1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-22 18:54:02 +01:00
llvm-mirror/test/CodeGen/AArch64/sve-ptest-removal-cmpgt.ll
Bradley Smith a0a984398f [AArch64][SVE] Add optimization to remove redundant ptest instructions
Co-Authored-by: Graham Hunter <graham.hunter@arm.com>
Co-Authored-by: Paul Walker <paul.walker@arm.com>

Differential Revision: https://reviews.llvm.org/D93292
2021-01-05 15:28:36 +00:00

55 lines
2.0 KiB
LLVM

; RUN: llc -mtriple=aarch64--linux-gnu -mattr=+sve %s -o - | FileCheck %s
;
; Compares
;
define i32 @cmpgt_nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b) {
; CHECK-LABEL: cmpgt_nxv16i8:
; CHECK: cmpgt p0.b, p0/z, z0.b, z1.b
; CHECK-NEXT: cset w0, ne
; CHECK-NEXT: ret
%1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
%2 = tail call i1 @llvm.aarch64.sve.ptest.any(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %1)
%conv = zext i1 %2 to i32
ret i32 %conv
}
;
; Immediate Compares
;
define i32 @cmpgt_imm_nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %a) {
; CHECK-LABEL: cmpgt_imm_nxv16i8:
; CHECK: cmpgt p0.b, p0/z, z0.b, #0
; CHECK-NEXT: cset w0, ne
; CHECK-NEXT: ret
%1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %a, <vscale x 16 x i8> zeroinitializer)
%2 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32 31)
%3 = tail call i1 @llvm.aarch64.sve.ptest.any(<vscale x 16 x i1> %2, <vscale x 16 x i1> %1)
%conv = zext i1 %3 to i32
ret i32 %conv
}
;
; Wide Compares
;
define i32 @cmpgt_wide_nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %a, <vscale x 2 x i64> %b) {
; CHECK-LABEL: cmpgt_wide_nxv16i8:
; CHECK: cmpgt p0.b, p0/z, z0.b, z1.d
; CHECK-NEXT: cset w0, ne
; CHECK-NEXT: ret
%1 = tail call <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv16i8(<vscale x 16 x i1> %pg, <vscale x 16 x i8> %a, <vscale x 2 x i64> %b)
%2 = tail call i1 @llvm.aarch64.sve.ptest.any(<vscale x 16 x i1> %pg, <vscale x 16 x i1> %1)
%conv = zext i1 %2 to i32
ret i32 %conv
}
declare <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
declare <vscale x 16 x i1> @llvm.aarch64.sve.cmpgt.wide.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 2 x i64>)
declare i1 @llvm.aarch64.sve.ptest.any(<vscale x 16 x i1>, <vscale x 16 x i1>)
declare <vscale x 16 x i1> @llvm.aarch64.sve.ptrue.nxv16i1(i32)