aboutsummaryrefslogtreecommitdiff
path: root/test/CodeGen/X86/shuffle-vs-trunc-256.ll
diff options
context:
space:
mode:
Diffstat (limited to 'test/CodeGen/X86/shuffle-vs-trunc-256.ll')
-rw-r--r--test/CodeGen/X86/shuffle-vs-trunc-256.ll184
1 files changed, 89 insertions, 95 deletions
diff --git a/test/CodeGen/X86/shuffle-vs-trunc-256.ll b/test/CodeGen/X86/shuffle-vs-trunc-256.ll
index 59a8aa47246c..5b46c73a3f65 100644
--- a/test/CodeGen/X86/shuffle-vs-trunc-256.ll
+++ b/test/CodeGen/X86/shuffle-vs-trunc-256.ll
@@ -1,6 +1,7 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefix=AVX --check-prefix=AVX1
-; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=AVX --check-prefix=AVX2
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=AVX --check-prefix=AVX2 --check-prefix=AVX2-SLOW
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2,+fast-variable-shuffle | FileCheck %s --check-prefix=AVX --check-prefix=AVX2 --check-prefix=AVX2-FAST
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefix=AVX512 --check-prefix=AVX512F
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl | FileCheck %s --check-prefix=AVX512 --check-prefix=AVX512VL
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512bw | FileCheck %s --check-prefix=AVX512 --check-prefix=AVX512BW
@@ -139,59 +140,17 @@ define void @shuffle_v16i16_to_v8i16(<16 x i16>* %L, <8 x i16>* %S) nounwind {
; AVX2-NEXT: vzeroupper
; AVX2-NEXT: retq
;
-; AVX512F-LABEL: shuffle_v16i16_to_v8i16:
-; AVX512F: # %bb.0:
-; AVX512F-NEXT: vmovdqa (%rdi), %ymm0
-; AVX512F-NEXT: vextracti128 $1, %ymm0, %xmm1
-; AVX512F-NEXT: vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
-; AVX512F-NEXT: vpshufb %xmm2, %xmm1, %xmm1
-; AVX512F-NEXT: vpshufb %xmm2, %xmm0, %xmm0
-; AVX512F-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
-; AVX512F-NEXT: vmovdqa %xmm0, (%rsi)
-; AVX512F-NEXT: vzeroupper
-; AVX512F-NEXT: retq
-;
-; AVX512VL-LABEL: shuffle_v16i16_to_v8i16:
-; AVX512VL: # %bb.0:
-; AVX512VL-NEXT: vmovdqa (%rdi), %ymm0
-; AVX512VL-NEXT: vextracti128 $1, %ymm0, %xmm1
-; AVX512VL-NEXT: vpshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; AVX512VL-NEXT: vpshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
-; AVX512VL-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; AVX512VL-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
-; AVX512VL-NEXT: vpshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
-; AVX512VL-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; AVX512VL-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
-; AVX512VL-NEXT: vmovdqa %xmm0, (%rsi)
-; AVX512VL-NEXT: vzeroupper
-; AVX512VL-NEXT: retq
-;
-; AVX512BW-LABEL: shuffle_v16i16_to_v8i16:
-; AVX512BW: # %bb.0:
-; AVX512BW-NEXT: vmovdqa (%rdi), %ymm0
-; AVX512BW-NEXT: vextracti128 $1, %ymm0, %xmm1
-; AVX512BW-NEXT: vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
-; AVX512BW-NEXT: vpshufb %xmm2, %xmm1, %xmm1
-; AVX512BW-NEXT: vpshufb %xmm2, %xmm0, %xmm0
-; AVX512BW-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
-; AVX512BW-NEXT: vmovdqa %xmm0, (%rsi)
-; AVX512BW-NEXT: vzeroupper
-; AVX512BW-NEXT: retq
-;
-; AVX512BWVL-LABEL: shuffle_v16i16_to_v8i16:
-; AVX512BWVL: # %bb.0:
-; AVX512BWVL-NEXT: vmovdqa (%rdi), %ymm0
-; AVX512BWVL-NEXT: vextracti128 $1, %ymm0, %xmm1
-; AVX512BWVL-NEXT: vpshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; AVX512BWVL-NEXT: vpshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
-; AVX512BWVL-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; AVX512BWVL-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
-; AVX512BWVL-NEXT: vpshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
