caio/clang-p2996
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
b20da5117fb66e7f9bcb56cb14af86dcafe2ee9a
clang-p2996/llvm/test/CodeGen/X86/horizontal-reduce-umin.ll
Matt Arsenault 4a36e96c3f RegAllocGreedy: Account for reserved registers in num regs heuristic 
This simple heuristic uses the estimated live range length combined
with the number of registers in the class to switch which heuristic to
use. This was taking the raw number of registers in the class, even
though not all of them may be available. AMDGPU heavily relies on
dynamically reserved numbers of registers based on user attributes to
satisfy occupancy constraints, so the raw number is highly misleading.

There are still a few problems here. In the original testcase that
made me notice this, the live range size is incorrect after the
scheduler rearranges instructions, since the instructions don't have
the original InstrDist offsets. Additionally, I think it would be more
appropriate to use the number of disjointly allocatable registers in
the class. For the AMDGPU register tuples, there are a large number of
registers in each tuple class, but only a small fraction can actually
be allocated at the same time since they all overlap with each
other. It seems we do not have a query that corresponds to the number
of independently allocatable registers. Relatedly, I'm still debugging
some allocation failures where overlapping tuples seem to not be
handled correctly.

The test changes are mostly noise. There are a handful of x86 tests
that look like regressions with an additional spill, and a handful
that now avoid a spill. The worst looking regression is likely
test/Thumb2/mve-vld4.ll which introduces a few additional
spills. test/CodeGen/AMDGPU/soft-clause-exceeds-register-budget.ll
shows a massive improvement by completely eliminating a large number
of spills inside a loop.
2021-09-14 21:00:29 -04:00

2212 lines
103 KiB
LLVM
Raw Blame History

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-apple-darwin -mattr=+sse2 | FileCheck %s --check-prefix=X86-SSE2
; RUN: llc < %s -mtriple=i686-apple-darwin -mattr=+sse4.2 | FileCheck %s --check-prefix=X86-SSE42
; RUN: llc < %s -mtriple=i686-apple-darwin -mattr=+avx | FileCheck %s --check-prefixes=X86-AVX,X86-AVX1
; RUN: llc < %s -mtriple=i686-apple-darwin -mattr=+avx2 | FileCheck %s --check-prefixes=X86-AVX,X86-AVX2
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+sse2 | FileCheck %s --check-prefix=X64-SSE2
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+sse4.2 | FileCheck %s --check-prefix=X64-SSE42
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx | FileCheck %s --check-prefixes=X64-AVX,X64-AVX1
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx2 | FileCheck %s --check-prefixes=X64-AVX,X64-AVX2
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx512f,+avx512bw,+avx512dq,+avx512vl | FileCheck %s --check-prefixes=X64-AVX,X64-AVX512
;
; 128-bit Vectors
;
define i64 @test_reduce_v2i64(<2 x i64> %a0) {
; X86-SSE2-LABEL: test_reduce_v2i64:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [2147483648,2147483648,2147483648,2147483648]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm3
; X86-SSE2-NEXT: pxor %xmm2, %xmm3
; X86-SSE2-NEXT: pxor %xmm1, %xmm2
; X86-SSE2-NEXT: movdqa %xmm2, %xmm4
; X86-SSE2-NEXT: pcmpgtd %xmm3, %xmm4
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
; X86-SSE2-NEXT: pcmpeqd %xmm3, %xmm2
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
; X86-SSE2-NEXT: pand %xmm5, %xmm2
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm4[1,1,3,3]
; X86-SSE2-NEXT: por %xmm2, %xmm3
; X86-SSE2-NEXT: pand %xmm3, %xmm0
; X86-SSE2-NEXT: pandn %xmm1, %xmm3
; X86-SSE2-NEXT: por %xmm0, %xmm3
; X86-SSE2-NEXT: movd %xmm3, %eax
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm3[1,1,1,1]
; X86-SSE2-NEXT: movd %xmm0, %edx
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v2i64:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: movdqa %xmm0, %xmm1
; X86-SSE42-NEXT: pshufd {{.*#+}} xmm2 = xmm0[2,3,2,3]
; X86-SSE42-NEXT: movdqa {{.*#+}} xmm0 = [0,2147483648,0,2147483648]
; X86-SSE42-NEXT: movdqa %xmm1, %xmm3
; X86-SSE42-NEXT: pxor %xmm0, %xmm3
; X86-SSE42-NEXT: pxor %xmm2, %xmm0
; X86-SSE42-NEXT: pcmpgtq %xmm3, %xmm0
; X86-SSE42-NEXT: blendvpd %xmm0, %xmm1, %xmm2
; X86-SSE42-NEXT: movd %xmm2, %eax
; X86-SSE42-NEXT: pextrd $1, %xmm2, %edx
; X86-SSE42-NEXT: retl
;
; X86-AVX1-LABEL: test_reduce_v2i64:
; X86-AVX1: ## %bb.0:
; X86-AVX1-NEXT: vpermilps {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-AVX1-NEXT: vmovddup {{.*#+}} xmm2 = [9223372036854775808,9223372036854775808]
; X86-AVX1-NEXT: ## xmm2 = mem[0,0]
; X86-AVX1-NEXT: vxorps %xmm2, %xmm0, %xmm3
; X86-AVX1-NEXT: vxorps %xmm2, %xmm1, %xmm2
; X86-AVX1-NEXT: vpcmpgtq %xmm3, %xmm2, %xmm2
; X86-AVX1-NEXT: vblendvpd %xmm2, %xmm0, %xmm1, %xmm0
; X86-AVX1-NEXT: vmovd %xmm0, %eax
; X86-AVX1-NEXT: vpextrd $1, %xmm0, %edx
; X86-AVX1-NEXT: retl
;
; X86-AVX2-LABEL: test_reduce_v2i64:
; X86-AVX2: ## %bb.0:
