This is an archive of the discontinued LLVM Phabricator instance.

Differential D90164

[mlir] NFC: properly align IR in comments
ClosedPublic

Authored by antiagainst on Oct 26 2020, 8:47 AM.

Details

Reviewers
aartbik
nicolasvasilache
Commits
rGf52b4a65f09c: [mlir] NFC: properly align IR in comments

Diff Detail

Event Timeline

antiagainst created this revision.Oct 26 2020, 8:47 AM
Herald added a reviewer: aartbik. · View Herald TranscriptOct 26 2020, 8:47 AM
Herald added a project: Restricted Project. · View Herald Transcript
Herald added subscribers: rdzhabarov, tatianashp, msifontes and 12 others. · View Herald Transcript
antiagainst requested review of this revision.Oct 26 2020, 8:47 AM
Herald added a reviewer: nicolasvasilache. · View Herald TranscriptOct 26 2020, 8:47 AM
Herald added subscribers: stephenneuendorffer, nicolasvasilache. · View Herald Transcript
Harbormaster completed remote builds in B76414: Diff 300688.Oct 26 2020, 9:15 AM
aartbik accepted this revision.Oct 26 2020, 10:13 AM
This revision is now accepted and ready to land.Oct 26 2020, 10:13 AM
Closed by commit rGf52b4a65f09c: [mlir] NFC: properly align IR in comments (authored by antiagainst). · Explain WhyOct 26 2020, 2:58 PM
This revision was automatically updated to reflect the committed changes.
antiagainst added a commit: rGf52b4a65f09c: [mlir] NFC: properly align IR in comments.

Revision Contents

PathSize
mlir/
include/
mlir/
Dialect/
Vector/
6 lines

Diff 300801

mlir/include/mlir/Dialect/Vector/VectorTransforms.h

Show First 20 Lines • Show All 234 Lines • ▼ Show 20 Lines
/// *all* values in [0 .. multiplicity - 1] (e.g. loop induction variable or /// *all* values in [0 .. multiplicity - 1] (e.g. loop induction variable or
/// SPMD id). This transformation only inserts /// SPMD id). This transformation only inserts
/// vector.extract_map/vector.insert_map. It is meant to be used with /// vector.extract_map/vector.insert_map. It is meant to be used with
/// canonicalizations pattern to propagate and fold the vector /// canonicalizations pattern to propagate and fold the vector
/// insert_map/extract_map operations. /// insert_map/extract_map operations.
/// Transforms: /// Transforms:
// %v = addf %a, %b : vector<32xf32> // %v = addf %a, %b : vector<32xf32>
/// to: /// to:
/// %v = addf %a, %b : vector<32xf32> %ev = /// %v = addf %a, %b : vector<32xf32>
/// vector.extract_map %v, %id, 32 : vector<32xf32> into vector<1xf32> %nv = /// %ev = vector.extract_map %v, %id, 32 : vector<32xf32> into vector<1xf32>
/// vector.insert_map %ev, %id, 32 : vector<1xf32> into vector<32xf32> /// %nv = vector.insert_map %ev, %id, 32 : vector<1xf32> into vector<32xf32>
Optional<DistributeOps> distributPointwiseVectorOp(OpBuilder &builder, Optional<DistributeOps> distributPointwiseVectorOp(OpBuilder &builder,
Operation *op, Value id, Operation *op, Value id,
int64_t multiplicity); int64_t multiplicity);
/// Canonicalize an extra element using the result of a pointwise operation. /// Canonicalize an extra element using the result of a pointwise operation.
/// Transforms: /// Transforms:
/// %v = addf %a, %b : vector32xf32> /// %v = addf %a, %b : vector32xf32>
/// %dv = vector.extract_map %v, %id, 32 : vector<32xf32> into vector<1xf32> /// %dv = vector.extract_map %v, %id, 32 : vector<32xf32> into vector<1xf32>
/// to: /// to:
▲ Show 20 Lines • Show All 195 Lines • Show Last 20 Lines