-; AVX512BWVL-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; AVX512BWVL-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
-; AVX512BWVL-NEXT: vmovdqa %xmm0, (%rsi)
-; AVX512BWVL-NEXT: vzeroupper
-; AVX512BWVL-NEXT: retq
+; AVX512-LABEL: shuffle_v16i16_to_v8i16:
+; AVX512: # %bb.0:
+; AVX512-NEXT: vmovdqa (%rdi), %ymm0
+; AVX512-NEXT: vextracti128 $1, %ymm0, %xmm1
+; AVX512-NEXT: vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; AVX512-NEXT: vpshufb %xmm2, %xmm1, %xmm1
+; AVX512-NEXT: vpshufb %xmm2, %xmm0, %xmm0
+; AVX512-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
+; AVX512-NEXT: vmovdqa %xmm0, (%rsi)
+; AVX512-NEXT: vzeroupper
+; AVX512-NEXT: retq
%vec = load <16 x i16>, <16 x i16>* %L
%strided.vec = shufflevector <16 x i16> %vec, <16 x i16> undef, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14>
store <8 x i16> %strided.vec, <8 x i16>* %S
@@ -290,13 +249,21 @@ define void @trunc_v4i64_to_v4i32(<8 x i32>* %L, <4 x i32>* %S) nounwind {
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
-; AVX2-LABEL: trunc_v4i64_to_v4i32:
-; AVX2: # %bb.0:
-; AVX2-NEXT: vpermilps {{.*#+}} ymm0 = mem[0,2,2,3,4,6,6,7]
-; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,2,2,3]
-; AVX2-NEXT: vmovaps %xmm0, (%rsi)
-; AVX2-NEXT: vzeroupper
-; AVX2-NEXT: retq
+; AVX2-SLOW-LABEL: trunc_v4i64_to_v4i32:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vpermilps {{.*#+}} ymm0 = mem[0,2,2,3,4,6,6,7]
+; AVX2-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,2,2,3]
+; AVX2-SLOW-NEXT: vmovaps %xmm0, (%rsi)
+; AVX2-SLOW-NEXT: vzeroupper
+; AVX2-SLOW-NEXT: retq
+;
+; AVX2-FAST-LABEL: trunc_v4i64_to_v4i32:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm0 = [0,2,4,6,4,6,6,7]
+; AVX2-FAST-NEXT: vpermps (%rdi), %ymm0, %ymm0
+; AVX2-FAST-NEXT: vmovaps %xmm0, (%rsi)
+; AVX2-FAST-NEXT: vzeroupper
+; AVX2-FAST-NEXT: retq
;
; AVX512F-LABEL: trunc_v4i64_to_v4i32:
; AVX512F: # %bb.0:
@@ -399,12 +366,9 @@ define void @shuffle_v32i8_to_v8i8(<32 x i8>* %L, <8 x i8>* %S) nounwind {
; AVX512BWVL: # %bb.0:
; AVX512BWVL-NEXT: vmovdqa (%rdi), %ymm0
; AVX512BWVL-NEXT: vextracti128 $1, %ymm0, %xmm1
-; AVX512BWVL-NEXT: vpshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; AVX512BWVL-NEXT: vpshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,6,7]
-; AVX512BWVL-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; AVX512BWVL-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
-; AVX512BWVL-NEXT: vpshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,6,6,7]
-; AVX512BWVL-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; AVX512BWVL-NEXT: vmovdqa {{.*#+}} xmm2 = [0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; AVX512BWVL-NEXT: vpshufb %xmm2, %xmm1, %xmm1
+; AVX512BWVL-NEXT: vpshufb %xmm2, %xmm0, %xmm0
; AVX512BWVL-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX512BWVL-NEXT: vpmovwb %xmm0, (%rsi)
; AVX512BWVL-NEXT: vzeroupper
@@ -490,18 +454,30 @@ define void @shuffle_v16i16_to_v4i16(<16 x i16>* %L, <4 x i16>* %S) nounwind {
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
-; AVX2-LABEL: shuffle_v16i16_to_v4i16:
-; AVX2: # %bb.0:
-; AVX2-NEXT: vmovdqa (%rdi), %ymm0
-; AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
-; AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
-; AVX2-NEXT: vpshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
-; AVX2-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
-; AVX2-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
-; AVX2-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; AVX2-NEXT: vmovq %xmm0, (%rsi)
-; AVX2-NEXT: vzeroupper
-; AVX2-NEXT: retq
+; AVX2-SLOW-LABEL: shuffle_v16i16_to_v4i16:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vmovdqa (%rdi), %ymm0