; X86-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-AVX2-NEXT: vpbroadcastq {{.*#+}} xmm2 = [9223372036854775808,9223372036854775808]
; X86-AVX2-NEXT: vpxor %xmm2, %xmm0, %xmm3
; X86-AVX2-NEXT: vpxor %xmm2, %xmm1, %xmm2
; X86-AVX2-NEXT: vpcmpgtq %xmm3, %xmm2, %xmm2
; X86-AVX2-NEXT: vblendvpd %xmm2, %xmm0, %xmm1, %xmm0
; X86-AVX2-NEXT: vmovd %xmm0, %eax
; X86-AVX2-NEXT: vpextrd $1, %xmm0, %edx
; X86-AVX2-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v2i64:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [9223372039002259456,9223372039002259456]
; X64-SSE2-NEXT: movdqa %xmm0, %xmm3
; X64-SSE2-NEXT: pxor %xmm2, %xmm3
; X64-SSE2-NEXT: pxor %xmm1, %xmm2
; X64-SSE2-NEXT: movdqa %xmm2, %xmm4
; X64-SSE2-NEXT: pcmpgtd %xmm3, %xmm4
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
; X64-SSE2-NEXT: pcmpeqd %xmm3, %xmm2
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
; X64-SSE2-NEXT: pand %xmm5, %xmm2
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm4[1,1,3,3]
; X64-SSE2-NEXT: por %xmm2, %xmm3
; X64-SSE2-NEXT: pand %xmm3, %xmm0
; X64-SSE2-NEXT: pandn %xmm1, %xmm3
; X64-SSE2-NEXT: por %xmm0, %xmm3
; X64-SSE2-NEXT: movq %xmm3, %rax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v2i64:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: movdqa %xmm0, %xmm1
; X64-SSE42-NEXT: pshufd {{.*#+}} xmm2 = xmm0[2,3,2,3]
; X64-SSE42-NEXT: movdqa {{.*#+}} xmm0 = [9223372036854775808,9223372036854775808]
; X64-SSE42-NEXT: movdqa %xmm1, %xmm3
; X64-SSE42-NEXT: pxor %xmm0, %xmm3
; X64-SSE42-NEXT: pxor %xmm2, %xmm0
; X64-SSE42-NEXT: pcmpgtq %xmm3, %xmm0
; X64-SSE42-NEXT: blendvpd %xmm0, %xmm1, %xmm2
; X64-SSE42-NEXT: movq %xmm2, %rax
; X64-SSE42-NEXT: retq
;
; X64-AVX1-LABEL: test_reduce_v2i64:
; X64-AVX1: ## %bb.0:
; X64-AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [9223372036854775808,9223372036854775808]
; X64-AVX1-NEXT: vpxor %xmm2, %xmm0, %xmm3
; X64-AVX1-NEXT: vpxor %xmm2, %xmm1, %xmm2
; X64-AVX1-NEXT: vpcmpgtq %xmm3, %xmm2, %xmm2
; X64-AVX1-NEXT: vblendvpd %xmm2, %xmm0, %xmm1, %xmm0
; X64-AVX1-NEXT: vmovq %xmm0, %rax
; X64-AVX1-NEXT: retq
;
; X64-AVX2-LABEL: test_reduce_v2i64:
; X64-AVX2: ## %bb.0:
; X64-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-AVX2-NEXT: vmovdqa {{.*#+}} xmm2 = [9223372036854775808,9223372036854775808]
; X64-AVX2-NEXT: vpxor %xmm2, %xmm0, %xmm3
; X64-AVX2-NEXT: vpxor %xmm2, %xmm1, %xmm2
; X64-AVX2-NEXT: vpcmpgtq %xmm3, %xmm2, %xmm2
; X64-AVX2-NEXT: vblendvpd %xmm2, %xmm0, %xmm1, %xmm0
; X64-AVX2-NEXT: vmovq %xmm0, %rax
; X64-AVX2-NEXT: retq
;
; X64-AVX512-LABEL: test_reduce_v2i64:
; X64-AVX512: ## %bb.0:
; X64-AVX512-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-AVX512-NEXT: vpminuq %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vmovq %xmm0, %rax
; X64-AVX512-NEXT: retq
%1 = shufflevector <2 x i64> %a0, <2 x i64> undef, <2 x i32> <i32 1, i32 undef>
%2 = icmp ult <2 x i64> %a0, %1
%3 = select <2 x i1> %2, <2 x i64> %a0, <2 x i64> %1
%4 = extractelement <2 x i64> %3, i32 0
ret i64 %4
}
define i32 @test_reduce_v4i32(<4 x i32> %a0) {
; X86-SSE2-LABEL: test_reduce_v4i32:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm0[2,3,2,3]
; X86-SSE2-NEXT: movdqa {{.*#+}} xmm1 = [2147483648,2147483648,2147483648,2147483648]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm4
; X86-SSE2-NEXT: pxor %xmm1, %xmm4
; X86-SSE2-NEXT: movdqa %xmm3, %xmm2
; X86-SSE2-NEXT: pxor %xmm1, %xmm2
; X86-SSE2-NEXT: pcmpgtd %xmm4, %xmm2
; X86-SSE2-NEXT: pand %xmm2, %xmm0
; X86-SSE2-NEXT: pandn %xmm3, %xmm2
; X86-SSE2-NEXT: por %xmm0, %xmm2
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm2[1,1,1,1]
; X86-SSE2-NEXT: movdqa %xmm2, %xmm3
; X86-SSE2-NEXT: pxor %xmm1, %xmm3
; X86-SSE2-NEXT: pxor %xmm0, %xmm1
; X86-SSE2-NEXT: pcmpgtd %xmm3, %xmm1
; X86-SSE2-NEXT: pand %xmm1, %xmm2
; X86-SSE2-NEXT: pandn %xmm0, %xmm1
; X86-SSE2-NEXT: por %xmm2, %xmm1
; X86-SSE2-NEXT: movd %xmm1, %eax
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v4i32:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-SSE42-NEXT: pminud %xmm0, %xmm1
; X86-SSE42-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
; X86-SSE42-NEXT: pminud %xmm1, %xmm0
; X86-SSE42-NEXT: movd %xmm0, %eax
; X86-SSE42-NEXT: retl
;
; X86-AVX-LABEL: test_reduce_v4i32:
; X86-AVX: ## %bb.0:
; X86-AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-AVX-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X86-AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X86-AVX-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X86-AVX-NEXT: vmovd %xmm0, %eax
; X86-AVX-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v4i32:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [2147483648,2147483648,2147483648,2147483648]
; X64-SSE2-NEXT: movdqa %xmm0, %xmm3
; X64-SSE2-NEXT: pxor %xmm2, %xmm3
; X64-SSE2-NEXT: movdqa %xmm1, %xmm4
; X64-SSE2-NEXT: pxor %xmm2, %xmm4
; X64-SSE2-NEXT: pcmpgtd %xmm3, %xmm4
; X64-SSE2-NEXT: pand %xmm4, %xmm0
; X64-SSE2-NEXT: pandn %xmm1, %xmm4
; X64-SSE2-NEXT: por %xmm0, %xmm4
