llvm-project.git/libc/test/integration/src/__support/GPU/scan_reduce.cpp, branch main Unnamed repository; edit this file 'description' to name the repository. [Clang] Fix cross-lane scan when given divergent lanes (#127703) 2025-02-19T22:46:59+00:00 Joseph Huber huberjn@outlook.com 2025-02-19T22:46:59+00:00 6cc7ca084a5bbb7ccf606cab12065604453dde59 Summary: The scan operation implemented here only works if there are contiguous ones in the executation mask that can be used to propagate the result. There are two solutions to this, one is to enter 'whole-wave-mode' and forcibly turn them back on, or to do this serially. This implementation does the latter because it's more portable, but checks to see if the parallel fast-path is applicable. Needs to be backported for correct behavior and because it fixes a failing libc test. 
Summary:
The scan operation implemented here only works if there are contiguous
ones in the executation mask that can be used to propagate the result.
There are two solutions to this, one is to enter 'whole-wave-mode' and
forcibly turn them back on, or to do this serially. This implementation
does the latter because it's more portable, but checks to see if the
parallel fast-path is applicable.

Needs to be backported for correct behavior and because it fixes a
failing libc test.
 [libc] Add utility functions for warp-level scan and reduction (#84866) 2024-03-12T15:40:49+00:00 Joseph Huber huberjn@outlook.com 2024-03-12T15:40:49+00:00 261e5648e70b363aecf86acfcd7fb416eb48fb7b Summary: The GPU uses a SIMT execution model. That means that each value actually belongs to a group of 32 or 64 other lanes executing next to it. These platforms offer some intrinsic fuctions to actually take elements from neighboring lanes. With these we can do parallel scans or reductions. These functions do not have an immediate user, but will be used in the allocator interface that is in-progress and are generally good to have. This patch is a precommit for these new utilitly functions. 
Summary:
The GPU uses a SIMT execution model. That means that each value actually
belongs to a group of 32 or 64 other lanes executing next to it. These
platforms offer some intrinsic fuctions to actually take elements from
neighboring lanes. With these we can do parallel scans or reductions.
These functions do not have an immediate user, but will be used in the
allocator interface that is in-progress and are generally good to have.
This patch is a precommit for these new utilitly functions.