Commit grafiek

76347 Commits

Auteur SHA1 Bericht Datum
Allen Hubbe 1494c24f9a gda ionic: collapsed cqe (#345)
* util: dlsym optional helper

Like DLSYM_HELPER, but does not return if the symbol is not found.

Signed-off-by: Allen Hubbe <allen.hubbe@amd.com>

* gda ionic: sync dv and fw headers

Sync dv and fw headers to match out-of-tree libionic and firmware.

Signed-off-by: Allen Hubbe <allen.hubbe@amd.com>

* gda ionic: collapsed cqe

Detect and enable collapsed cqe if supported by drivers and firmware.
Fall back to regular completion queue.

Signed-off-by: Allen Hubbe <allen.hubbe@amd.com>

---------

Signed-off-by: Allen Hubbe <allen.hubbe@amd.com>
2026-01-06 20:42:15 -05:00
Aurelien Bouteiller bcdf60def6 Enable new a2a (pr 334) on ionic as well (#366)
* Enable new a2a (pr 334) on ionic as well

* Apply suggestions from AI code review

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

* Apply suggestions from code review

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

---------

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

[ROCm/rocshmem commit: 82d91433c9]
2026-01-06 20:41:51 -05:00
Aurelien Bouteiller 82d91433c9 Enable new a2a (pr 334) on ionic as well (#366)
* Enable new a2a (pr 334) on ionic as well

* Apply suggestions from AI code review

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

* Apply suggestions from code review

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

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
2026-01-06 20:41:51 -05:00
Stella Laurenzo 81eed26ec6 [amdsmi] Add include dirs for libdrm. (#2504)
This has started failing on various developer build systems. Looking at it, it is not precisely clear how this ever worked given that nothing appears to be adding the DRM include dirs.

I'd prefer that we remove this delay loading (at least for TheRock builds where it is never needed), but in the meantime, this does fix the issue and is verified on an affected system.

Fixes https://github.com/ROCm/TheRock/issues/2744
2026-01-06 15:18:20 -08:00
Yazen AL Musaffar cb372748f8 [ROCM-SMI] [SWDEV-569731] rsmi tests failing on Frequency/Power/GpuMetrics ReadOnly Fix (#2303)
* Updated unsupported metric version file for rocm_smi_tests Frequency/Power/GpuMetrics ReadOnly tests

Signed-off-by: yalmusaf_amdeng <Yazen.ALMusaffar@amd.com>
2026-01-06 16:46:38 -06:00
Gerardo Hernandez 50644f5aef SWDEV-508225 remove assertions when loading fat binary (#2013)
* SWDEV-508225 - do not assert() after calling digestFatBinary() if it fails. Otherwise this causes assertions to trigger easily in systems that have an APU and a discrete GPU and the code was compiled for the discrete one

* SWDEV-508225 - fix that when using a non-existent ordinal in HIP_VISIBLE_DEVICES, getCurrentArch() would crash
2026-01-06 21:53:32 +00:00
Daniel Oliveira 32fde0f73d [SWDEV-568613] Add gpu_metrics 1.0 support for older GPUs (#2444)
fix: Add gpu_metrics 1.0 support which is still used by some hardware

Code changes related to the following:
  * APIs
  * Unit tests

Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
2026-01-06 14:25:13 -06:00
systems-assistant[bot] c6b7448227 Add support for get and set APIs for CPUISOFreqPolicy and DFCState Co… (#1901)
* Add support for get and set APIs for CPUISOFreqPolicy and DFCState Control

  - Add support for get and set APIs for CPUISOFreqPolicy and DFCState Control
    in AMD SMI and also in the CLI tool

* CHANGELOG.md file updated

* SWDEV-562837: Update amdsmi-py-api.md as per the new APIs

Updated amdsmi-py-api.md as per the new APIs added.

---------

Signed-off-by: Soumya <sranjanr@amd.com>
Signed-off-by: gabrpham <Gabriel.Pham@amd.com>
Co-authored-by: Saka Sitharammurthy <SitharamMurthy.Saka@amd.com>
2026-01-06 10:37:07 -06:00
SakaSitharammurthy 6c98c49362 [SWDEV-568731] Updated example code in amdsmi-py-api.md file (#2311)
Addresses:
- SWDEV-568731
- SWDEV-568724
- SWDEV-568695

Signed-off-by: Saka, SitharamMurthy <SitharamMurthy.Saka@amd.com>
2026-01-06 10:34:36 -06:00
Edgar Gabriel e38f98fad5 fix reduction test for gfx1201 (#374)
* fix reduction for gfx942 and 1201

match the synchronizaation of internal_putmem_wg and internal_getmem_wg
to their non-internal counterparts. the internal_putmem_wg is used in
the ipc reduction