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm4[1,1,1,1]
; X64-SSE2-NEXT: movdqa %xmm4, %xmm1
; X64-SSE2-NEXT: pxor %xmm2, %xmm1
; X64-SSE2-NEXT: pxor %xmm0, %xmm2
; X64-SSE2-NEXT: pcmpgtd %xmm1, %xmm2
; X64-SSE2-NEXT: pand %xmm2, %xmm4
; X64-SSE2-NEXT: pandn %xmm0, %xmm2
; X64-SSE2-NEXT: por %xmm4, %xmm2
; X64-SSE2-NEXT: movd %xmm2, %eax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v4i32:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-SSE42-NEXT: pminud %xmm0, %xmm1
; X64-SSE42-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
; X64-SSE42-NEXT: pminud %xmm1, %xmm0
; X64-SSE42-NEXT: movd %xmm0, %eax
; X64-SSE42-NEXT: retq
;
; X64-AVX-LABEL: test_reduce_v4i32:
; X64-AVX: ## %bb.0:
; X64-AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-AVX-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X64-AVX-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX-NEXT: vmovd %xmm0, %eax
; X64-AVX-NEXT: retq
%1 = shufflevector <4 x i32> %a0, <4 x i32> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
%2 = icmp ult <4 x i32> %a0, %1
%3 = select <4 x i1> %2, <4 x i32> %a0, <4 x i32> %1
%4 = shufflevector <4 x i32> %3, <4 x i32> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
%5 = icmp ult <4 x i32> %3, %4
%6 = select <4 x i1> %5, <4 x i32> %3, <4 x i32> %4
%7 = extractelement <4 x i32> %6, i32 0
ret i32 %7
}
define i16 @test_reduce_v8i16(<8 x i16> %a0) {
; X86-SSE2-LABEL: test_reduce_v8i16:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
; X86-SSE2-NEXT: psrld $16, %xmm1
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: movd %xmm0, %eax
; X86-SSE2-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v8i16:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: phminposuw %xmm0, %xmm0
; X86-SSE42-NEXT: movd %xmm0, %eax
; X86-SSE42-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-SSE42-NEXT: retl
;
; X86-AVX-LABEL: test_reduce_v8i16:
; X86-AVX: ## %bb.0:
; X86-AVX-NEXT: vphminposuw %xmm0, %xmm0
; X86-AVX-NEXT: vmovd %xmm0, %eax
; X86-AVX-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-AVX-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v8i16:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: movdqa %xmm0, %xmm1
; X64-SSE2-NEXT: psrld $16, %xmm1
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: movd %xmm0, %eax
; X64-SSE2-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v8i16:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: phminposuw %xmm0, %xmm0
; X64-SSE42-NEXT: movd %xmm0, %eax
; X64-SSE42-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-SSE42-NEXT: retq
;
; X64-AVX-LABEL: test_reduce_v8i16:
; X64-AVX: ## %bb.0:
; X64-AVX-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX-NEXT: vmovd %xmm0, %eax
; X64-AVX-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-AVX-NEXT: retq
%1 = shufflevector <8 x i16> %a0, <8 x i16> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
%2 = icmp ult <8 x i16> %a0, %1
%3 = select <8 x i1> %2, <8 x i16> %a0, <8 x i16> %1
%4 = shufflevector <8 x i16> %3, <8 x i16> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%5 = icmp ult <8 x i16> %3, %4
%6 = select <8 x i1> %5, <8 x i16> %3, <8 x i16> %4
%7 = shufflevector <8 x i16> %6, <8 x i16> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%8 = icmp ult <8 x i16> %6, %7
%9 = select <8 x i1> %8, <8 x i16> %6, <8 x i16> %7
%10 = extractelement <8 x i16> %9, i32 0
ret i16 %10
}
define i8 @test_reduce_v16i8(<16 x i8> %a0) {
; X86-SSE2-LABEL: test_reduce_v16i8:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-SSE2-NEXT: pminub %xmm0, %xmm1
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
; X86-SSE2-NEXT: pminub %xmm1, %xmm0
; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
; X86-SSE2-NEXT: psrld $16, %xmm1
; X86-SSE2-NEXT: pminub %xmm0, %xmm1
; X86-SSE2-NEXT: movdqa %xmm1, %xmm0
; X86-SSE2-NEXT: psrlw $8, %xmm0
; X86-SSE2-NEXT: pminub %xmm1, %xmm0
; X86-SSE2-NEXT: movd %xmm0, %eax
; X86-SSE2-NEXT: ## kill: def $al killed $al killed $eax
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v16i8:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: movdqa %xmm0, %xmm1
; X86-SSE42-NEXT: psrlw $8, %xmm1
; X86-SSE42-NEXT: pminub %xmm0, %xmm1
; X86-SSE42-NEXT: phminposuw %xmm1, %xmm0
; X86-SSE42-NEXT: movd %xmm0, %eax
; X86-SSE42-NEXT: ## kill: def $al killed $al killed $eax
; X86-SSE42-NEXT: retl
;
; X86-AVX-LABEL: test_reduce_v16i8:
; X86-AVX: ## %bb.0:
; X86-AVX-NEXT: vpsrlw $8, %xmm0, %xmm1
; X86-AVX-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X86-AVX-NEXT: vphminposuw %xmm0, %xmm0
; X86-AVX-NEXT: vmovd %xmm0, %eax
; X86-AVX-NEXT: ## kill: def $al killed $al killed $eax
; X86-AVX-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v16i8:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-SSE2-NEXT: pminub %xmm0, %xmm1
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
; X64-SSE2-NEXT: pminub %xmm1, %xmm0
; X64-SSE2-NEXT: movdqa %xmm0, %xmm1
; X64-SSE2-NEXT: psrld $16, %xmm1
; X64-SSE2-NEXT: pminub %xmm0, %xmm1
