aboutsummaryrefslogtreecommitdiff
path: root/llvm/lib/Support/Parallel.cpp
diff options
context:
space:
mode:
authorAndy Davis <andydavis@google.com>2020-03-31 08:21:04 -0700
committerAndy Davis <andydavis@google.com>2020-03-31 08:39:17 -0700
commit31a346cc35c829e207fbf24d8f8f1ee22ca9f140 (patch)
tree610b416045a73e9967dd414a601f88c2f5d2634d /llvm/lib/Support/Parallel.cpp
parent34756a1c709d1ef0292413cddaeeef8a21ed2124 (diff)
downloadllvm-31a346cc35c829e207fbf24d8f8f1ee22ca9f140.zip
llvm-31a346cc35c829e207fbf24d8f8f1ee22ca9f140.tar.gz
llvm-31a346cc35c829e207fbf24d8f8f1ee22ca9f140.tar.bz2
[MLIR][Vector] Add support for TupleGetOp folding through InsertSlicesOp and ExtractSlicesOp.
Summary: Add support for TupleGetOp folding through InsertSlicesOp and ExtractSlicesOp. Vector-to-vector transformations for unrolling and lowering to hardware vectors can generate chains of structured vector operations (InsertSlicesOp, ExtractSlicesOp and ShapeCastOp) between the producer of a hardware vector value and its consumer. Because InsertSlicesOp, ExtractSlicesOp and ShapeCastOp are structured, we can track the location (tuple index and vector offsets) of the consumer vector value through the chain of structured operations to the producer, enabling a much more powerful producer-consumer fowarding of values through structured ops and tuple, which in turn enables a more powerful TupleGetOp folding transformation. Reviewers: nicolasvasilache, aartbik Reviewed By: aartbik Subscribers: grosul1, mehdi_amini, rriddle, jpienaar, burmako, shauheen, antiagainst, arpith-jacob, mgester, lucyrfox, liufengdb, Joonsoo, llvm-commits Tags: #llvm Differential Revision: https://reviews.llvm.org/D76889
Diffstat (limited to 'llvm/lib/Support/Parallel.cpp')
0 files changed, 0 insertions, 0 deletions