450ff0624 Update rocm_smi_device.cc of delete duplicate content (#214)
It has the duplicate content in line 488 in the same file.

Co-authored-by: qiwei_ji <qiwei_ji@outlook.com>
2025-09-10 14:50:13 -05:00
Bindhiya Kanangot Balakrishnan c7b6bb9600 SWDEV-476667 - Get pcie bw from GPU metric (#853)
The sysfs pcie bandwidth file pcie_bw is deprecated
in newer asics. This change will get pcie BW from
GPU metric for version 1.5 or later.

Signed-off-by: Bindhiya Kanangot Balakrishnan <Bindhiya.KanangotBalakrishnan@amd.com>
2025-09-10 14:48:31 -05:00
Ammar ELWazir 2a9700fcd7 [ROCProfiler-Register/Systems/Compute] Fixing License file name usage (#927)
ROCProfiler-Register/Systems/Compute: The license file name in the CMake install module and other locations was originally LICENSE, but it was recently changed to LICENSE.md, requiring an update to the CMake install module and all other relevant locations.
2025-09-10 15:46:39 -04:00
isaki001 9fa7a738da add reduce/broadcast algo/proto selection table for multi-node gfx940 (#1889)
[ROCm/rccl commit: 9c36439354]
2025-09-10 14:25:23 -05:00
isaki001 9c36439354 add reduce/broadcast algo/proto selection table for multi-node gfx940 (#1889) 2025-09-10 14:25:23 -05:00
Mythreya Kuricheti aa90a83e42 [aqlprofile] Misc formatting (#810)
* Remove dupicated code segment and apply formatting to some code regions
2025-09-10 12:11:53 -07:00
Omri Mor 8b4689f129 Update AUTHORS (#246)
* Update AUTHORS and CODEOWNERS

[ROCm/rocshmem commit: f677e5eb59]
2025-09-10 11:32:23 -07:00
Omri Mor f677e5eb59 Update AUTHORS (#246)
* Update AUTHORS and CODEOWNERS
2025-09-10 11:32:23 -07:00
Omri Mor 07fc061008 add missing copyright header to rocshmem_info.cpp (#245)
[ROCm/rocshmem commit: 0fd966611d]
2025-09-10 10:11:06 -07:00
Omri Mor 0fd966611d add missing copyright header to rocshmem_info.cpp (#245) 2025-09-10 10:11:06 -07:00
nunnikri b6ffa43226 SWDEV-553044 : Adding missing copyright header in rocm-core (#875)
SWDEV-553044 : Adding missing copyright header in rocm-core
2025-09-10 10:10:16 -07:00
Avinash Kethineedi 6860bc1275 GDA get* APIs (#243)
* feat(GDA): add `get*` and `get*_nbi` APIs for mlx and bnxt NICs
   - implemented thread, wave and wg variants of `get*` and `get_nbi`.

* test(GDA): enable functional tests for `get*` and `get*_nbi` APIs

[ROCm/rocshmem commit: 671f8187f4]
2025-09-10 11:24:53 -05:00
Avinash Kethineedi 671f8187f4 GDA get* APIs (#243)
* feat(GDA): add `get*` and `get*_nbi` APIs for mlx and bnxt NICs
   - implemented thread, wave and wg variants of `get*` and `get_nbi`.

* test(GDA): enable functional tests for `get*` and `get*_nbi` APIs
2025-09-10 11:24:53 -05:00
Julia Jiang 8bc97e3273 SWDEV-551652 - Adding changelog for HIP 7.0.2 (#849) 2025-09-10 09:22:40 -07:00
Julia Jiang ae9a7f34c0 SWDEV-504376 - Update definition for hipHostAllocWriteCombined (#887) 2025-09-10 09:22:24 -07:00
Julia Jiang 4981fa2139 SWDEV-552580 - Update instructions to build HIP from source in monorepo (#806) 2025-09-10 09:21:27 -07:00
Joseph Macaranas 696881ae82 LICENSE clean up (#919)
- Clean up and standardization of MIT licenses after discussion with legal team.
- Update README.md with blurb for top-level files.
- MIT License explicitly mentioned for relevant projects.
- Removal of years.
- Copyright attribution should be to `Advanced Micro Devices, Inc.` and not `AMD ROCm(TM) Software`
- Removal of `All rights reserved.`
- Reduce line width of the text for readability.
- Add clear visual separators for additional licenses.
- Convert text files to markdown format for aforementioned separators.
- Update build scripts to point to renamed files.
- Fixed SMI doc references