; X64-SSE2-NEXT: movdqa %xmm1, %xmm0
; X64-SSE2-NEXT: psrlw $8, %xmm0
; X64-SSE2-NEXT: pminub %xmm1, %xmm0
; X64-SSE2-NEXT: movd %xmm0, %eax
; X64-SSE2-NEXT: ## kill: def $al killed $al killed $eax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v16i8:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: movdqa %xmm0, %xmm1
; X64-SSE42-NEXT: psrlw $8, %xmm1
; X64-SSE42-NEXT: pminub %xmm0, %xmm1
; X64-SSE42-NEXT: phminposuw %xmm1, %xmm0
; X64-SSE42-NEXT: movd %xmm0, %eax
; X64-SSE42-NEXT: ## kill: def $al killed $al killed $eax
; X64-SSE42-NEXT: retq
;
; X64-AVX-LABEL: test_reduce_v16i8:
; X64-AVX: ## %bb.0:
; X64-AVX-NEXT: vpsrlw $8, %xmm0, %xmm1
; X64-AVX-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X64-AVX-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX-NEXT: vmovd %xmm0, %eax
; X64-AVX-NEXT: ## kill: def $al killed $al killed $eax
; X64-AVX-NEXT: retq
%1 = shufflevector <16 x i8> %a0, <16 x i8> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%2 = icmp ult <16 x i8> %a0, %1
%3 = select <16 x i1> %2, <16 x i8> %a0, <16 x i8> %1
%4 = shufflevector <16 x i8> %3, <16 x i8> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%5 = icmp ult <16 x i8> %3, %4
%6 = select <16 x i1> %5, <16 x i8> %3, <16 x i8> %4
%7 = shufflevector <16 x i8> %6, <16 x i8> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%8 = icmp ult <16 x i8> %6, %7
%9 = select <16 x i1> %8, <16 x i8> %6, <16 x i8> %7
%10 = shufflevector <16 x i8> %9, <16 x i8> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%11 = icmp ult <16 x i8> %9, %10
%12 = select <16 x i1> %11, <16 x i8> %9, <16 x i8> %10
%13 = extractelement <16 x i8> %12, i32 0
ret i8 %13
}
;
; 256-bit Vectors
;
define i64 @test_reduce_v4i64(<4 x i64> %a0) {
; X86-SSE2-LABEL: test_reduce_v4i64:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [2147483648,2147483648,2147483648,2147483648]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm3
; X86-SSE2-NEXT: pxor %xmm2, %xmm3
; X86-SSE2-NEXT: movdqa %xmm1, %xmm4
; X86-SSE2-NEXT: pxor %xmm2, %xmm4
; X86-SSE2-NEXT: movdqa %xmm4, %xmm5
; X86-SSE2-NEXT: pcmpgtd %xmm3, %xmm5
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm5[0,0,2,2]
; X86-SSE2-NEXT: pcmpeqd %xmm3, %xmm4
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3]
; X86-SSE2-NEXT: pand %xmm6, %xmm4
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm5[1,1,3,3]
; X86-SSE2-NEXT: por %xmm4, %xmm3
; X86-SSE2-NEXT: pand %xmm3, %xmm0
; X86-SSE2-NEXT: pandn %xmm1, %xmm3
; X86-SSE2-NEXT: por %xmm0, %xmm3
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm3[2,3,2,3]
; X86-SSE2-NEXT: movdqa %xmm3, %xmm1
; X86-SSE2-NEXT: pxor %xmm2, %xmm1
; X86-SSE2-NEXT: pxor %xmm0, %xmm2
; X86-SSE2-NEXT: movdqa %xmm2, %xmm4
; X86-SSE2-NEXT: pcmpgtd %xmm1, %xmm4
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
; X86-SSE2-NEXT: pcmpeqd %xmm1, %xmm2
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm2[1,1,3,3]
; X86-SSE2-NEXT: pand %xmm5, %xmm1
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm4[1,1,3,3]
; X86-SSE2-NEXT: por %xmm1, %xmm2
; X86-SSE2-NEXT: pand %xmm2, %xmm3
; X86-SSE2-NEXT: pandn %xmm0, %xmm2
; X86-SSE2-NEXT: por %xmm3, %xmm2
; X86-SSE2-NEXT: movd %xmm2, %eax
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm2[1,1,1,1]
; X86-SSE2-NEXT: movd %xmm0, %edx
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v4i64:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: movdqa %xmm0, %xmm3
; X86-SSE42-NEXT: movdqa {{.*#+}} xmm2 = [0,2147483648,0,2147483648]
; X86-SSE42-NEXT: movdqa %xmm0, %xmm4
; X86-SSE42-NEXT: pxor %xmm2, %xmm4
; X86-SSE42-NEXT: movdqa %xmm1, %xmm0
; X86-SSE42-NEXT: pxor %xmm2, %xmm0
; X86-SSE42-NEXT: pcmpgtq %xmm4, %xmm0
; X86-SSE42-NEXT: blendvpd %xmm0, %xmm3, %xmm1
; X86-SSE42-NEXT: pshufd {{.*#+}} xmm3 = xmm1[2,3,2,3]
; X86-SSE42-NEXT: movdqa %xmm1, %xmm0
; X86-SSE42-NEXT: pxor %xmm2, %xmm0
; X86-SSE42-NEXT: pxor %xmm3, %xmm2
; X86-SSE42-NEXT: pcmpgtq %xmm0, %xmm2
; X86-SSE42-NEXT: movdqa %xmm2, %xmm0
; X86-SSE42-NEXT: blendvpd %xmm0, %xmm1, %xmm3
; X86-SSE42-NEXT: movd %xmm3, %eax
; X86-SSE42-NEXT: pextrd $1, %xmm3, %edx
; X86-SSE42-NEXT: retl
;
; X86-AVX1-LABEL: test_reduce_v4i64:
; X86-AVX1: ## %bb.0:
; X86-AVX1-NEXT: vmovddup {{.*#+}} xmm1 = [9223372036854775808,9223372036854775808]
; X86-AVX1-NEXT: ## xmm1 = mem[0,0]
; X86-AVX1-NEXT: vxorps %xmm1, %xmm0, %xmm2
; X86-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
; X86-AVX1-NEXT: vxorps %xmm1, %xmm3, %xmm4
; X86-AVX1-NEXT: vpcmpgtq %xmm2, %xmm4, %xmm2
; X86-AVX1-NEXT: vblendvpd %xmm2, %xmm0, %xmm3, %xmm0
; X86-AVX1-NEXT: vpermilps {{.*#+}} xmm2 = xmm0[2,3,2,3]
; X86-AVX1-NEXT: vxorpd %xmm1, %xmm0, %xmm3
; X86-AVX1-NEXT: vxorpd %xmm1, %xmm2, %xmm1
; X86-AVX1-NEXT: vpcmpgtq %xmm3, %xmm1, %xmm1
; X86-AVX1-NEXT: vblendvpd %xmm1, %xmm0, %xmm2, %xmm0
; X86-AVX1-NEXT: vmovd %xmm0, %eax
; X86-AVX1-NEXT: vpextrd $1, %xmm0, %edx
; X86-AVX1-NEXT: vzeroupper
; X86-AVX1-NEXT: retl
;
; X86-AVX2-LABEL: test_reduce_v4i64:
; X86-AVX2: ## %bb.0:
; X86-AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; X86-AVX2-NEXT: vpbroadcastq {{.*#+}} xmm2 = [9223372036854775808,9223372036854775808]
; X86-AVX2-NEXT: vpxor %xmm2, %xmm0, %xmm3
; X86-AVX2-NEXT: vpxor %xmm2, %xmm1, %xmm4
; X86-AVX2-NEXT: vpcmpgtq %xmm3, %xmm4, %xmm3
; X86-AVX2-NEXT: vblendvpd %xmm3, %xmm0, %xmm1, %xmm0
; X86-AVX2-NEXT: vpermilps {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-AVX2-NEXT: vxorpd %xmm2, %xmm0, %xmm3
; X86-AVX2-NEXT: vxorpd %xmm2, %xmm1, %xmm2
; X86-AVX2-NEXT: vpcmpgtq %xmm3, %xmm2, %xmm2
; X86-AVX2-NEXT: vblendvpd %xmm2, %xmm0, %xmm1, %xmm0
; X86-AVX2-NEXT: vmovd %xmm0, %eax
; X86-AVX2-NEXT: vpextrd $1, %xmm0, %edx
; X86-AVX2-NEXT: vzeroupper
; X86-AVX2-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v4i64:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [9223372039002259456,9223372039002259456]
; X64-SSE2-NEXT: movdqa %xmm0, %xmm3
; X64-SSE2-NEXT: pxor %xmm2, %xmm3
; X64-SSE2-NEXT: movdqa %xmm1, %xmm4
; X64-SSE2-NEXT: pxor %xmm2, %xmm4
; X64-SSE2-NEXT: movdqa %xmm4, %xmm5
; X64-SSE2-NEXT: pcmpgtd %xmm3, %xmm5
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm5[0,0,2,2]
; X64-SSE2-NEXT: pcmpeqd %xmm3, %xmm4
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm4[1,1,3,3]
; X64-SSE2-NEXT: pand %xmm6, %xmm3
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm5[1,1,3,3]
; X64-SSE2-NEXT: por %xmm3, %xmm4
; X64-SSE2-NEXT: pand %xmm4, %xmm0
; X64-SSE2-NEXT: pandn %xmm1, %xmm4
; X64-SSE2-NEXT: por %xmm0, %xmm4
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm4[2,3,2,3]
; X64-SSE2-NEXT: movdqa %xmm4, %xmm1
; X64-SSE2-NEXT: pxor %xmm2, %xmm1
; X64-SSE2-NEXT: pxor %xmm0, %xmm2
; X64-SSE2-NEXT: movdqa %xmm2, %xmm3
; X64-SSE2-NEXT: pcmpgtd %xmm1, %xmm3
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm3[0,0,2,2]
; X64-SSE2-NEXT: pcmpeqd %xmm1, %xmm2
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm2[1,1,3,3]
; X64-SSE2-NEXT: pand %xmm5, %xmm1
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm3[1,1,3,3]
; X64-SSE2-NEXT: por %xmm1, %xmm2
; X64-SSE2-NEXT: pand %xmm2, %xmm4
; X64-SSE2-NEXT: pandn %xmm0, %xmm2
; X64-SSE2-NEXT: por %xmm4, %xmm2
; X64-SSE2-NEXT: movq %xmm2, %rax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v4i64:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: movdqa %xmm0, %xmm2
; X64-SSE42-NEXT: movdqa {{.*#+}} xmm3 = [9223372036854775808,9223372036854775808]
; X64-SSE42-NEXT: movdqa %xmm0, %xmm4
; X64-SSE42-NEXT: pxor %xmm3, %xmm4
; X64-SSE42-NEXT: movdqa %xmm1, %xmm0
; X64-SSE42-NEXT: pxor %xmm3, %xmm0
; X64-SSE42-NEXT: pcmpgtq %xmm4, %xmm0
; X64-SSE42-NEXT: blendvpd %xmm0, %xmm2, %xmm1
; X64-SSE42-NEXT: pshufd {{.*#+}} xmm2 = xmm1[2,3,2,3]
; X64-SSE42-NEXT: movdqa %xmm1, %xmm0
; X64-SSE42-NEXT: pxor %xmm3, %xmm0
; X64-SSE42-NEXT: pxor %xmm2, %xmm3
; X64-SSE42-NEXT: pcmpgtq %xmm0, %xmm3
; X64-SSE42-NEXT: movdqa %xmm3, %xmm0
; X64-SSE42-NEXT: blendvpd %xmm0, %xmm1, %xmm2
; X64-SSE42-NEXT: movq %xmm2, %rax
; X64-SSE42-NEXT: retq
;
; X64-AVX1-LABEL: test_reduce_v4i64:
; X64-AVX1: ## %bb.0:
; X64-AVX1-NEXT: vmovdqa {{.*#+}} xmm1 = [9223372036854775808,9223372036854775808]
; X64-AVX1-NEXT: vpxor %xmm1, %xmm0, %xmm2
; X64-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
; X64-AVX1-NEXT: vpxor %xmm1, %xmm3, %xmm4
; X64-AVX1-NEXT: vpcmpgtq %xmm2, %xmm4, %xmm2
; X64-AVX1-NEXT: vblendvpd %xmm2, %xmm0, %xmm3, %xmm0
; X64-AVX1-NEXT: vpermilps {{.*#+}} xmm2 = xmm0[2,3,2,3]
; X64-AVX1-NEXT: vxorpd %xmm1, %xmm0, %xmm3
; X64-AVX1-NEXT: vxorpd %xmm1, %xmm2, %xmm1
; X64-AVX1-NEXT: vpcmpgtq %xmm3, %xmm1, %xmm1
; X64-AVX1-NEXT: vblendvpd %xmm1, %xmm0, %xmm2, %xmm0
; X64-AVX1-NEXT: vmovq %xmm0, %rax
; X64-AVX1-NEXT: vzeroupper
; X64-AVX1-NEXT: retq
;
; X64-AVX2-LABEL: test_reduce_v4i64:
; X64-AVX2: ## %bb.0:
; X64-AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; X64-AVX2-NEXT: vpbroadcastq {{.*#+}} ymm2 = [9223372036854775808,9223372036854775808,9223372036854775808,9223372036854775808]
; X64-AVX2-NEXT: vpxor %xmm2, %xmm0, %xmm3
; X64-AVX2-NEXT: vpxor %xmm2, %xmm1, %xmm4
; X64-AVX2-NEXT: vpcmpgtq %xmm3, %xmm4, %xmm3
; X64-AVX2-NEXT: vblendvpd %xmm3, %xmm0, %xmm1, %xmm0
; X64-AVX2-NEXT: vpermilps {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-AVX2-NEXT: vxorpd %xmm2, %xmm0, %xmm3
; X64-AVX2-NEXT: vxorpd %xmm2, %xmm1, %xmm2
; X64-AVX2-NEXT: vpcmpgtq %xmm3, %xmm2, %xmm2
; X64-AVX2-NEXT: vblendvpd %xmm2, %xmm0, %xmm1, %xmm0
; X64-AVX2-NEXT: vmovq %xmm0, %rax
; X64-AVX2-NEXT: vzeroupper
; X64-AVX2-NEXT: retq
;
; X64-AVX512-LABEL: test_reduce_v4i64:
; X64-AVX512: ## %bb.0:
; X64-AVX512-NEXT: vextracti128 $1, %ymm0, %xmm1
; X64-AVX512-NEXT: vpminuq %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-AVX512-NEXT: vpminuq %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vmovq %xmm0, %rax
; X64-AVX512-NEXT: vzeroupper
; X64-AVX512-NEXT: retq
%1 = shufflevector <4 x i64> %a0, <4 x i64> undef, <4 x i32> <i32 2, i32 3, i32 undef, i32 undef>
%2 = icmp ult <4 x i64> %a0, %1
%3 = select <4 x i1> %2, <4 x i64> %a0, <4 x i64> %1
