e6bf48d110
As of rev ea222be0d, LLVMs assembler will actually try to honour the "fill value" part of p2align directives. X86 printed these as 0x90, which isn't actually what it wanted: we want multi-byte nops for .text padding. Compiling via a textual assembly file produces single-byte nop padding since ea222be0d but the built-in assembler will produce multi-byte nops. This divergent behaviour is undesirable. To fix: don't set the byte padding field for x86, which allows the assembler to pick multi-byte nops. Test that we get the same multi-byte padding when compiled via textual assembly or directly to object file. Added same-align-bytes-with-llasm-llobj.ll to that effect, updated numerous other tests to not contain check-lines for the explicit padding.
347 lines
14 KiB
LLVM
347 lines
14 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=CHECK,SSE,SSE2
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.2 | FileCheck %s --check-prefixes=CHECK,SSE,SSE4
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=CHECK,AVX,AVX1
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,AVX,AVX2
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl,+avx512f | FileCheck %s --check-prefixes=CHECK,AVX,AVX512
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl,+avx512bw | FileCheck %s --check-prefixes=CHECK,AVX,AVX512
|
|
|
|
; trunc(concat(x,y)) -> pack
|
|
|
|
define <8 x i16> @trunc_concat_packssdw_128(<4 x i32> %a0, <4 x i32> %a1) nounwind {
|
|
; SSE-LABEL: trunc_concat_packssdw_128:
|
|
; SSE: # %bb.0:
|
|
; SSE-NEXT: psrad $17, %xmm0
|
|
; SSE-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
|
|
; SSE-NEXT: packssdw %xmm1, %xmm0
|
|
; SSE-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: trunc_concat_packssdw_128:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpsrad $17, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1
|
|
; AVX1-NEXT: vpackssdw %xmm1, %xmm0, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: trunc_concat_packssdw_128:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpsrad $17, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpbroadcastd {{.*#+}} xmm2 = [15,15,15,15]
|
|
; AVX2-NEXT: vpand %xmm2, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpackssdw %xmm1, %xmm0, %xmm0
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512-LABEL: trunc_concat_packssdw_128:
|
|
; AVX512: # %bb.0:
|
|
; AVX512-NEXT: vpsrad $17, %xmm0, %xmm0
|
|
; AVX512-NEXT: vpandd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm1, %xmm1
|
|
; AVX512-NEXT: vpackssdw %xmm1, %xmm0, %xmm0
|
|
; AVX512-NEXT: retq
|
|
%1 = ashr <4 x i32> %a0, <i32 17, i32 17, i32 17, i32 17>
|
|
%2 = and <4 x i32> %a1, <i32 15, i32 15, i32 15, i32 15>
|
|
%3 = shufflevector <4 x i32> %1, <4 x i32> %2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
|
|
%4 = trunc <8 x i32> %3 to <8 x i16>
|
|
ret <8 x i16> %4
|
|
}
|
|
|
|
define <8 x i16> @trunc_concat_packusdw_128(<4 x i32> %a0, <4 x i32> %a1) nounwind {
|
|
; SSE2-LABEL: trunc_concat_packusdw_128:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: psrld $17, %xmm0
|
|
; SSE2-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
|
|
; SSE2-NEXT: packssdw %xmm1, %xmm0
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE4-LABEL: trunc_concat_packusdw_128:
|
|
; SSE4: # %bb.0:
|
|
; SSE4-NEXT: psrld $17, %xmm0
|
|
; SSE4-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
|
|
; SSE4-NEXT: packusdw %xmm1, %xmm0
|
|
; SSE4-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: trunc_concat_packusdw_128:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpsrld $17, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1
|
|