Co-authored-by: Maisam Arif <Maisam.Arif@amd.com>
2025-09-10 12:06:14 -04:00
systems-assistant[bot] f34abcad37 SWDEV-548453 - Fix memory leak (#528) 2025-09-10 21:17:30 +05:30
amd-hsivasun bd12f54c72 [Ex CI] enable rocm-core (#916)
* Fix syntax error in azure_resolve_subtree_deps.py

* Fixed typo

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
2025-09-10 11:47:00 -04:00
Wenkai Du 2577b33de8 Enable LL128 and use same tuning table for gfx942 4 NICs (#1898)
[ROCm/rccl commit: c2bccf9156]
2025-09-10 11:11:15 -04:00
Wenkai Du c2bccf9156 Enable LL128 and use same tuning table for gfx942 4 NICs (#1898) 2025-09-10 11:11:15 -04:00
ywang103-amd 4c614d3d21 remove test_scientific_notation_trigger_above_upper_bound for its (#913) 2025-09-10 10:18:42 -04:00
yugang-amd 526784233b remove broken link etc. (#234)
[ROCm/rocshmem commit: 4a760d741a]
2025-09-10 09:48:28 -04:00
yugang-amd 4a760d741a remove broken link etc. (#234) 2025-09-10 09:48:28 -04:00
Yiltan b79b9f4b60 Unify common BNXT and MLX5 initialization code (#233)
Co-authored-by: Aurelien Bouteiller <aurelien.bouteiller@amd.com>

[ROCm/rocshmem commit: cb39f7a313]
2025-09-10 09:13:36 -04:00
Yiltan cb39f7a313 Unify common BNXT and MLX5 initialization code (#233)
Co-authored-by: Aurelien Bouteiller <aurelien.bouteiller@amd.com>
2025-09-10 09:13:36 -04:00
Giovanni Lenzi Baraldi 9849073836 SWDEV-540648: Adding realtime clock to v3 tool. Update decoder header. (#666)
* SWDEV-540648: Adding realtime clock to v3 tool. Update header for decoder.

* Adding tests

* Review comments

* Review comment
2025-09-10 12:39:27 +02:00
systems-assistant[bot] 4e9acd492d SWDEV-487647 - Add HIP tests for Dynamic Alloca (#569)
Co-authored-by: Pravin Jagtap <Pravin.Jagtap@amd.com>
Co-authored-by: Aaditya <Aaditya.AlokDeshpande@amd.com>
Co-authored-by: Arsenault, Matthew <Matthew.Arsenault@amd.com>
2025-09-10 13:25:41 +05:30
Godavarthy Surya, Anusha 1be5c9870a SWDEV-524745 - Part-I Add multi device support for hip graph. Update nodes with DevId. (#812)
- The graph nodes have been updated to capture the device ID from the capture stream or the current device when explicitly added.
- Update the device ID for the memcpy node, ensuring that the device where the memory is allocated is taken into account for H2D and D2H pinned operations.

Co-authored-by: Anusha GodavarthySurya <Anusha.GodavarthySurya@amd.com>
2025-09-10 11:35:25 +05:30
systems-assistant[bot] 75602772aa SWDEV-538606 - Handle updateStreams from multiple threads (#505) 2025-09-10 11:24:52 +05:30
Venkateshwar Reddy Kandula 61583e3125 [AQLProfile] Fix caching of agents with new available data (#752)
* > hsa_agent not provided by new api/rocprofiler-sdk and causes every device to have same id,
in cases where gfxip is same and config is different pm4factory doesn't know the difference. This fix uses gfxip and CU count as a key for cache.

* Change comparison from gfxip to name in instances_fncomp_t

Updated comparison in instances_fncomp_t to use 'name' for backward compatibility with rocprofv2.