* move specialization to internal_putmem

[ROCm/rocshmem commit: 8d2504d6c1]
2026-01-06 10:15:38 -06:00
Edgar Gabriel 8d2504d6c1 fix reduction test for gfx1201 (#374)
* fix reduction for gfx942 and 1201

match the synchronizaation of internal_putmem_wg and internal_getmem_wg
to their non-internal counterparts. the internal_putmem_wg is used in
the ipc reduction

* move specialization to internal_putmem
2026-01-06 10:15:38 -06:00
pghoshamd 637b0d71f0 SWDEV-569319 Replace ScopedAcquire with stdcpp wrappers (#2146)
* SWDEV-569319 Replace ScopedAcquire with stdcpp wrappers

* Remove KernelMutex and KernelSharedMutex abstractions with std::mutex and std::shared_mutex

* Replaced unique_locks with lock_guards

* More changes

* Replace new and deletes with smart pointers

* Replaced some more with shared ptrs

* Replacements with smart pointers - pt 2

* missed change
2026-01-06 10:59:34 -05:00
Mustafa Abduljabbar 5bba932529 [WarpSpeed] Improve handling for auto and manual modes (#2125)
* Force ring in WarpSpeed manual mode and log event

* Skip usage for non-ring in WarpSpeed auto mode

* Enable WarpSpeed when its CU count is set

[ROCm/rccl commit: 93fdcb160c]
2026-01-06 10:21:49 -05:00
Mustafa Abduljabbar 93fdcb160c [WarpSpeed] Improve handling for auto and manual modes (#2125)
* Force ring in WarpSpeed manual mode and log event

* Skip usage for non-ring in WarpSpeed auto mode

* Enable WarpSpeed when its CU count is set
2026-01-06 10:21:49 -05:00
vedithal-amd e005f8487b [rocprofiler-compute] Add gfx arch. based pre-processor guards and runtime checks in rocflop.cpp (#2487)
* Remove MFMA functionality in rocflop sample since its not supported in MI50

* Add gfx arc based support for MFMA and SMFMAC in rocflop.cpp

* Add --int32 usage doc

* Address review comments
2026-01-06 10:17:54 -05:00
Nusrat Islam 49d9f8cc27 use memcpy for local copies (#2121)
Co-authored-by: Nusrat Islam <nusislam@dell300x-ccs-aus-k13-09.cs-aus.dcgpu>

[ROCm/rccl commit: b4a86ef680]
2026-01-06 09:00:57 -06:00
Nusrat Islam b4a86ef680 use memcpy for local copies (#2121)
Co-authored-by: Nusrat Islam <nusislam@dell300x-ccs-aus-k13-09.cs-aus.dcgpu>
2026-01-06 09:00:57 -06:00
Edgar Gabriel cc727261de disable the putmem_signal_on_stream on RO (#376)
it fails in about 50% of the cases. Will revisit later why it fails,
but RO is at the moment lower priority, so disabling the test for now.

[ROCm/rocshmem commit: ed2f75f1de]
2026-01-06 08:10:46 -06:00
Edgar Gabriel ed2f75f1de disable the putmem_signal_on_stream on RO (#376)
it fails in about 50% of the cases. Will revisit later why it fails,
but RO is at the moment lower priority, so disabling the test for now.
2026-01-06 08:10:46 -06:00
Jonathan R. Madsen 7fcea905f3 [rocprofiler-sdk] Fix double-buffering emplace and flush synchronization (#2334)
* Fix buffer tracing synchronization lock

- PR #529 (in rocprofiler-sdk-internal) introduced waiting on the syncer flag when emplacing in a buffer to prevent the overwriting buffer records currently being processed in a buffer flush callback
- The above fix introduced a block on the both buffers when a buffer flush callback was being executed instead of a block on the buffer being flushed.

* Add rocpd tests for duplicate records

* Address code review comments
2026-01-06 06:06:18 -06:00
habajpai-amd 9e4d1c31c7 fix: prevent static initialization deadlock in thread_data (#2474)
* fix: prevent static initialization deadlock in thread_data

* update comment
2026-01-06 16:39:32 +05:30
Longlong Yao c34ec1e52f wsl/librocdxg: Change scratch memory allocation
Calculate the actual scratch memory size required based on the
packet information for kernel dispatch.