%4 = shufflevector <4 x i64> %3, <4 x i64> undef, <4 x i32> <i32 1, i32 undef, i32 undef, i32 undef>
%5 = icmp ult <4 x i64> %3, %4
%6 = select <4 x i1> %5, <4 x i64> %3, <4 x i64> %4
%7 = extractelement <4 x i64> %6, i32 0
ret i64 %7
}
define i32 @test_reduce_v8i32(<8 x i32> %a0) {
; X86-SSE2-LABEL: test_reduce_v8i32:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [2147483648,2147483648,2147483648,2147483648]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm4
; X86-SSE2-NEXT: pxor %xmm2, %xmm4
; X86-SSE2-NEXT: movdqa %xmm1, %xmm3
; X86-SSE2-NEXT: pxor %xmm2, %xmm3
; X86-SSE2-NEXT: pcmpgtd %xmm4, %xmm3
; X86-SSE2-NEXT: pand %xmm3, %xmm0
; X86-SSE2-NEXT: pandn %xmm1, %xmm3
; X86-SSE2-NEXT: por %xmm0, %xmm3
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm3[2,3,2,3]
; X86-SSE2-NEXT: movdqa %xmm3, %xmm4
; X86-SSE2-NEXT: pxor %xmm2, %xmm4
; X86-SSE2-NEXT: movdqa %xmm1, %xmm0
; X86-SSE2-NEXT: pxor %xmm2, %xmm0
; X86-SSE2-NEXT: pcmpgtd %xmm4, %xmm0
; X86-SSE2-NEXT: pand %xmm0, %xmm3
; X86-SSE2-NEXT: pandn %xmm1, %xmm0
; X86-SSE2-NEXT: por %xmm3, %xmm0
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm3
; X86-SSE2-NEXT: pxor %xmm2, %xmm3
; X86-SSE2-NEXT: pxor %xmm1, %xmm2
; X86-SSE2-NEXT: pcmpgtd %xmm3, %xmm2
; X86-SSE2-NEXT: pand %xmm2, %xmm0
; X86-SSE2-NEXT: pandn %xmm1, %xmm2
; X86-SSE2-NEXT: por %xmm0, %xmm2
; X86-SSE2-NEXT: movd %xmm2, %eax
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v8i32:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: pminud %xmm1, %xmm0
; X86-SSE42-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-SSE42-NEXT: pminud %xmm0, %xmm1
; X86-SSE42-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
; X86-SSE42-NEXT: pminud %xmm1, %xmm0
; X86-SSE42-NEXT: movd %xmm0, %eax
; X86-SSE42-NEXT: retl
;
; X86-AVX1-LABEL: test_reduce_v8i32:
; X86-AVX1: ## %bb.0:
; X86-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
; X86-AVX1-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X86-AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-AVX1-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X86-AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X86-AVX1-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X86-AVX1-NEXT: vmovd %xmm0, %eax
; X86-AVX1-NEXT: vzeroupper
; X86-AVX1-NEXT: retl
;
; X86-AVX2-LABEL: test_reduce_v8i32:
; X86-AVX2: ## %bb.0:
; X86-AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; X86-AVX2-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X86-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-AVX2-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X86-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X86-AVX2-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X86-AVX2-NEXT: vmovd %xmm0, %eax
; X86-AVX2-NEXT: vzeroupper
; X86-AVX2-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v8i32:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: movdqa {{.*#+}} xmm2 = [2147483648,2147483648,2147483648,2147483648]
; X64-SSE2-NEXT: movdqa %xmm0, %xmm3
; X64-SSE2-NEXT: pxor %xmm2, %xmm3
; X64-SSE2-NEXT: movdqa %xmm1, %xmm4
; X64-SSE2-NEXT: pxor %xmm2, %xmm4
; X64-SSE2-NEXT: pcmpgtd %xmm3, %xmm4
; X64-SSE2-NEXT: pand %xmm4, %xmm0
; X64-SSE2-NEXT: pandn %xmm1, %xmm4
; X64-SSE2-NEXT: por %xmm0, %xmm4
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm4[2,3,2,3]
; X64-SSE2-NEXT: movdqa %xmm4, %xmm1
; X64-SSE2-NEXT: pxor %xmm2, %xmm1
; X64-SSE2-NEXT: movdqa %xmm0, %xmm3
; X64-SSE2-NEXT: pxor %xmm2, %xmm3
; X64-SSE2-NEXT: pcmpgtd %xmm1, %xmm3
; X64-SSE2-NEXT: pand %xmm3, %xmm4
; X64-SSE2-NEXT: pandn %xmm0, %xmm3
; X64-SSE2-NEXT: por %xmm4, %xmm3
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm3[1,1,1,1]
; X64-SSE2-NEXT: movdqa %xmm3, %xmm1
; X64-SSE2-NEXT: pxor %xmm2, %xmm1
; X64-SSE2-NEXT: pxor %xmm0, %xmm2
; X64-SSE2-NEXT: pcmpgtd %xmm1, %xmm2
; X64-SSE2-NEXT: pand %xmm2, %xmm3
; X64-SSE2-NEXT: pandn %xmm0, %xmm2
; X64-SSE2-NEXT: por %xmm3, %xmm2
; X64-SSE2-NEXT: movd %xmm2, %eax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v8i32:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: pminud %xmm1, %xmm0
; X64-SSE42-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-SSE42-NEXT: pminud %xmm0, %xmm1
; X64-SSE42-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
; X64-SSE42-NEXT: pminud %xmm1, %xmm0
; X64-SSE42-NEXT: movd %xmm0, %eax
; X64-SSE42-NEXT: retq
;
; X64-AVX1-LABEL: test_reduce_v8i32:
; X64-AVX1: ## %bb.0:
; X64-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
; X64-AVX1-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-AVX1-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X64-AVX1-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX1-NEXT: vmovd %xmm0, %eax
; X64-AVX1-NEXT: vzeroupper
; X64-AVX1-NEXT: retq
;
; X64-AVX2-LABEL: test_reduce_v8i32:
; X64-AVX2: ## %bb.0:
; X64-AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; X64-AVX2-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-AVX2-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X64-AVX2-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT: vmovd %xmm0, %eax
