diff options
| author | Erich Keane <ekeane@nvidia.com> | 2025-04-21 11:51:26 -0700 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-04-21 11:51:26 -0700 |
| commit | 0ae9dac262afccea1e1a2e02520f737ab38b286c (patch) | |
| tree | fc76c215d83a7f0b42e409c92d153b5391f8957a /llvm/lib/Analysis/AssumptionCache.cpp | |
| parent | e0c1e23b99e9719d0a01ab7dfc0807d891004bd4 (diff) | |
| download | llvm-0ae9dac262afccea1e1a2e02520f737ab38b286c.zip llvm-0ae9dac262afccea1e1a2e02520f737ab38b286c.tar.gz llvm-0ae9dac262afccea1e1a2e02520f737ab38b286c.tar.bz2 | |
[OpenACC][CIR] Lower 'num_workers' for parallel/kernels (#136578)
This patch also includes the first one to handle 'device_type' properly,
which is where most of the 'challenge' here comes from.
From the best I can tell: we must keep two lists of the same size, 1 of
all of the 'num_workers' items, and 1 of the 'device_type' value for
that 'num_workers'. Additionally, the 'device_type' list can only handle
single 'device_type' values, so we have to duplicate the 'num_workers'
items in cases where there are multiple applicable 'device_type' values.
This patch accomplishes this by keeping the two in sync, and saving the
current 'device_type' in the visitor.
Diffstat (limited to 'llvm/lib/Analysis/AssumptionCache.cpp')
0 files changed, 0 insertions, 0 deletions