If the required size exceeds the total allocated memory, scratch
memory must be reallocated. Otherwise, no action is needed.

miopen_gtest: Full/GPU_MIOpenDriverRegressionTest_FP16.MIOpenDriverRegressionHalf/0

Signed-off-by: Longlong Yao <Longlong.Yao@amd.com>
Reviewed-by: Flora Cui <flora.cui@amd.com>
Reviewed-by: Horatio Zhang <Hongkun.Zhang@amd.com>
2026-01-06 10:12:04 +08:00
Longlong Yao c3f55c8e59 wsl/librocdxg: Change scratch memory allocation
Calculate the actual scratch memory size required based on the
packet information for kernel dispatch.

If the required size exceeds the total allocated memory, scratch
memory must be reallocated. Otherwise, no action is needed.

miopen_gtest: Full/GPU_MIOpenDriverRegressionTest_FP16.MIOpenDriverRegressionHalf/0

Signed-off-by: Longlong Yao <Longlong.Yao@amd.com>
Reviewed-by: Flora Cui <flora.cui@amd.com>
Reviewed-by: Horatio Zhang <Hongkun.Zhang@amd.com>
2026-01-06 10:12:04 +08:00
Aurelien Bouteiller abb1e0684a Do not hardcode wf_size==64 in ionic provider (#367)
* Do not hardcode wf_size==64 in ionic provider

* Simpler same_qp_mask in ionic


[ROCm/rocshmem commit: 0c496d83d6]
2026-01-05 18:36:58 -05:00
Aurelien Bouteiller 0c496d83d6 Do not hardcode wf_size==64 in ionic provider (#367)
* Do not hardcode wf_size==64 in ionic provider

* Simpler same_qp_mask in ionic
2026-01-05 18:36:58 -05:00
Jeff Jiang bb47dd2537 Utils - Fix for the incorrect MD5 results when user crop is enabled. (#687)
* * rocDecode: Fixed the incorrect MD5 results when user crop is enabled.
 - When the user cropping is enabled, the cropping rect needs to be sent to the output surface info struct, which is used by the MD5 calculation.

* * rocDecode: Minor change based on review comment.

[ROCm/rocdecode commit: 8fbb195a72]
2026-01-05 13:11:59 -08:00
Jeff Jiang 8fbb195a72 Utils - Fix for the incorrect MD5 results when user crop is enabled. (#687)
* * rocDecode: Fixed the incorrect MD5 results when user crop is enabled.
 - When the user cropping is enabled, the cropping rect needs to be sent to the output surface info struct, which is used by the MD5 calculation.

* * rocDecode: Minor change based on review comment.
2026-01-05 13:11:59 -08:00
Jason Bonnell 1d5a6e9bfe Update rocprofiler workflows to use new mi325 runner names (#2467)
* Update rocprofiler workflows to use new runner naming for mi325

* Add input options to workflow_dispatch for rocprofiler-systems CI workflow

* Update runner name on therock-ci-linux.yml as well
2026-01-05 15:41:01 -05:00
AidanBeltonS 39d8432893 SWDEV-566854 - Improve memory object handling (#1939)
* Improve memory object handling for memcpy

* update

* Pass offsets and make hip_graph changes

* Update projects/clr/hipamd/src/hip_memory.cpp

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

* Remove unnecessary command overload

* Update based on feedback

* Fix failing hipGraphTests

* Fix graph bugs

* Fix failing memcpy tests

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
2026-01-05 18:05:56 +00:00
Joseph Macaranas 11d9472e5f Bump TheRock SHA for CI 20251230 (#2466)
* Bump TheRock SHA for CI 20251230
* Remove patch and align workflows between OS
2026-01-05 13:00:37 -05:00
Benjamin Welton 7871f53563 Add gfx950 support to ValuPipeIssueUtil counter (#2396)
Add gfx950 (MI350) to the ValuPipeIssueUtil counter definition to
enable RDC_FI_PROF_VALU_PIPE_ISSUE_UTIL telemetry field support on
MI350 hardware.
2026-01-05 09:37:34 -08:00
Julia Jiang 88f4bb1988 SWDEV-564412 - fix test failure on hipSetValidDevices_with_hipMemcpyPeer (#2150) 2026-01-05 12:36:31 -05:00
Julia Jiang 0f0504d79d SWDEV-564412-Fix soft hang in HIP sub-test hipMemVmm_Uncached (#2223) 2026-01-05 12:36:08 -05:00
Julia Jiang 3568e0df02 SWDEV-563487 - Fix catch tests failures on Windows (#2097) 2026-01-05 12:35:41 -05:00
Shadi Dashmiz 2789ea429a SWDEV-565300: Fix coherency range mode in mem pool pointers (#2296)
Signed-off-by: sdashmiz <shadi.dashmiz@amd.com>
2026-01-05 11:33:11 -05:00
Avinash de23e1db6d Navi4 LL enablement and tuning (#2095)
* LL enablement for gfx1201