; AVX1-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: trunc_concat_packusdw_128:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpsrld $17, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpbroadcastd {{.*#+}} xmm2 = [15,15,15,15]
|
|
; AVX2-NEXT: vpand %xmm2, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512-LABEL: trunc_concat_packusdw_128:
|
|
; AVX512: # %bb.0:
|
|
; AVX512-NEXT: vpsrld $17, %xmm0, %xmm0
|
|
; AVX512-NEXT: vpandd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm1, %xmm1
|
|
; AVX512-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
|
; AVX512-NEXT: retq
|
|
%1 = lshr <4 x i32> %a0, <i32 17, i32 17, i32 17, i32 17>
|
|
%2 = and <4 x i32> %a1, <i32 15, i32 15, i32 15, i32 15>
|
|
%3 = shufflevector <4 x i32> %1, <4 x i32> %2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
|
|
%4 = trunc <8 x i32> %3 to <8 x i16>
|
|
ret <8 x i16> %4
|
|
}
|
|
|
|
define <16 x i8> @trunc_concat_packsswb_128(<8 x i16> %a0, <8 x i16> %a1) nounwind {
|
|
; SSE-LABEL: trunc_concat_packsswb_128:
|
|
; SSE: # %bb.0:
|
|
; SSE-NEXT: psraw $15, %xmm0
|
|
; SSE-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
|
|
; SSE-NEXT: packsswb %xmm1, %xmm0
|
|
; SSE-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: trunc_concat_packsswb_128:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpsraw $15, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1
|
|
; AVX1-NEXT: vpacksswb %xmm1, %xmm0, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: trunc_concat_packsswb_128:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpsraw $15, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1
|
|
; AVX2-NEXT: vpacksswb %xmm1, %xmm0, %xmm0
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512-LABEL: trunc_concat_packsswb_128:
|
|
; AVX512: # %bb.0:
|
|
; AVX512-NEXT: vpsraw $15, %xmm0, %xmm0
|
|
; AVX512-NEXT: vpandd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm1, %xmm1
|
|
; AVX512-NEXT: vpacksswb %xmm1, %xmm0, %xmm0
|
|
; AVX512-NEXT: retq
|
|
%1 = ashr <8 x i16> %a0, <i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15>
|
|
%2 = and <8 x i16> %a1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
|
|
%3 = shufflevector <8 x i16> %1, <8 x i16> %2, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
|
|
%4 = trunc <16 x i16> %3 to <16 x i8>
|
|
ret <16 x i8> %4
|
|
}
|
|
|
|
define <16 x i8> @trunc_concat_packuswb_128(<8 x i16> %a0, <8 x i16> %a1) nounwind {
|
|
; SSE-LABEL: trunc_concat_packuswb_128:
|
|
; SSE: # %bb.0:
|
|
; SSE-NEXT: psrlw $15, %xmm0
|
|
; SSE-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
|
|
; SSE-NEXT: packuswb %xmm1, %xmm0
|
|
; SSE-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: trunc_concat_packuswb_128:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1
|
|
; AVX1-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: trunc_concat_packuswb_128:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1
|
|
; AVX2-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512-LABEL: trunc_concat_packuswb_128:
|
|
; AVX512: # %bb.0:
|
|
; AVX512-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX512-NEXT: vpandd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm1, %xmm1
|
|
; AVX512-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
|
|
; AVX512-NEXT: retq
|
|
%1 = lshr <8 x i16> %a0, <i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15>
|
|
%2 = and <8 x i16> %a1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
|
|
%3 = shufflevector <8 x i16> %1, <8 x i16> %2, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
|
|
%4 = trunc <16 x i16> %3 to <16 x i8>
|
|
ret <16 x i8> %4
|
|
}
|
|
|
|
; concat(trunc(x),trunc(y)) -> pack
|
|
|
|
