AI/rocm-systems
2
0
Rozštěpit 0
Zdrojový kód Úkoly Pull requesty Akce Balíčky Projekty Vydání Wiki Aktivita
64 531 Commity 3 Větve 0 Značky
f7dba2eb5c1d377dd90b5e261b1c769fe2fe59cf
Graf commitů

2480 Commity

Autor SHA1 Zpráva Datum
Rahul Manocha 538f1c3b74 SWDEV-556205 - fix segfault in hiprtc (#1058) 
Co-authored-by: Rahul Manocha <rmanocha@amd.com>
 2025-09-26 09:06:37 -07:00
Rahul Manocha 2bc561d404 SWDEV-557057 - fix for datatype for hipMemcpy3DBatchAsync (#1114) 
Co-authored-by: Rahul Manocha <rmanocha@amd.com>
 2025-09-25 13:53:23 -07:00
Godavarthy Surya, Anusha fb72d7f851 SWDEV-524746 - Part-II Add multi device support for hip graph. Updated kernel arg manager for each device (#813) 
- Updated kernel arg manager to support allocating kernel args on multiple devices for single graph.
- Updated AQL path to capture on the device where graph node is added.

Co-authored-by: Anusha GodavarthySurya <Anusha.GodavarthySurya@amd.com>
 2025-09-25 20:38:18 +05:30
Ioannis Assiouras 97bc3af918 SWDEV-550882 - Add support for hipIpcMemLazyEnablePeerAccess (#817) 2025-09-23 00:05:51 +01:00
Shadi Dashmiz 9b350754cc SWDEV-555084: Fix the python script (#996) 
- no need to manually updated the newly generated hip_prof_str.h

Signed-off-by: shadi <shadi.dashmiz@amd.com>
 2025-09-22 08:41:19 -04:00
Todd tiantuo Li 7137c7f3d8 SWDEV-541478 - return hipSuccess for hipTexObjectCreate TypePitch2D with zero width or height (#712) 2025-09-19 20:48:01 -07:00
Jatin Chaudhary e79eaaa8a5 SWDEV-546287 - Implement hipLibrary load/unload (#975) 2025-09-19 22:23:49 +01:00
JonathanLichtnerAMD f31afe1d20 [HIP CLR] Make hipMemPtrGetInfo consistent with malloc and hipMalloc (#1005) 
hipMemPtrGetInfo was returning the error hipErrorInvalidValue if it
was called on a nullptr.  However, this does not match the malloc
convention where a nullptr has size zero;  for example,
malloc_usable_size() returns zero if called on a nullptr.

This commit changes hipMemPtrGetInfo to set the size to zero and
return hipSuccess when called with a nullptr.  (This also fits with
hipMalloc and hipFree usage, since hipMalloc of size zero results in a
nullptr, and hipFree of a nullptr is successful.)
 2025-09-19 12:53:41 -06:00
Godavarthy Surya, Anusha 538528d1e5 SWDEV-548417 - Fix Memleaks in Graph (#973) 
Command enqueued on the graph internal stream are not released add stream during graphExec release

Co-authored-by: Rahul Manocha <rmanocha@amd.com>
 2025-09-19 17:45:01 +05:30
Godavarthy Surya, Anusha ce560304a8 SWDEV-548417 - Fix Memleaks in Graph (#713) 
Co-authored-by: Anusha GodavarthySurya <Anusha.GodavarthySurya@amd.com>
 2025-09-19 17:39:36 +05:30
Jaydeep 9f5b390db4 SWDEV-555484 - getQueueId uses hsa_queue's id which is not necessary to be bound by GPU_MAX_HW_QUEUES and hence accessing array beyond size cause data curruption. (#1040) 2025-09-19 14:31:27 +05:30
Jaydeep 99613f1009 SWDEV-555484 - Invalidate capturing stream only for null/legacy stream. (#1032) 2025-09-19 14:31:17 +05:30
SaleelK 149dc17c90 clr: Optimize doorbell ring (#1030) 
*Lay foundation to batch packets efficiently for graphs
*Dynamically copy packets with max threshold set with
DEBUG_HIP_GRAPH_BATCH_SIZE, if not stagger packet copy with pow2
*Default threshold for DEBUG_HIP_GRAPH_BATCH_SIZE is 256
*If TS are not collected for a signal for reuse, create a new signal.
This can potentially increase signal footprint if the handler doesn't run
fast enough.
 2025-09-18 15:02:10 -07:00
systems-assistant[bot] 0018a4e70c SWDEV-541623 - cuda parity hipLaunchCooperativeKernelMultiDevice and hipExtLaunchMultiKernelMultiDevice (#415) 
* SWDEV-541623 - cuda parity hipLaunchCooperativeKernelMultiDevice and hipExtLaunchMultiKernelMultiDevice

numDevices does not match the system devices

* SWDEV-541623 -  enable Unit_hipExtLaunchMultiKernelMultiDevice_Negative_MultiKernelSameDevice

