maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/Transforms/ScalarizeMaskedMemIntrin/X86/expand-masked-gather.ll

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

67 lines
3.5 KiB
LLVM
Raw Normal View History

Revert "Temporarily Revert "Add basic loop fusion pass."" The reversion apparently deleted the test/Transforms directory. Will be re-reverting again. llvm-svn: 358552
2019-04-17 12:52:47 +08:00
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -S %s -scalarize-masked-mem-intrin -mtriple=x86_64-linux-gnu | FileCheck %s
define <2 x i64> @scalarize_v2i64(<2 x i64*> %p, <2 x i1> %mask, <2 x i64> %passthru) {
; CHECK-LABEL: @scalarize_v2i64(
[ScalarizeMaskedMemIntrin] Bitcast the mask to the scalar domain and use scalar bit tests for the branches. X86 at least is able to use movmsk or kmov to move the mask to the scalar domain. Then we can just use test instructions to test individual bits. This is more efficient than extracting each mask element individually. I special cased v1i1 to use the previous behavior. This avoids poor type legalization of bitcast of v1i1 to i1. I've skipped expandload/compressstore as I think we need to handle constant masks for those better first. Many tests end up with duplicate test instructions due to tail duplication in the branch folding pass. But the same thing happens when constructing similar code in C. So its not unique to the scalarization. Not sure if this lowering code will also be good for other targets, but we're only testing X86 today. Differential Revision: https://reviews.llvm.org/D65319 llvm-svn: 367489
2019-08-01 06:58:15 +08:00
; CHECK-NEXT: [[SCALAR_MASK:%.*]] = bitcast <2 x i1> [[MASK:%.*]] to i2
; CHECK-NEXT: [[TMP1:%.*]] = and i2 [[SCALAR_MASK]], 1
; CHECK-NEXT: [[TMP2:%.*]] = icmp ne i2 [[TMP1]], 0
; CHECK-NEXT: br i1 [[TMP2]], label [[COND_LOAD:%.*]], label [[ELSE:%.*]]
Revert "Temporarily Revert "Add basic loop fusion pass."" The reversion apparently deleted the test/Transforms directory. Will be re-reverting again. llvm-svn: 358552
2019-04-17 12:52:47 +08:00
; CHECK: cond.load:
; CHECK-NEXT: [[PTR0:%.*]] = extractelement <2 x i64*> [[P:%.*]], i64 0
; CHECK-NEXT: [[LOAD0:%.*]] = load i64, i64* [[PTR0]], align 8
; CHECK-NEXT: [[RES0:%.*]] = insertelement <2 x i64> [[PASSTHRU:%.*]], i64 [[LOAD0]], i64 0
; CHECK-NEXT: br label [[ELSE]]
; CHECK: else:
; CHECK-NEXT: [[RES_PHI_ELSE:%.*]] = phi <2 x i64> [ [[RES0]], [[COND_LOAD]] ], [ [[PASSTHRU]], [[TMP0:%.*]] ]
[ScalarizeMaskedMemIntrin] Bitcast the mask to the scalar domain and use scalar bit tests for the branches. X86 at least is able to use movmsk or kmov to move the mask to the scalar domain. Then we can just use test instructions to test individual bits. This is more efficient than extracting each mask element individually. I special cased v1i1 to use the previous behavior. This avoids poor type legalization of bitcast of v1i1 to i1. I've skipped expandload/compressstore as I think we need to handle constant masks for those better first. Many tests end up with duplicate test instructions due to tail duplication in the branch folding pass. But the same thing happens when constructing similar code in C. So its not unique to the scalarization. Not sure if this lowering code will also be good for other targets, but we're only testing X86 today. Differential Revision: https://reviews.llvm.org/D65319 llvm-svn: 367489