* Single node LL/Simple tuning

* multinode algo/prto default choice

* First iteration of Table tuning

* gfx924 tuning table correction

* Addressing PR comments and prefix match fix


[ROCm/rccl commit: 9545ae04b2]
2026-01-05 10:17:12 -06:00
Avinash 9545ae04b2 Navi4 LL enablement and tuning (#2095)
* LL enablement for gfx1201

* Single node LL/Simple tuning

* multinode algo/prto default choice

* First iteration of Table tuning

* gfx924 tuning table correction

* Addressing PR comments and prefix match fix
2026-01-05 10:17:12 -06:00
Flora Cui 7d501366cb wsl/librocdxg: fix wgp count calc
Signed-off-by: Flora Cui <flora.cui@amd.com>
Reviewed-by: Longlong Yao <Longlong.Yao@amd.com>
2026-01-05 16:04:18 +08:00
jamessiddeley-amd 53fd27c0ed [rocprofiler-compute] Improve roofline logging for roofline.csv (#2390)
* enhanced roofline log output for graceful exit

* addressed comment, added block filtering

* ruff format
2026-01-02 14:41:28 -05:00
Swati Rawat 3f004c9237 Update using-rocprofv3-with-openmp.rst (#2473) 2026-01-02 22:29:39 +05:30
Sv. Lockal afaa412d9d [rocprofiler-register] Fix compilation with libc++ (#1241)
`tests/rocprofiler/rocprofiler.cpp` uses `std::string` without including `<string>` directly.
This works with libstdc++ due to transitive includes, but fails with libc++.

Closes #1240
2026-01-02 22:26:56 +05:30
Ioannis Assiouras aecc845456 SWDEV-573589 - Fixed performance regression due to the increase of the signal pool (#2470) 2026-01-02 12:50:56 +00:00
Nusrat Islam 57f81914d8 gfx950: restrict maxChannels to 48 for multi-node collectives (#2116)
* gfx950: restrict maxChannels to 48 for multi-node collectives

* change env name for reduced CU config

---------

Co-authored-by: Nusrat Islam <nusislam@dell300x-ccs-aus-k13-09.cs-aus.dcgpu>
Co-authored-by: Islam <nusislam@amd.com>

[ROCm/rccl commit: f756aa9add]
2025-12-31 09:28:19 -06:00
Nusrat Islam f756aa9add gfx950: restrict maxChannels to 48 for multi-node collectives (#2116)
* gfx950: restrict maxChannels to 48 for multi-node collectives

* change env name for reduced CU config

---------

Co-authored-by: Nusrat Islam <nusislam@dell300x-ccs-aus-k13-09.cs-aus.dcgpu>
Co-authored-by: Islam <nusislam@amd.com>
2025-12-31 09:28:19 -06:00
Joseph Narlo 03f714dd25 [SWDEV-567254] Sync Unified and Linux header (#2220)
* [SWDEV-567254] Sync Unified and Linux header

Signed-off-by: Joseph Narlo <joseph.narlo@amd.com>

* Latest sync changes

* Sync

* Add back guest_windows tag

* Sync

---------

Signed-off-by: Joseph Narlo <joseph.narlo@amd.com>
Co-authored-by: amd-josnarlo <josnarlo.amd.com>
2025-12-30 13:27:55 -06:00
vedithal-amd ca32193c84 Fix test cases (#2462) 2025-12-30 11:39:20 -05:00
amd-jiali 7d25ecc65c Add an environment variable to allow user explicitly turn off direct AllGather (#2119)
Co-authored-by: Jiali Li <jialili@amd.com>

[ROCm/rccl commit: 935208ad09]
2025-12-29 16:43:40 -08:00
amd-jiali 935208ad09 Add an environment variable to allow user explicitly turn off direct AllGather (#2119)
Co-authored-by: Jiali Li <jialili@amd.com>
2025-12-29 16:43:40 -08:00
Jimbo a59d46ffbf SWDEV-567545 - Implement block_rank in co-op grid groups (#2182)
* SWDEV-567545 - Implement block_rank in co-op grid groups
2025-12-29 11:39:23 -05:00
Adam Pryor 5bf6e366dd [SWDEV-548460] Add RDC Policy Reset Message (#2180)
* [SWDEV-548460] Add RDC Policy Reset Message

* [rdc] Bump version to 1.3.0

Signed-off-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>

* chore: [rdc] Format CMakeLists.txt

Signed-off-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>

---------

Signed-off-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>
Co-authored-by: Galantsev, Dmitrii <dmitrii.galantsev@amd.com>
2025-12-29 08:31:13 -08:00