[llvm] 29f97ec - [AArch64] Mul fold tests for D130874. NFC
David Green via llvm-commits
llvm-commits at lists.llvm.org
Tue Aug 2 04:29:45 PDT 2022
Author: David Green
Date: 2022-08-02T12:29:40+01:00
New Revision: 29f97ec8450ebae4e02581757937d26ed4c81aac
URL: https://github.com/llvm/llvm-project/commit/29f97ec8450ebae4e02581757937d26ed4c81aac
DIFF: https://github.com/llvm/llvm-project/commit/29f97ec8450ebae4e02581757937d26ed4c81aac.diff
LOG: [AArch64] Mul fold tests for D130874. NFC
Added:
llvm/test/CodeGen/AArch64/mulcmle.ll
Modified:
Removed:
################################################################################
diff --git a/llvm/test/CodeGen/AArch64/mulcmle.ll b/llvm/test/CodeGen/AArch64/mulcmle.ll
new file mode 100644
index 000000000000..3b2c09cef851
--- /dev/null
+++ b/llvm/test/CodeGen/AArch64/mulcmle.ll
@@ -0,0 +1,149 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=aarch64-none-eabi %s -o - | FileCheck %s
+
+define <1 x i64> @v1i64(<1 x i64> %a) {
+; CHECK-LABEL: v1i64:
+; CHECK: // %bb.0:
+; CHECK-NEXT: movi v1.2s, #1
+; CHECK-NEXT: ushr d0, d0, #31
+; CHECK-NEXT: and v0.8b, v0.8b, v1.8b
+; CHECK-NEXT: fmov x8, d0
+; CHECK-NEXT: lsl x9, x8, #32
+; CHECK-NEXT: sub x8, x9, x8
+; CHECK-NEXT: fmov d0, x8
+; CHECK-NEXT: ret
+ %b = lshr <1 x i64> %a, <i64 31>
+ %c = and <1 x i64> %b, <i64 4294967297>
+ %d = mul nuw <1 x i64> %c, <i64 4294967295>
+ ret <1 x i64> %d
+}
+
+define <2 x i64> @v2i64(<2 x i64> %a) {
+; CHECK-LABEL: v2i64:
+; CHECK: // %bb.0:
+; CHECK-NEXT: movi v1.4s, #1
+; CHECK-NEXT: ushr v0.2d, v0.2d, #31
+; CHECK-NEXT: and v0.16b, v0.16b, v1.16b
+; CHECK-NEXT: fmov x9, d0
+; CHECK-NEXT: mov x8, v0.d[1]
+; CHECK-NEXT: lsl x10, x9, #32
+; CHECK-NEXT: sub x9, x10, x9
+; CHECK-NEXT: lsl x10, x8, #32
+; CHECK-NEXT: sub x8, x10, x8
+; CHECK-NEXT: fmov d0, x9
+; CHECK-NEXT: mov v0.d[1], x8
+; CHECK-NEXT: ret
+ %b = lshr <2 x i64> %a, <i64 31, i64 31>
+ %c = and <2 x i64> %b, <i64 4294967297, i64 4294967297>
+ %d = mul nuw <2 x i64> %c, <i64 4294967295, i64 4294967295>
+ ret <2 x i64> %d
+}
+
+define <2 x i32> @v2i32(<2 x i32> %a) {
+; CHECK-LABEL: v2i32:
+; CHECK: // %bb.0:
+; CHECK-NEXT: movi v1.4h, #1
+; CHECK-NEXT: movi d2, #0x00ffff0000ffff
+; CHECK-NEXT: ushr v0.2s, v0.2s, #15
+; CHECK-NEXT: and v0.8b, v0.8b, v1.8b
+; CHECK-NEXT: mul v0.2s, v0.2s, v2.2s
+; CHECK-NEXT: ret
+ %b = lshr <2 x i32> %a, <i32 15, i32 15>
+ %c = and <2 x i32> %b, <i32 65537, i32 65537>
+ %d = mul nuw <2 x i32> %c, <i32 65535, i32 65535>
+ ret <2 x i32> %d
+}
+
+define <4 x i32> @v4i32(<4 x i32> %a) {
+; CHECK-LABEL: v4i32:
+; CHECK: // %bb.0:
+; CHECK-NEXT: movi v1.8h, #1
+; CHECK-NEXT: ushr v0.4s, v0.4s, #15
+; CHECK-NEXT: movi v2.2d, #0x00ffff0000ffff
+; CHECK-NEXT: and v0.16b, v0.16b, v1.16b
+; CHECK-NEXT: mul v0.4s, v0.4s, v2.4s
+; CHECK-NEXT: ret
+ %b = lshr <4 x i32> %a, <i32 15, i32 15, i32 15, i32 15>
+ %c = and <4 x i32> %b, <i32 65537, i32 65537, i32 65537, i32 65537>
+ %d = mul nuw <4 x i32> %c, <i32 65535, i32 65535, i32 65535, i32 65535>
+ ret <4 x i32> %d
+}
+