; X64-AVX2-NEXT: vzeroupper
; X64-AVX2-NEXT: retq
;
; X64-AVX512-LABEL: test_reduce_v8i32:
; X64-AVX512: ## %bb.0:
; X64-AVX512-NEXT: vextracti128 $1, %ymm0, %xmm1
; X64-AVX512-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-AVX512-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X64-AVX512-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vmovd %xmm0, %eax
; X64-AVX512-NEXT: vzeroupper
; X64-AVX512-NEXT: retq
%1 = shufflevector <8 x i32> %a0, <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
%2 = icmp ult <8 x i32> %a0, %1
%3 = select <8 x i1> %2, <8 x i32> %a0, <8 x i32> %1
%4 = shufflevector <8 x i32> %3, <8 x i32> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%5 = icmp ult <8 x i32> %3, %4
%6 = select <8 x i1> %5, <8 x i32> %3, <8 x i32> %4
%7 = shufflevector <8 x i32> %6, <8 x i32> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%8 = icmp ult <8 x i32> %6, %7
%9 = select <8 x i1> %8, <8 x i32> %6, <8 x i32> %7
%10 = extractelement <8 x i32> %9, i32 0
ret i32 %10
}
define i16 @test_reduce_v16i16(<16 x i16> %a0) {
; X86-SSE2-LABEL: test_reduce_v16i16:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
; X86-SSE2-NEXT: psrld $16, %xmm1
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: movd %xmm0, %eax
; X86-SSE2-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v16i16:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: pminuw %xmm1, %xmm0
; X86-SSE42-NEXT: phminposuw %xmm0, %xmm0
; X86-SSE42-NEXT: movd %xmm0, %eax
; X86-SSE42-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-SSE42-NEXT: retl
;
; X86-AVX1-LABEL: test_reduce_v16i16:
; X86-AVX1: ## %bb.0:
; X86-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
; X86-AVX1-NEXT: vpminuw %xmm1, %xmm0, %xmm0
; X86-AVX1-NEXT: vphminposuw %xmm0, %xmm0
; X86-AVX1-NEXT: vmovd %xmm0, %eax
; X86-AVX1-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-AVX1-NEXT: vzeroupper
; X86-AVX1-NEXT: retl
;
; X86-AVX2-LABEL: test_reduce_v16i16:
; X86-AVX2: ## %bb.0:
; X86-AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; X86-AVX2-NEXT: vpminuw %xmm1, %xmm0, %xmm0
; X86-AVX2-NEXT: vphminposuw %xmm0, %xmm0
; X86-AVX2-NEXT: vmovd %xmm0, %eax
; X86-AVX2-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-AVX2-NEXT: vzeroupper
; X86-AVX2-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v16i16:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: movdqa %xmm0, %xmm1
; X64-SSE2-NEXT: psrld $16, %xmm1
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: movd %xmm0, %eax
; X64-SSE2-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v16i16:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: pminuw %xmm1, %xmm0
; X64-SSE42-NEXT: phminposuw %xmm0, %xmm0
; X64-SSE42-NEXT: movd %xmm0, %eax
; X64-SSE42-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-SSE42-NEXT: retq
;
; X64-AVX1-LABEL: test_reduce_v16i16:
; X64-AVX1: ## %bb.0:
; X64-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
; X64-AVX1-NEXT: vpminuw %xmm1, %xmm0, %xmm0
; X64-AVX1-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX1-NEXT: vmovd %xmm0, %eax
; X64-AVX1-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-AVX1-NEXT: vzeroupper
; X64-AVX1-NEXT: retq
;
; X64-AVX2-LABEL: test_reduce_v16i16:
; X64-AVX2: ## %bb.0:
; X64-AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; X64-AVX2-NEXT: vpminuw %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX2-NEXT: vmovd %xmm0, %eax
; X64-AVX2-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-AVX2-NEXT: vzeroupper
; X64-AVX2-NEXT: retq
;
; X64-AVX512-LABEL: test_reduce_v16i16:
; X64-AVX512: ## %bb.0:
; X64-AVX512-NEXT: vextracti128 $1, %ymm0, %xmm1
; X64-AVX512-NEXT: vpminuw %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX512-NEXT: vmovd %xmm0, %eax
; X64-AVX512-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-AVX512-NEXT: vzeroupper
; X64-AVX512-NEXT: retq
%1 = shufflevector <16 x i16> %a0, <16 x i16> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%2 = icmp ult <16 x i16> %a0, %1
%3 = select <16 x i1> %2, <16 x i16> %a0, <16 x i16> %1
%4 = shufflevector <16 x i16> %3, <16 x i16> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%5 = icmp ult <16 x i16> %3, %4
%6 = select <16 x i1> %5, <16 x i16> %3, <16 x i16> %4
%7 = shufflevector <16 x i16> %6, <16 x i16> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%8 = icmp ult <16 x i16> %6, %7
%9 = select <16 x i1> %8, <16 x i16> %6, <16 x i16> %7
%10 = shufflevector <16 x i16> %9, <16 x i16> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%11 = icmp ult <16 x i16> %9, %10
%12 = select <16 x i1> %11, <16 x i16> %9, <16 x i16> %10
%13 = extractelement <16 x i16> %12, i32 0
ret i16 %13
}
define i8 @test_reduce_v32i8(<32 x i8> %a0) {
; X86-SSE2-LABEL: test_reduce_v32i8:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: pminub %xmm1, %xmm0
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-SSE2-NEXT: pminub %xmm0, %xmm1
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