---------

Co-authored-by: agunashe <ajay.gunashekar@amd.com>
 2025-09-17 08:33:59 -07:00
systems-assistant[bot] c85200fc42 SWDEV-541096 - add hipEventWaitDefault and hipEventWaitExternal flags (#507) 
Co-authored-by: Li, Todd tiantuo <Toddtiantuo.Li@amd.com>
 2025-09-11 14:50:55 -07:00
systems-assistant[bot] 3e1e2408a9 SWDEV-541427 - Fix forked stream joining to parent stream that is not origin stream(BeginCaptureStream) (#449) 
Co-authored-by: Anusha GodavarthySurya <Anusha.GodavarthySurya@amd.com>
Co-authored-by: systems-assistant[bot] <systems-assistant[bot]@users.noreply.github.com>
Co-authored-by: Godavarthy Surya, Anusha <agodavar@amd.com>
 2025-09-11 16:57:33 +05:30
systems-assistant[bot] 0647cf1d28 SWDEV-542700 - Return an error if stream capture is attempted on the null stream while a stream capture is active. (#450) 
Co-authored-by: Anusha GodavarthySurya <Anusha.GodavarthySurya@amd.com>
Co-authored-by: systems-assistant[bot] <systems-assistant[bot]@users.noreply.github.com>
Co-authored-by: Godavarthy Surya, Anusha <agodavar@amd.com>
 2025-09-11 16:57:22 +05:30
Ioannis Assiouras 35629e433d SWDEV-546146 - Added support for hipMemLocationTypeHost in hipMemSetAccess (#682) 2025-09-10 23:06:20 +01:00
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
Satyanvesh Dittakavi 85065dab32 SWDEV-550521 - Add the JIT options for HIPRTC linker APIs (#762) 
* SWDEV-550521 - Add the JIT options for HIPRTC linker APIs

* Address review comments about using C++ datatypes
 2025-09-09 12:24:08 +05:30
SaleelK e197aa83ba SWDEV-543723 - Execute permission for kernArg buf (#728) 
- Refactor deviceLocalAlloc arguments
- Refactor hostAlloc code, have cleaner interface
- Kern args buffer need to have execute flag set as CP enforces this on
  certain newer HW.
 2025-09-08 12:21:30 -07:00
vstojilj f17e332fe0 Release graph if hipStreamEndCapture fails (#738) 2025-09-08 16:32:03 +02:00
Todd tiantuo Li c8ecf77a94 Update dispatch table to move 7.1 new APIs under HIP_RUNTIME_API_TABLE_STEP_VERSION 14 (#790) 2025-09-05 14:14:43 -07:00
Jimbo 3d9d35a1f8 SWDEV-553375 - Allow hipMemAllocationTypeUncached in hipMemGetAllocationGranularity (#847) 2025-09-05 10:31:20 -04:00
Ajay GunaShekar 2202dcfe80 SWDEV-552613 - Windows: Use Direct Dispatch only on HSA ROCr Backend (#809) 
* SWDEV-552613 - Disable Direct Dispatch on Windows

* SWDEV-552613 -  Use Direct Dispatch on HSA backend only

---------

Co-authored-by: GunaShekar <agunashe@amd.com>
Co-authored-by: Christophe Paquot <35546540+chrispaquot@users.noreply.github.com>
 2025-09-04 14:07:12 -07:00
SaleelK c4537e8050 SWDEV-553126 - Improve logging (#835) 
* Ability to mask COPY api usage in logs
* Show total graph nodes in logs
* Add another log level for detailed debug
 2025-09-04 10:08:41 -07:00
marandje 79bda80049 SWDEV-549686 - Resolve memory leaks in texture unit-tests (#711) 2025-09-04 17:59:18 +02:00
systems-assistant[bot] 12ca3c9043 SWDEV-548482 - Address memory leaks in memory tests (#438) 
* SWDEV-548482 - Address memory leaks in memory tests

