Mirror intel/llvm commits #2813
Conversation
This is an optimisation that uses a direct memcopy from host to device.
If number of work group is too large, local/private shadow memory will consume a lot of device memory and easily cause out of resource issue. So we'd better set a limitation for it. Also include other changes in this commit: 1.Allocate private base/shadow per subgroup instead of workgroup for GPU device. 2.Refine the algorithm to calculate Subgroup linear Id.
`EXPECT` doesn't return from the current function, so a number of tests would keep going if they failed, in one case segfaulting due to writing to memory that was not allocated correctly. Standardise the test suite so that it always returns on the first failure using `ASSERT` where possible.
V2 stores device handles in a vector of size `Platform->getNumDevices()`. If there are more devices than that, we may have out-of-range acces.
Unified Runtime -> intel/llvm Repo Move NoticeInformationThe source code of Unified Runtime has been moved to intel/llvm under the unified-runtime top-level directory, The code will be mirrored to oneapi-src/unified-runtime and the specification will continue to be hosted at oneapi-src.github.io/unified-runtime. The contribution guide will be updated with new instructions for contributing to Unified Runtime. PR MigrationAll open PRs including this one will be marked with the Should you wish to continue with your PR you will need to migrate it to intel/llvm. If your PR should remain open and not be closed automatically, you can remove the This is an automated comment. |
5 participants
Automated changes by create-pull-request GitHub action