Pinned copy can cause big performance drops, because slow pinning under Windows.
Use up to 128MB for staging transfers. Change staging buffer size to 4MB.
Linux path should still have the old defaults.
Change-Id: I954edceb3ec89e8e670be116aa2d0a9564c8b11c
Add a env var ROC_USE_FGS_KERNARG to toggle kernel arg placement
By default its in Fine Grain Kernel arg segment for supported asics.
Change-Id: I3d57ed69a1a4db2b392b0438ead499f3ddca4716
- The logic will trace compute, sdma read/write operations and
apply signals when necessary
- ROC_CPU_WAIT_FOR_SIGNAL, ROC_SYSTEM_SCOPE_SIGNAL
and ROC_SKIP_COPY_SYNC were added to control the tracking
Change-Id: I9e8e6174c63bf7784f7ab00964e2918c8667d364
- Correct GSL path to report targets using the TargetID syntax.
- Correct GSL path to check compatibility of code objects when
loading.
- Add concept of an device isa and create a registery used by ROCm,
PAL and GSL.
- Support XNACK and SRAMECC target features consistently for PAL and ROCm.
- Correct logic for NullDevices and asserts to avoid memory coruption.
- Allow all NullDevices to be created for HIP.
- Numerous other code improvements.
Change-Id: I40abf3d2b22249c1492d1af5919665f8184f4e0e
The change reuses HSA signals for dispatches as a wait signal.
Skipping the barrier requires to disable L2 cache for sysmem
allocations and extra tracking for HDP access with the large bar.
ROC_BARRIER_SYNC=0 activates the new logic. Barrier sync is
still used by default.
ROC_ACTIVE_WAIT=1 enables unconditional active wait in ROCr.
The change also consolidated ROCr wait logic under single function.
Change-Id: I6bd1be30aa88258da1b1f9de319ef5a45852afd8
Fix a typo with the name define, when compilation wasn't enabled.
Force CPU prefetch if system was forced in runtime
Change-Id: Id4b578f9fa44a45426fdb5d8ecb1da803aa42313
- Expose ROCclr interfaces for HIP usage
- ROCr interfaces aren't available in staging, thus control the
build with AMD_HMM_SUPPORT define
Change-Id: Iadc2bcc230e78d3b0dc22b235189c8cc80843446
Remove queue limitation since we loop through HW queues now.
Add a DevLogError if we fail to create the hsa_queue. A ticket showed a regression there.
Change-Id: I4f58e405f88e75600a762f6d6352838c969cdb5e
This workaround is to avoid performance penalty of SDMA engine
taking a while to clock up from a lower DPM state. Add env var
GPU_FORCE_BLIT_COPY_SIZE (1024 by default for HIP in KB). Forcing
Src and Dst agent to be amdgpu makes ROCr take blit copy path for
what otherwise should have been SDMA copy
Change-Id: I222f687155f86000d17d66d25182e490b6710463
1. Enable pitch workaround
2. When we use copy image, we don't need to create the custom pitch image
3. wrtBackImageBuffer_ stores device memory object, not amd image object.
Tests:
conformance kernel read / write test pass with this code change.
Change-Id: I7dca3127adde6ac83e78dd270a2256ebed55c60d
German
Saleel Kudchadker
Sourabh Betigeri
Satyanvesh Dittakavi
Jason Tang
agunashe
Tony Tye
kjayapra-amd
Jason Tang
Aakash Sudhanwa
Christophe Paquot
Alex Xie
Laurent Morichetti