define <8 x i16> @concat_trunc_packssdw_128(<4 x i32> %a0, <4 x i32> %a1) nounwind {
|
|
; SSE2-LABEL: concat_trunc_packssdw_128:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: psrad $17, %xmm0
|
|
; SSE2-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
|
|
; SSE2-NEXT: packssdw %xmm0, %xmm0
|
|
; SSE2-NEXT: packuswb %xmm1, %xmm1
|
|
; SSE2-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE4-LABEL: concat_trunc_packssdw_128:
|
|
; SSE4: # %bb.0:
|
|
; SSE4-NEXT: psrad $17, %xmm0
|
|
; SSE4-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
|
|
; SSE4-NEXT: packssdw %xmm1, %xmm0
|
|
; SSE4-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: concat_trunc_packssdw_128:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpsrad $17, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1
|
|
; AVX1-NEXT: vpackssdw %xmm1, %xmm0, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: concat_trunc_packssdw_128:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpsrad $17, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpbroadcastd {{.*#+}} xmm2 = [15,15,15,15]
|
|
; AVX2-NEXT: vpand %xmm2, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpackssdw %xmm1, %xmm0, %xmm0
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512-LABEL: concat_trunc_packssdw_128:
|
|
; AVX512: # %bb.0:
|
|
; AVX512-NEXT: vpsrad $17, %xmm0, %xmm0
|
|
; AVX512-NEXT: vpandd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm1, %xmm1
|
|
; AVX512-NEXT: vpackssdw %xmm1, %xmm0, %xmm0
|
|
; AVX512-NEXT: retq
|
|
%1 = ashr <4 x i32> %a0, <i32 17, i32 17, i32 17, i32 17>
|
|
%2 = and <4 x i32> %a1, <i32 15, i32 15, i32 15, i32 15>
|
|
%3 = trunc <4 x i32> %1 to <4 x i16>
|
|
%4 = trunc <4 x i32> %2 to <4 x i16>
|
|
%5 = shufflevector <4 x i16> %3, <4 x i16> %4, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
|
|
ret <8 x i16> %5
|
|
}
|
|
|
|
define <8 x i16> @concat_trunc_packusdw_128(<4 x i32> %a0, <4 x i32> %a1) nounwind {
|
|
; SSE2-LABEL: concat_trunc_packusdw_128:
|
|
; SSE2: # %bb.0:
|
|
; SSE2-NEXT: psrld $17, %xmm0
|
|
; SSE2-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
|
|
; SSE2-NEXT: packssdw %xmm0, %xmm0
|
|
; SSE2-NEXT: packuswb %xmm1, %xmm1
|
|
; SSE2-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
|
|
; SSE2-NEXT: retq
|
|
;
|
|
; SSE4-LABEL: concat_trunc_packusdw_128:
|
|
; SSE4: # %bb.0:
|
|
; SSE4-NEXT: psrld $17, %xmm0
|
|
; SSE4-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
|
|
; SSE4-NEXT: packusdw %xmm1, %xmm0
|
|
; SSE4-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: concat_trunc_packusdw_128:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpsrld $17, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1
|
|
; AVX1-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: concat_trunc_packusdw_128:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpsrld $17, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpbroadcastd {{.*#+}} xmm2 = [15,15,15,15]
|
|
; AVX2-NEXT: vpand %xmm2, %xmm1, %xmm1
|
|
; AVX2-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512-LABEL: concat_trunc_packusdw_128:
|
|
; AVX512: # %bb.0:
|
|
; AVX512-NEXT: vpsrld $17, %xmm0, %xmm0
|
|
; AVX512-NEXT: vpandd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm1, %xmm1
|
|
; AVX512-NEXT: vpackusdw %xmm1, %xmm0, %xmm0
|
|
; AVX512-NEXT: retq
|
|
%1 = lshr <4 x i32> %a0, <i32 17, i32 17, i32 17, i32 17>
|
|
%2 = and <4 x i32> %a1, <i32 15, i32 15, i32 15, i32 15>
|
|
%3 = trunc <4 x i32> %1 to <4 x i16>
|
|
%4 = trunc <4 x i32> %2 to <4 x i16>
|
|
%5 = shufflevector <4 x i16> %3, <4 x i16> %4, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
|
|
ret <8 x i16> %5
|
|
}
|
|
|
|
