Add support for these 2 new queries:
- HSA_AMD_AGENT_INFO_SCRATCH_LIMIT_MAX
Maximum amount of scratch memory allowed on this agent
- HSA_AMD_AGENT_INFO_SCRATCH_LIMIT_CURRENT
Current limit for scratch memory on this agent
Add support for these 2 new queries:
- HSA_AMD_AGENT_INFO_SCRATCH_LIMIT_MAX
Maximum amount of scratch memory allowed on this agent
- HSA_AMD_AGENT_INFO_SCRATCH_LIMIT_CURRENT
Current limit for scratch memory on this agent
[ROCm/ROCR-Runtime commit: 107b48fb15]
Updating ROCr code to match new handshake protocol with CP FW for
asynchronous scratch reclaim.
Increase previous limits when scratch reclaim feature is available.
Updating ROCr code to match new handshake protocol with CP FW for
asynchronous scratch reclaim.
Increase previous limits when scratch reclaim feature is available.
[ROCm/ROCR-Runtime commit: aa2f98e6f9]
* Fix a couple typos. Add note about limitations with AI / ML workloads.
* Updated note based on feedback.
[ROCm/rocprofiler-systems commit: b1d0c431c7]
Add dynamic arrays of devices to avoid array overflow.
Fixed the following tests:
Unit_hipDeviceGetByPCIBusId_WrongBusID)
Unit_hipSetDevice_ValidVisibleDeviceList
Unit_hipSetDevice_MinRvdMaxHvdDevicesList
Change-Id: I73d022321609d20ef2bd419f2cff287f6002a4d2
Allow IPC signals to be registered with hsa_amd_signal_async_handler.
This forces AsyncEventsLoop to switch to polling instead of interrupts.
[ROCm/ROCR-Runtime commit: fa8be44df9]
Changes: - Fixed Device Name (market name)
- Added new API rsmi_dev_market_name_get()
- Updated tests
- Updated amdgpu_drm.h to match latest mainline kernel
- Fixed subsystem ID to only show hex value (not subsystem name)
- rocm_smi_lib now has a recommended requirement for libdrm
Change-Id: Ic438529e16c8c3dbbdd620da664918148c40c997
[ROCm/rocm_smi_lib commit: 6a5e94c451]
Changes: - Fixed Device Name (market name)
- Added new API rsmi_dev_market_name_get()
- Updated tests
- Updated amdgpu_drm.h to match latest mainline kernel
- Fixed subsystem ID to only show hex value (not subsystem name)
- rocm_smi_lib now has a recommended requirement for libdrm
Change-Id: Ic438529e16c8c3dbbdd620da664918148c40c997
- Use correct header in device_library_decl
- use std:: instead of __hip_internal:: for host compilation
- hide device specific stuff behind __clang__ and __HIP__ check
Change-Id: I2f3647e00555ed0e79f9954a459c41394c3cd49b
[ROCm/clr commit: c3f49c8788]
- Use correct header in device_library_decl
- use std:: instead of __hip_internal:: for host compilation
- hide device specific stuff behind __clang__ and __HIP__ check
Change-Id: I2f3647e00555ed0e79f9954a459c41394c3cd49b
- Also add a cache, which allows compiled code objects to be reused
instead of compiling again. This should improve performance on
multigpu systems.
Change-Id: Ib135d616c076b77f8aaf28de275d408b38021d89
[ROCm/clr commit: 0391aec14a]
- Also add a cache, which allows compiled code objects to be reused
instead of compiling again. This should improve performance on
multigpu systems.
Change-Id: Ib135d616c076b77f8aaf28de275d408b38021d89
David Yat Sin
Kandula, Venkateshwar reddy
David Galiffi
Jimbo Xie
Dragoslav Sicarov
Charis Poag
Choudhary, Rahul
Jatin Chaudhary
Elwazir, Ammar
Madsen, Jonathan
Tim Gu
agunashe
Longlong Yao
ywang103-amd
Anusha GodavarthySurya
Adel Johar
Fei Zheng