+; AVX2-SLOW-NEXT: vextracti128 $1, %ymm0, %xmm1
+; AVX2-SLOW-NEXT: vpshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
+; AVX2-SLOW-NEXT: vpshuflw {{.*#+}} xmm1 = xmm1[0,2,2,3,4,5,6,7]
+; AVX2-SLOW-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
+; AVX2-SLOW-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[0,2,2,3,4,5,6,7]
+; AVX2-SLOW-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; AVX2-SLOW-NEXT: vmovq %xmm0, (%rsi)
+; AVX2-SLOW-NEXT: vzeroupper
+; AVX2-SLOW-NEXT: retq
+;
+; AVX2-FAST-LABEL: shuffle_v16i16_to_v4i16:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovdqa (%rdi), %ymm0
+; AVX2-FAST-NEXT: vextracti128 $1, %ymm0, %xmm1
+; AVX2-FAST-NEXT: vmovdqa {{.*#+}} xmm2 = [0,1,8,9,8,9,10,11,8,9,10,11,12,13,14,15]
+; AVX2-FAST-NEXT: vpshufb %xmm2, %xmm1, %xmm1
+; AVX2-FAST-NEXT: vpshufb %xmm2, %xmm0, %xmm0
+; AVX2-FAST-NEXT: vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; AVX2-FAST-NEXT: vmovq %xmm0, (%rsi)
+; AVX2-FAST-NEXT: vzeroupper
+; AVX2-FAST-NEXT: retq
;
; AVX512F-LABEL: shuffle_v16i16_to_v4i16:
; AVX512F: # %bb.0:
@@ -563,14 +539,23 @@ define void @trunc_v4i64_to_v4i16(<16 x i16>* %L, <4 x i16>* %S) nounwind {
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
-; AVX2-LABEL: trunc_v4i64_to_v4i16:
-; AVX2: # %bb.0:
-; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = mem[0,2,2,3,4,6,6,7]
-; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,2,3]
-; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
-; AVX2-NEXT: vmovq %xmm0, (%rsi)
-; AVX2-NEXT: vzeroupper
-; AVX2-NEXT: retq
+; AVX2-SLOW-LABEL: trunc_v4i64_to_v4i16:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vpshufd {{.*#+}} ymm0 = mem[0,2,2,3,4,6,6,7]
+; AVX2-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,2,3]
+; AVX2-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; AVX2-SLOW-NEXT: vmovq %xmm0, (%rsi)
+; AVX2-SLOW-NEXT: vzeroupper
+; AVX2-SLOW-NEXT: retq
+;
+; AVX2-FAST-LABEL: trunc_v4i64_to_v4i16:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovdqa {{.*#+}} ymm0 = [0,2,4,6,4,6,6,7]
+; AVX2-FAST-NEXT: vpermd (%rdi), %ymm0, %ymm0
+; AVX2-FAST-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,1,4,5,8,9,12,13,8,9,12,13,12,13,14,15]
+; AVX2-FAST-NEXT: vmovq %xmm0, (%rsi)
+; AVX2-FAST-NEXT: vzeroupper
+; AVX2-FAST-NEXT: retq
;
; AVX512F-LABEL: trunc_v4i64_to_v4i16:
; AVX512F: # %bb.0:
@@ -693,14 +678,23 @@ define void @trunc_v4i64_to_v4i8(<32 x i8>* %L, <4 x i8>* %S) nounwind {
; AVX1-NEXT: vzeroupper
; AVX1-NEXT: retq
;
-; AVX2-LABEL: trunc_v4i64_to_v4i8:
-; AVX2: # %bb.0:
-; AVX2-NEXT: vpshufd {{.*#+}} ymm0 = mem[0,2,2,3,4,6,6,7]
-; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,2,3]
-; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u]
-; AVX2-NEXT: vmovd %xmm0, (%rsi)
-; AVX2-NEXT: vzeroupper
-; AVX2-NEXT: retq
+; AVX2-SLOW-LABEL: trunc_v4i64_to_v4i8:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vpshufd {{.*#+}} ymm0 = mem[0,2,2,3,4,6,6,7]
+; AVX2-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,2,2,3]
+; AVX2-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u]
+; AVX2-SLOW-NEXT: vmovd %xmm0, (%rsi)
+; AVX2-SLOW-NEXT: vzeroupper
+; AVX2-SLOW-NEXT: retq
+;
+; AVX2-FAST-LABEL: trunc_v4i64_to_v4i8:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovdqa {{.*#+}} ymm0 = [0,2,4,6,4,6,6,7]
+; AVX2-FAST-NEXT: vpermd (%rdi), %ymm0, %ymm0
+; AVX2-FAST-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u]
+; AVX2-FAST-NEXT: vmovd %xmm0, (%rsi)
+; AVX2-FAST-NEXT: vzeroupper
+; AVX2-FAST-NEXT: retq
;
; AVX512F-LABEL: trunc_v4i64_to_v4i8:
; AVX512F: # %bb.0:
generated by cgit v1.2.3 (git 2.25.1) at 2024-06-19 23:23:26 +0000