define <16 x i8> @concat_trunc_packsswb_128(<8 x i16> %a0, <8 x i16> %a1) nounwind {
|
|
; SSE-LABEL: concat_trunc_packsswb_128:
|
|
; SSE: # %bb.0:
|
|
; SSE-NEXT: psraw $15, %xmm0
|
|
; SSE-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
|
|
; SSE-NEXT: packsswb %xmm1, %xmm0
|
|
; SSE-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: concat_trunc_packsswb_128:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpsraw $15, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1
|
|
; AVX1-NEXT: vpacksswb %xmm1, %xmm0, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: concat_trunc_packsswb_128:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpsraw $15, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1
|
|
; AVX2-NEXT: vpacksswb %xmm1, %xmm0, %xmm0
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512-LABEL: concat_trunc_packsswb_128:
|
|
; AVX512: # %bb.0:
|
|
; AVX512-NEXT: vpsraw $15, %xmm0, %xmm0
|
|
; AVX512-NEXT: vpandd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm1, %xmm1
|
|
; AVX512-NEXT: vpacksswb %xmm1, %xmm0, %xmm0
|
|
; AVX512-NEXT: retq
|
|
%1 = ashr <8 x i16> %a0, <i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15>
|
|
%2 = and <8 x i16> %a1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
|
|
%3 = trunc <8 x i16> %1 to <8 x i8>
|
|
%4 = trunc <8 x i16> %2 to <8 x i8>
|
|
%5 = shufflevector <8 x i8> %3, <8 x i8> %4, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
|
|
ret <16 x i8> %5
|
|
}
|
|
|
|
define <16 x i8> @concat_trunc_packuswb_128(<8 x i16> %a0, <8 x i16> %a1) nounwind {
|
|
; SSE-LABEL: concat_trunc_packuswb_128:
|
|
; SSE: # %bb.0:
|
|
; SSE-NEXT: psrlw $15, %xmm0
|
|
; SSE-NEXT: pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
|
|
; SSE-NEXT: packuswb %xmm1, %xmm0
|
|
; SSE-NEXT: retq
|
|
;
|
|
; AVX1-LABEL: concat_trunc_packuswb_128:
|
|
; AVX1: # %bb.0:
|
|
; AVX1-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX1-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1
|
|
; AVX1-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
|
|
; AVX1-NEXT: retq
|
|
;
|
|
; AVX2-LABEL: concat_trunc_packuswb_128:
|
|
; AVX2: # %bb.0:
|
|
; AVX2-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX2-NEXT: vpand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1, %xmm1
|
|
; AVX2-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
|
|
; AVX2-NEXT: retq
|
|
;
|
|
; AVX512-LABEL: concat_trunc_packuswb_128:
|
|
; AVX512: # %bb.0:
|
|
; AVX512-NEXT: vpsrlw $15, %xmm0, %xmm0
|
|
; AVX512-NEXT: vpandd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm1, %xmm1
|
|
; AVX512-NEXT: vpackuswb %xmm1, %xmm0, %xmm0
|
|
; AVX512-NEXT: retq
|
|
%1 = lshr <8 x i16> %a0, <i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15>
|
|
%2 = and <8 x i16> %a1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
|
|
%3 = trunc <8 x i16> %1 to <8 x i8>
|
|
%4 = trunc <8 x i16> %2 to <8 x i8>
|
|
%5 = shufflevector <8 x i8> %3, <8 x i8> %4, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
|
|
ret <16 x i8> %5
|
|
}
|
|
|
|
; Fuzz test - don't pack a v1i32 comparison result.
|
|
define void @autogen_SD10339(<1 x i32> %I49) {
|
|
; CHECK-LABEL: autogen_SD10339:
|
|
; CHECK: # %bb.0: # %BB
|
|
; CHECK-NEXT: .p2align 4
|
|
; CHECK-NEXT: .LBB8_1: # %CF
|
|
; CHECK-NEXT: # =>This Inner Loop Header: Depth=1
|
|
; CHECK-NEXT: movw $1, 0
|
|
; CHECK-NEXT: jmp .LBB8_1
|
|
BB:
|
|
%Cmp53 = icmp uge <1 x i32> %I49, zeroinitializer
|
|
br label %CF
|
|
|
|
CF: ; preds = %CF, %BB
|
|
%ZE166 = zext <1 x i1> %Cmp53 to <1 x i16>
|
|
store <1 x i16> %ZE166, ptr null, align 2
|
|
br label %CF
|
|
}
|
|
|
|
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
|
|
; AVX: {{.*}}
|