[PATCH] D100370: [AArch64][SVE] Fix bug in lowering of fixed-length integer vector divides

[Paul Walker via Phabricator via llvm-commits]

paulwalker-arm added inline comments.


================
Comment at: llvm/test/CodeGen/AArch64/sve-fixed-length-int-div.ll:1
 ; RUN: llc -aarch64-sve-vector-bits-min=128  -asm-verbose=0 < %s | FileCheck %s -D#VBYTES=16 -check-prefix=NO_SVE
+; RUN: llc -aarch64-sve-vector-bits-min=256  -asm-verbose=0 < %s | FileCheck %s -D#VBYTES=32 -check-prefixes=CHECK,VBITS_EQ_256
----------------
Given the updates I wondered if this is still necessary?


================
Comment at: llvm/test/CodeGen/AArch64/sve-fixed-length-int-div.ll:4
 ; RUN: llc -aarch64-sve-vector-bits-min=384  -asm-verbose=0 < %s | FileCheck %s -D#VBYTES=32
-; RUN: llc -aarch64-sve-vector-bits-min=512  -asm-verbose=0 < %s | FileCheck %s -D#VBYTES=64 -check-prefixes=CHECK,VBITS_GE_512
+; RUN: llc -aarch64-sve-vector-bits-min=512  -asm-verbose=0 < %s | FileCheck %s -D#VBYTES=64 -check-prefixes=CHECK,VBITS_EQ_512,VBITS_GE_512
 ; RUN: llc -aarch64-sve-vector-bits-min=640  -asm-verbose=0 < %s | FileCheck %s -D#VBYTES=64 -check-prefixes=CHECK,VBITS_GE_512
----------------
OCD perhaps but can you move this and `VBITS_EQ_1024` to the end of the list so the `GE` parts remain aligned.


================
Comment at: llvm/test/CodeGen/AArch64/sve-fixed-length-int-div.ll:74
+
+; LESS-THAN HALF VECTOR:
+; VBITS_GE_512: ptrue [[PG:p[0-9]+]].s, vl16
----------------
To be precise this should be `QUARTER VECTOR OR SMALLER`. Same goes for the other instances that follow.


================
Comment at: llvm/test/CodeGen/AArch64/sve-fixed-length-int-div.ll:251
+
+; LESS-THAN-OR-EQUAL-TO HALF VECTOR:
+; VBITS_GE_2048: ptrue [[PG1:p[0-9]+]].b, vl128
----------------
What is different here from the previous tests? I would have thought `HALF VECTOR` would carry the same meaning.


================
Comment at: llvm/test/CodeGen/AArch64/sve-fixed-length-int-rem.ll:1
 ; RUN: llc -aarch64-sve-vector-bits-min=128  -asm-verbose=0 < %s | FileCheck %s -D#VBYTES=16 -check-prefix=NO_SVE
+; RUN: llc -aarch64-sve-vector-bits-min=256  -asm-verbose=0 < %s | FileCheck %s -D#VBYTES=32 -check-prefixes=CHECK,VBITS_EQ_256
----------------
As above. Is this still used?  I imagine most of my `div.ll` comments will apply to this file.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D100370/new/

https://reviews.llvm.org/D100370