+define <8 x i32> @v8i32(<8 x i32> %a) {
+; CHECK-LABEL: v8i32:
+; CHECK: // %bb.0:
+; CHECK-NEXT: movi v2.8h, #1
+; CHECK-NEXT: ushr v1.4s, v1.4s, #15
+; CHECK-NEXT: movi v3.2d, #0x00ffff0000ffff
+; CHECK-NEXT: ushr v0.4s, v0.4s, #15
+; CHECK-NEXT: and v1.16b, v1.16b, v2.16b
+; CHECK-NEXT: and v0.16b, v0.16b, v2.16b
+; CHECK-NEXT: mul v0.4s, v0.4s, v3.4s
+; CHECK-NEXT: mul v1.4s, v1.4s, v3.4s
+; CHECK-NEXT: ret
+ %b = lshr <8 x i32> %a, <i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15, i32 15>
+ %c = and <8 x i32> %b, <i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537, i32 65537>
+ %d = mul nuw <8 x i32> %c, <i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535, i32 65535>
+ ret <8 x i32> %d
+}
+
+define <4 x i16> @v4i16(<4 x i16> %a) {
+; CHECK-LABEL: v4i16:
+; CHECK: // %bb.0:
+; CHECK-NEXT: movi v1.8b, #1
+; CHECK-NEXT: movi d2, #0xff00ff00ff00ff
+; CHECK-NEXT: ushr v0.4h, v0.4h, #7
+; CHECK-NEXT: and v0.8b, v0.8b, v1.8b
+; CHECK-NEXT: mul v0.4h, v0.4h, v2.4h
+; CHECK-NEXT: ret
+ %b = lshr <4 x i16> %a, <i16 7, i16 7, i16 7, i16 7>
+ %c = and <4 x i16> %b, <i16 257, i16 257, i16 257, i16 257>
+ %d = mul nuw <4 x i16> %c, <i16 255, i16 255, i16 255, i16 255>
+ ret <4 x i16> %d
+}
+
+define <8 x i16> @v8i16(<8 x i16> %a) {
+; CHECK-LABEL: v8i16:
+; CHECK: // %bb.0:
+; CHECK-NEXT: movi v1.16b, #1
+; CHECK-NEXT: ushr v0.8h, v0.8h, #7
+; CHECK-NEXT: movi v2.2d, #0xff00ff00ff00ff
+; CHECK-NEXT: and v0.16b, v0.16b, v1.16b
+; CHECK-NEXT: mul v0.8h, v0.8h, v2.8h
+; CHECK-NEXT: ret
+ %b = lshr <8 x i16> %a, <i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7, i16 7>
+ %c = and <8 x i16> %b, <i16 257, i16 257, i16 257, i16 257, i16 257, i16 257, i16 257, i16 257>
+ %d = mul nuw <8 x i16> %c, <i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255, i16 255>
+ ret <8 x i16> %d
+}
+
+define <8 x i8> @v8i8(<8 x i8> %a) {
+; CHECK-LABEL: v8i8:
+; CHECK: // %bb.0:
+; CHECK-NEXT: movi v1.8b, #17
+; CHECK-NEXT: movi v2.8b, #15
+; CHECK-NEXT: ushr v0.8b, v0.8b, #3
+; CHECK-NEXT: and v0.8b, v0.8b, v1.8b
+; CHECK-NEXT: mul v0.8b, v0.8b, v2.8b
+; CHECK-NEXT: ret
+ %b = lshr <8 x i8> %a, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>
+ %c = and <8 x i8> %b, <i8 17, i8 17, i8 17, i8 17, i8 17, i8 17, i8 17, i8 17>
+ %d = mul nuw <8 x i8> %c, <i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15>
+ ret <8 x i8> %d
+}
+
+define <16 x i8> @v16i8(<16 x i8> %a) {
+; CHECK-LABEL: v16i8:
+; CHECK: // %bb.0:
+; CHECK-NEXT: movi v1.16b, #17
+; CHECK-NEXT: ushr v0.16b, v0.16b, #3
+; CHECK-NEXT: movi v2.16b, #15
+; CHECK-NEXT: and v0.16b, v0.16b, v1.16b
+; CHECK-NEXT: mul v0.16b, v0.16b, v2.16b
+; CHECK-NEXT: ret
+ %b = lshr <16 x i8> %a, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>
+ %c = and <16 x i8> %b, <i8 17, i8 17, i8 17, i8 17, i8 17, i8 17, i8 17, i8 17, i8 17, i8 17, i8 17, i8 17, i8 17, i8 17, i8 17, i8 17>
+ %d = mul nuw <16 x i8> %c, <i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15, i8 15>
+ ret <16 x i8> %d
+}
+
More information about the llvm-commits
mailing list