* SWDEV-547453 - Do not alter the dev_ptr if operation is not successfull

* SWDEV-548482 - Minor tweaks

* SWDEV-548482 - Move eventlist release after the command is created

---------

Co-authored-by: Marko Arandjelovic <Marko.Arandjelovic@amd.com>
 2025-09-02 17:29:41 +02:00
Ioannis Assiouras 1017532916 SWDEV-546631 - Fix hipLaunchHostFunction in stream capture for windows (#654) 2025-08-28 07:51:50 +01:00
Sourabh U Betigeri 5201efe050 SWDEV-545259 - Adds clang-cl support and missing dispatch table entries (#760) 2025-08-27 09:35:58 -07:00
systems-assistant[bot] ded5b86e83 SWDEV-540609 - capture of MIOpen OCL kernels needs remainder globalWorkSize (#431) 
Co-authored-by: Rakesh Roy <rakesh.roy@amd.com>
 2025-08-26 16:11:31 -04:00
systems-assistant[bot] 7601798fa7 SWDEV-545953 - Add Implementation for hipStreamGetId (#434) 
Authored-by: Satyanvesh Dittakavi <Satyanvesh.Dittakavi@amd.com>
 2025-08-26 22:47:55 +05:30
Jimbo c03048d68e Implement hipMemAllocationTypeUncached in hipMemCreate (#747) 
* 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
 2025-08-26 11:34:49 -04:00
Danylo Lytovchenko 2ff2316227 Adjust clang format to the new versions, revert broken macro layout (#714) 2025-08-22 17:23:22 +02:00
Todd tiantuo Li bbf0d34b48 SWDEV-550231 - fix hipGetDriverEntryPoint_common (#651) 2025-08-21 18:57:16 -07:00
systems-assistant[bot] 4d5db26646 SWDEV-549707 - Fixed memleak in hipMemImportFromShareableHandle (#442) 
Co-authored-by: Ioannis Assiouras <Ioannis.Assiouras@amd.com>
 2025-08-20 19:28:59 +01:00
Danylo Lytovchenko f7338717ae SWDEV-470698 - fix formatting, add format check workflow (#657) 2025-08-20 19:58:06 +05:30
Rahul Manocha f4158ed3b7 SWDEV-546295 - fix param conversions for hipMemcpy3DBatchAsync (#641) 
Co-authored-by: Rahul Manocha <rmanocha@amd.com>
 2025-08-19 20:30:03 -07:00
Andryeyev, German 72b9408fed SWDEV-547108 - Fix compilation errors under Windows (#867) 
Interop and numa are not enabled.

[ROCm/clr commit: 0ac913e64c]
 2025-08-17 02:33:31 -04:00
Betigeri, Sourabh 35e48d1eaf SWDEV-546293 - hipMemPrefetchAsync_v2 and hipMemAdvise_v2 implementation (#869) 
SWDEV-546293 - hipMemPrefetchAsync hipMemAdvise_v2

Please enter the commit message for your changes. Lines starting

[ROCm/clr commit: cbee74a80e]
 2025-08-15 22:40:04 -07:00
Li, Todd tiantuo ad9eb56dd4 SWDEV-546285 - add hipGetDriverEntryPoint (#855) 
[ROCm/clr commit: 789e2029ca]
 2025-08-15 20:08:21 -07:00
Xie, Jiabao(Jimbo) 5ce7103555 SWDEV-547589 - Add hipDeviceMallocUncached to hipMemCreate (#815) 
Co-authored-by: Jimbo Xie <jiabaxie@amd.com>

[ROCm/clr commit: bd1e489f1f]
 2025-08-15 16:02:26 -07:00
Kudchadker, Saleel 9bb6fb0b79 SWDEV-537480 - Fix file handle cleanup (#523) 
[ROCm/clr commit: b90f3c2507]
 2025-08-15 15:32:12 -07:00
Xie, Jiabao(Jimbo) 42a97a2c16 SWDEV-477219 - fix using external api calls on hipEventRecordWithFlags (#650) 
Co-authored-by: Jimbo Xie <jiabaxie@amd.com>

[ROCm/clr commit: e8128f975f]
 2025-08-15 13:40:52 -07:00
Manocha, Rahul 8fb910f69b SWDEV-546295 - Add new APIs hipMemcpy3DBatchAsync and hipMemcpy3DPeer (#818) 
Co-authored-by: Rahul Manocha <rmanocha@amd.com>

[ROCm/clr commit: e3cf50c96d]
 2025-08-15 13:18:29 -07:00
GunaShekar, Ajay 0d63a5abd3 SWDEV-546178 - hipModuleLoadFatBinary implementation (#785) 
* SWDEV-546178 - hipModuleLoadFatBinary implementation

* SWDEV-546178 - hipModuleLoadFatBinary CHANGELOG update

* SWDEV-546178 - fix roctracer build issue due to hip_prof_str.h

[ROCm/clr commit: f6736d57a5]
 2025-08-15 13:06:24 -07:00
Manocha, Rahul b3ccf487da SWDEV-545952 - API definitions for hipStreamSet/GetAttribute (#831) 
Co-authored-by: Rahul Manocha <rmanocha@amd.com>

[ROCm/clr commit: 0f49c4a97f]
 2025-08-15 12:51:35 -07:00
GunaShekar, Ajay daa2814cf3 SWDEV-541623 - cuda parity hipTexObjectCreate when Array is null (#734) 
[ROCm/clr commit: 54e592beec]
 2025-08-15 12:35:33 -07:00
Arandjelovic, Marko 4e4caf3ff8 SWDEV-512604 - Use the same context during Image creation (#800) 
[ROCm/clr commit: be73d024e2]
 2025-08-15 18:31:11 +02:00