2019-08-01 06:58:15 +08:00
; CHECK-NEXT: [[TMP3:%.*]] = and i2 [[SCALAR_MASK]], -2
; CHECK-NEXT: [[TMP4:%.*]] = icmp ne i2 [[TMP3]], 0
; CHECK-NEXT: br i1 [[TMP4]], label [[COND_LOAD1:%.*]], label [[ELSE2:%.*]]
Revert "Temporarily Revert "Add basic loop fusion pass."" The reversion apparently deleted the test/Transforms directory. Will be re-reverting again. llvm-svn: 358552
2019-04-17 12:52:47 +08:00
; CHECK: cond.load1:
; CHECK-NEXT: [[PTR1:%.*]] = extractelement <2 x i64*> [[P]], i64 1
; CHECK-NEXT: [[LOAD1:%.*]] = load i64, i64* [[PTR1]], align 8
; CHECK-NEXT: [[RES1:%.*]] = insertelement <2 x i64> [[RES_PHI_ELSE]], i64 [[LOAD1]], i64 1
; CHECK-NEXT: br label [[ELSE2]]
; CHECK: else2:
; CHECK-NEXT: [[RES_PHI_ELSE3:%.*]] = phi <2 x i64> [ [[RES1]], [[COND_LOAD1]] ], [ [[RES_PHI_ELSE]], [[ELSE]] ]
; CHECK-NEXT: ret <2 x i64> [[RES_PHI_ELSE3]]
;
%ret = call <2 x i64> @llvm.masked.gather.v2i64.v2p0i64(<2 x i64*> %p, i32 8, <2 x i1> %mask, <2 x i64> %passthru)
ret <2 x i64> %ret
}
define <2 x i64> @scalarize_v2i64_ones_mask(<2 x i64*> %p, <2 x i64> %passthru) {
; CHECK-LABEL: @scalarize_v2i64_ones_mask(
; CHECK-NEXT: [[PTR0:%.*]] = extractelement <2 x i64*> [[P:%.*]], i64 0
; CHECK-NEXT: [[LOAD0:%.*]] = load i64, i64* [[PTR0]], align 8
; CHECK-NEXT: [[RES0:%.*]] = insertelement <2 x i64> [[PASSTHRU:%.*]], i64 [[LOAD0]], i64 0
; CHECK-NEXT: [[PTR1:%.*]] = extractelement <2 x i64*> [[P]], i64 1
; CHECK-NEXT: [[LOAD1:%.*]] = load i64, i64* [[PTR1]], align 8
; CHECK-NEXT: [[RES1:%.*]] = insertelement <2 x i64> [[RES0]], i64 [[LOAD1]], i64 1
; CHECK-NEXT: ret <2 x i64> [[RES1]]
;
%ret = call <2 x i64> @llvm.masked.gather.v2i64.v2p0i64(<2 x i64*> %p, i32 8, <2 x i1> <i1 true, i1 true>, <2 x i64> %passthru)
ret <2 x i64> %ret
}
define <2 x i64> @scalarize_v2i64_zero_mask(<2 x i64*> %p, <2 x i64> %passthru) {
; CHECK-LABEL: @scalarize_v2i64_zero_mask(
; CHECK-NEXT: ret <2 x i64> [[PASSTHRU:%.*]]
;
%ret = call <2 x i64> @llvm.masked.gather.v2i64.v2p0i64(<2 x i64*> %p, i32 8, <2 x i1> <i1 false, i1 false>, <2 x i64> %passthru)
ret <2 x i64> %ret
}
define <2 x i64> @scalarize_v2i64_const_mask(<2 x i64*> %p, <2 x i64> %passthru) {
; CHECK-LABEL: @scalarize_v2i64_const_mask(
; CHECK-NEXT: [[PTR1:%.*]] = extractelement <2 x i64*> [[P:%.*]], i64 1
; CHECK-NEXT: [[LOAD1:%.*]] = load i64, i64* [[PTR1]], align 8
; CHECK-NEXT: [[RES1:%.*]] = insertelement <2 x i64> [[PASSTHRU:%.*]], i64 [[LOAD1]], i64 1
; CHECK-NEXT: ret <2 x i64> [[RES1]]
;
%ret = call <2 x i64> @llvm.masked.gather.v2i64.v2p0i64(<2 x i64*> %p, i32 8, <2 x i1> <i1 false, i1 true>, <2 x i64> %passthru)
ret <2 x i64> %ret
}
declare <2 x i64> @llvm.masked.gather.v2i64.v2p0i64(<2 x i64*>, i32, <2 x i1>, <2 x i64>)