1. Before, the signal pool is increased depending on the usage
2. After, a static number of signals are allocated to the pool
Only these are used by hip in a stream
3. If the signals required are more than the pool size, the
stream has to wait to make sure all the signals are available
4. Once they are available, the stream can use them
5. Removed HIP_NUM_SIGNALS_PER_STREAM because of redundancy with HIP_STREAM_SIGNALS
6. Increased signal count from 2 to 32.
Future Work: Dynamically increase the pool size depending on the number of
streams allocated by the application. And, null stream should have more signals
Change-Id: I6be36e084f26bb04766fabf776c7210aee0f9e91
[ROCm/clr commit: 9c7ee12822]
Remove dead depFutures, enqueueBarrier call.
Rename some parms to reflect usage.
Add comments to better explain tricky parts of sync code.
Change-Id: I763296421d9c2b3b58fc8cef5f010b12ab49553c
[ROCm/clr commit: ef61aae878]
1. Did not change the logic in allocSignal
2. Added guard to wait on signal limit
Change-Id: I78f29097e6a584b3c3d78319dac19869067bd1fe
[ROCm/clr commit: 1b2a24d0b8]
1. The number of kernels that can use signals are increased to 128
2. The kernel count is now specific to the stream
Change-Id: Ie6d1aa3f437aad8f08c3333fe48bd3f46e551e60
[ROCm/clr commit: 2e754d27dc]
1. The patch uses HIP signal pools to sync between copy and kernel commands
2. The hsa_signal_create is removed
3. Left the redundant enqueueBarrier method just in case
Change-Id: I3dff3e8ee57fff3cd49bec802ff735ed128e5ca1
[ROCm/clr commit: 0232e6bbb4]
ihipStream_t::copySync use GPU agent in memory async copy API, even
if the src/dst memory does not belong to GPU, which cause the hsa
runtime to choose a slower copy engine.
SWDEV-95191
Change-Id: If3cab3d493c0c96ed63721cdcf28247a1193887c
[ROCm/clr commit: dc0a787984]
_ Use fields from GRID_LAUNCH_20 structure
(See USE_GRID_LAUNCH_20 define, currently set to 0)
"1" will require HCC support.
- Remove old DISABLE_GRID_LAUNCH support.
Change-Id: I584ce648d217251789a6283cf27feb24cb7dc8d1
[ROCm/clr commit: e27b5cc927]
- Complete translation tables for cudaError <-> hipError_t.
- Remove some odd errors that were not correctly translated or not used.
- Add HIPCHECK_API to test infrastructure. Used for negative testing
an API ; if a mismatch occurs it shows the expected return error
code. Can also print a warning rather than error.
- Enable hipMemoryAllocate on NV system, and review error coded.
- Add hipErrorName to nvcc.
Change-Id: I680427dcf32a5796d5913cf9e7f3b4c6f6b91599
Conflicts:
tests/src/CMakeLists.txt
Bug fixes and improved docs for hipFree and hipHostFree.
- Passing NULL pointer initialized runtime and return hipSuccess
(not an error like before).
- add negative test for this. (hipMemoryAllocate, improved)
- Match NVCC errors for invalid pointers, add to test.
- Update hipFree and hipHostFree docs.
- hipGetDevicePointer always set *devicePointer=NULL, even for
invalid flags.
- Gate shared memory usage on specific HCC work-week.
Change-Id: I533b4fd3280a3d6cdbf05eb768976f0c7506c012
[ROCm/clr commit: 6a2a140f34]
The completion future of a particular kernel is lost if there are
multiple kernels in the stream. This can cause a racing condition where
the signal associated with the unreferenced completion_future might get
released by hcc runtime.
[ROCm/clr commit: 6773a64b22]
- devicereset would lose track of default stream and thus subsequent
synchronization calls might not actually sychronize.
- Also deviceReset now correctly frees streams.
- fix waits in P2P staging copy - first phase (Device0-to-Staging) must
wait for second phase (Staging to Device1) to finish draining the
buffer.
[ROCm/clr commit: 6abfa13c34]
- add P2P staging buffer copy.
- If copy device does not have sufficient access permissions, fall back
to staging buffer.
- improve docs for which copy device is used.
[ROCm/clr commit: c3bd85595d]
- set USE_PEER_TO_PEER=3 (requires HCC "am_memtracker_update_peers")
- when enabling peer, turn it on for previously allocated memory.
- hipDeviceCanAccessPeer is no longer self-ware (self does not qualify
as a peer)
- device peerlist always includes self, so when we call allow_access
we never remove self access.
- hipDeviceReset() removes old peer mappings.
[ROCm/clr commit: 5af4c901c6]
- set USE_PEER_TO_PEER=3 (requires HCC "am_memtracker_update_peers")
- when enabling peer, turn it on for previously allocated memory.
- hipDeviceCanAccessPeer is no longer self-ware (self does not qualify
as a peer)
- device peerlist always includes self, so when we call allow_access
we never remove self access.
- hipDeviceReset() removes old peer mappings.
[ROCm/clr commit: 9e7efd7c65]
Aditya Atluri
Ben Sander
Rahul Garg
Maneesh Gupta
Fan Cao
Aditya Atluri
bwicakso
pensun