[Mlir-commits] [mlir] [mlir][vector][nfc] Fix typos in "VectorEmulateNarrowType.cpp" (PR #125415)

Andrzej WarzyƄski llvmlistbot at llvm.org
Sun Feb 2 07:26:16 PST 2025


https://github.com/banach-space created https://github.com/llvm/llvm-project/pull/125415

Updates `emulatedVectorLoad` that was introduced in #115922.
Specifically, ATM `emulatedVectorLoad` mixes "emulated type" and
"container type". This only became clear after #123526 in which the
concepts of "emulated" and "container" types were introduced.

This is an NFC change and simply updates the variable naming.


>From 3954c8f5dafae1f3435d490b3f114420dabbe925 Mon Sep 17 00:00:00 2001
From: Andrzej Warzynski <andrzej.warzynski at arm.com>
Date: Sun, 2 Feb 2025 15:21:41 +0000
Subject: [PATCH] [mlir][vector][nfc] Fix typos in
 "VectorEmulateNarrowType.cpp"

Updates `emulatedVectorLoad` that was introduced in #115922.
Specifically, ATM `emulatedVectorLoad` mixes "emulated type" and
"container type". This only became clear after #123526 in which the
concepts of "emulated" and "container" types were introduced.

This is an NFC change and simply updates the variable naming.
---
 .../Transforms/VectorEmulateNarrowType.cpp    | 24 ++++++++++---------
 1 file changed, 13 insertions(+), 11 deletions(-)

diff --git a/mlir/lib/Dialect/Vector/Transforms/VectorEmulateNarrowType.cpp b/mlir/lib/Dialect/Vector/Transforms/VectorEmulateNarrowType.cpp
index 63365cb5446124..0d310dc8be2fe9 100644
--- a/mlir/lib/Dialect/Vector/Transforms/VectorEmulateNarrowType.cpp
+++ b/mlir/lib/Dialect/Vector/Transforms/VectorEmulateNarrowType.cpp
@@ -278,23 +278,25 @@ static Value dynamicallyInsertSubVector(RewriterBase &rewriter, Location loc,
   return dest;
 }
 
-/// Returns the op sequence for an emulated sub-byte data type vector load.
-/// specifically, use `emulatedElemType` for loading a vector of `origElemType`.
-/// The load location is given by `base` and `linearizedIndices`, and the
-/// load size is given by `numEmulatedElementsToLoad`.
+/// Emulate a vector load for `emulatedElemTy` using `containerElemTy`
+///
+/// Specifically, use `containerElemTy` for loading a vector of
+/// `emulatedElemTy`. The load location is given by `base` and
+/// `linearizedIndices`, and the load size is given by
+/// `numEmulatedElementsToLoad`.
 static VectorValue emulatedVectorLoad(OpBuilder &rewriter, Location loc,
                                       Value base,
                                       OpFoldResult linearizedIndices,
-                                      int64_t numEmultedElementsToLoad,
-                                      Type origElemType,
-                                      Type emulatedElemType) {
-  auto scale = emulatedElemType.getIntOrFloatBitWidth() /
-               origElemType.getIntOrFloatBitWidth();
+                                      int64_t numContainerElemsToLoad,
+                                      Type emulatedElemTy,
+                                      Type containerElemTy) {
+  auto scale = containerElemTy.getIntOrFloatBitWidth() /
+               emulatedElemTy.getIntOrFloatBitWidth();
   auto newLoad = rewriter.create<vector::LoadOp>(
-      loc, VectorType::get(numEmultedElementsToLoad, emulatedElemType), base,
+      loc, VectorType::get(numContainerElemsToLoad, containerElemTy), base,
       getValueOrCreateConstantIndexOp(rewriter, loc, linearizedIndices));
   return rewriter.create<vector::BitCastOp>(
-      loc, VectorType::get(numEmultedElementsToLoad * scale, origElemType),
+      loc, VectorType::get(numContainerElemsToLoad * scale, emulatedElemTy),
       newLoad);
 }
 



More information about the Mlir-commits mailing list