---------

Co-authored-by: Venkateshwar Reddy Kandula <venkateshwar.kandula1306@gmail.com>
2025-09-09 22:45:05 -05:00
Venkateshwar Reddy Kandula 1494243e4d Update CODEOWNERS to remove specific project owners (#918)
Removed ownership entries for aqlprofile and rocprofiler from CODEOWNERS.
2025-09-09 16:54:40 -05:00
vedithal-amd 85a557673d Handle mutually exclusive report section filters (#710)
* Make --roof-only, --block and --set mutually exclusive from each other

* Update help output and documentation
  * Add sanitize function for checking profiler options

* Update filter blocks arguments when --set or --roof-only is provided

* Update filter_blocks in profiling_config.yaml based on --set option
  * Log Filtered Sections instead of Report Sections and Set Selection

* Move soc class function calls from rocprof compute base class to profiler base class

* Fix bug in panel level filtering using --filter-block option

* Remove roofline specific pmc files

* Move microbenchmark entry point from gfx specific soc class to base soc class

* Run microbenchmarks only if block 4 is selected or roof only is selected; skip for mi100
2025-09-09 17:48:20 -04:00
Joseph Macaranas 5738922604 CODEOWNERS: remove rocprofiler-sdk entries (#917)
- rocprofiler-sdk team self-manages reviewer selection
2025-09-09 17:31:52 -04:00
Pham, Gabriel 954d4860c1 [SWDEV-551309] Adjusted amdsmitst and reset command (#654)
* Adjusted amdsmitst and reset command to account for separation of power profile and perf level behavior
* Updated test to reset power profile to previous user setting
* Removed performance level from reset_profile_results in reset --profile command
* Updated Changelog with change to reset profile behavior

---------

Signed-off-by: gabrpham_amdeng <Gabriel.Pham@amd.com>
2025-09-09 16:11:07 -05:00
Pham, Gabriel e9ee0bccf2 [SWDEV-551309] Adjusted amdsmitst and reset command (#654)
* Adjusted amdsmitst and reset command to account for separation of power profile and perf level behavior
* Updated test to reset power profile to previous user setting
* Removed performance level from reset_profile_results in reset --profile command
* Updated Changelog with change to reset profile behavior

---------

Signed-off-by: gabrpham_amdeng <Gabriel.Pham@amd.com>

[ROCm/amdsmi commit: 954d4860c1]
2025-09-09 16:11:07 -05:00
Benjamin Welton ed5b2ac165 Fix deadlock in InterceptQueue::Submit when packet count exceeds queue capacity (#855)
InterceptQueue::Submit had an "all-or-nothing" packet submission policy that
could cause infinite retry loops when the number of packets to submit exceeded
the available queue slots. When 504+ packets needed submission to a ~500-slot
queue, the system would:
1. Set submitted_count=0 (submit nothing)
2. Add retry barrier packet
3. Trigger async handler via StoreRelaxed
4. Attempt to submit overflow packets
5. Fail again due to same space constraints
6. Repeat

Solution:
Added partial packet submission capability during overflow processing while
preserving the original "all-or-nothing" behavior for normal operations.
When processing overflow packets and insufficient space exists for all packets,
the system now submits as many packets as possible rather than none.

The fix:
- Detects overflow processing via !overflow_.empty()
- Allows partial submission: submitted_count = free_slots - barrier_reservation
- Maintains atomicity guarantees for normal packet rewrites
- Prevents infinite retry loops by ensuring forward progress

This resolves deadlocks in high-throughput scenarios while maintaining
backward compatibility and the original design intent for packet rewrite
atomicity.
2025-09-09 14:06:29 -07:00
Arif, Maisam fd5eb4e963 [SWDEV-550075] Updated README to link to amd-smi virtualization repo (#664)
Co-authored-by: Peter Park <peter.park@amd.com>
2025-09-09 16:05:01 -05:00
Arif, Maisam 1a36f2ad0b [SWDEV-550075] Updated README to link to amd-smi virtualization repo (#664)
Co-authored-by: Peter Park <peter.park@amd.com>

[ROCm/amdsmi commit: fd5eb4e963]
2025-09-09 16:05:01 -05:00
SaleelK c8e91b3f3e clr: Fix condition for taking shader path (#884)
* SWDEV-551080
* Fix condition for taking shader path, the size check was moved
  incorrectly
* Also account for a bitmask returned for preferred engines
2025-09-09 13:13:29 -07:00