; X86-SSE2-NEXT: pminub %xmm1, %xmm0
; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
; X86-SSE2-NEXT: psrld $16, %xmm1
; X86-SSE2-NEXT: pminub %xmm0, %xmm1
; X86-SSE2-NEXT: movdqa %xmm1, %xmm0
; X86-SSE2-NEXT: psrlw $8, %xmm0
; X86-SSE2-NEXT: pminub %xmm1, %xmm0
; X86-SSE2-NEXT: movd %xmm0, %eax
; X86-SSE2-NEXT: ## kill: def $al killed $al killed $eax
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v32i8:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: pminub %xmm1, %xmm0
; X86-SSE42-NEXT: movdqa %xmm0, %xmm1
; X86-SSE42-NEXT: psrlw $8, %xmm1
; X86-SSE42-NEXT: pminub %xmm0, %xmm1
; X86-SSE42-NEXT: phminposuw %xmm1, %xmm0
; X86-SSE42-NEXT: movd %xmm0, %eax
; X86-SSE42-NEXT: ## kill: def $al killed $al killed $eax
; X86-SSE42-NEXT: retl
;
; X86-AVX1-LABEL: test_reduce_v32i8:
; X86-AVX1: ## %bb.0:
; X86-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
; X86-AVX1-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X86-AVX1-NEXT: vpsrlw $8, %xmm0, %xmm1
; X86-AVX1-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X86-AVX1-NEXT: vphminposuw %xmm0, %xmm0
; X86-AVX1-NEXT: vmovd %xmm0, %eax
; X86-AVX1-NEXT: ## kill: def $al killed $al killed $eax
; X86-AVX1-NEXT: vzeroupper
; X86-AVX1-NEXT: retl
;
; X86-AVX2-LABEL: test_reduce_v32i8:
; X86-AVX2: ## %bb.0:
; X86-AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; X86-AVX2-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X86-AVX2-NEXT: vpsrlw $8, %xmm0, %xmm1
; X86-AVX2-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X86-AVX2-NEXT: vphminposuw %xmm0, %xmm0
; X86-AVX2-NEXT: vmovd %xmm0, %eax
; X86-AVX2-NEXT: ## kill: def $al killed $al killed $eax
; X86-AVX2-NEXT: vzeroupper
; X86-AVX2-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v32i8:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: pminub %xmm1, %xmm0
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-SSE2-NEXT: pminub %xmm0, %xmm1
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
; X64-SSE2-NEXT: pminub %xmm1, %xmm0
; X64-SSE2-NEXT: movdqa %xmm0, %xmm1
; X64-SSE2-NEXT: psrld $16, %xmm1
; X64-SSE2-NEXT: pminub %xmm0, %xmm1
; X64-SSE2-NEXT: movdqa %xmm1, %xmm0
; X64-SSE2-NEXT: psrlw $8, %xmm0
; X64-SSE2-NEXT: pminub %xmm1, %xmm0
; X64-SSE2-NEXT: movd %xmm0, %eax
; X64-SSE2-NEXT: ## kill: def $al killed $al killed $eax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v32i8:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: pminub %xmm1, %xmm0
; X64-SSE42-NEXT: movdqa %xmm0, %xmm1
; X64-SSE42-NEXT: psrlw $8, %xmm1
; X64-SSE42-NEXT: pminub %xmm0, %xmm1
; X64-SSE42-NEXT: phminposuw %xmm1, %xmm0
; X64-SSE42-NEXT: movd %xmm0, %eax
; X64-SSE42-NEXT: ## kill: def $al killed $al killed $eax
; X64-SSE42-NEXT: retq
;
; X64-AVX1-LABEL: test_reduce_v32i8:
; X64-AVX1: ## %bb.0:
; X64-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm1
; X64-AVX1-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X64-AVX1-NEXT: vpsrlw $8, %xmm0, %xmm1
; X64-AVX1-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X64-AVX1-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX1-NEXT: vmovd %xmm0, %eax
; X64-AVX1-NEXT: ## kill: def $al killed $al killed $eax
; X64-AVX1-NEXT: vzeroupper
; X64-AVX1-NEXT: retq
;
; X64-AVX2-LABEL: test_reduce_v32i8:
; X64-AVX2: ## %bb.0:
; X64-AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; X64-AVX2-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT: vpsrlw $8, %xmm0, %xmm1
; X64-AVX2-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX2-NEXT: vmovd %xmm0, %eax
; X64-AVX2-NEXT: ## kill: def $al killed $al killed $eax
; X64-AVX2-NEXT: vzeroupper
; X64-AVX2-NEXT: retq
;
; X64-AVX512-LABEL: test_reduce_v32i8:
; X64-AVX512: ## %bb.0:
; X64-AVX512-NEXT: vextracti128 $1, %ymm0, %xmm1
; X64-AVX512-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vpsrlw $8, %xmm0, %xmm1
; X64-AVX512-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX512-NEXT: vmovd %xmm0, %eax
; X64-AVX512-NEXT: ## kill: def $al killed $al killed $eax
; X64-AVX512-NEXT: vzeroupper
; X64-AVX512-NEXT: retq
%1 = shufflevector <32 x i8> %a0, <32 x i8> undef, <32 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%2 = icmp ult <32 x i8> %a0, %1
%3 = select <32 x i1> %2, <32 x i8> %a0, <32 x i8> %1
%4 = shufflevector <32 x i8> %3, <32 x i8> undef, <32 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%5 = icmp ult <32 x i8> %3, %4
%6 = select <32 x i1> %5, <32 x i8> %3, <32 x i8> %4
%7 = shufflevector <32 x i8> %6, <32 x i8> undef, <32 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%8 = icmp ult <32 x i8> %6, %7
%9 = select <32 x i1> %8, <32 x i8> %6, <32 x i8> %7
%10 = shufflevector <32 x i8> %9, <32 x i8> undef, <32 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%11 = icmp ult <32 x i8> %9, %10
%12 = select <32 x i1> %11, <32 x i8> %9, <32 x i8> %10
%13 = shufflevector <32 x i8> %12, <32 x i8> undef, <32 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%14 = icmp ult <32 x i8> %12, %13
%15 = select <32 x i1> %14, <32 x i8> %12, <32 x i8> %13
%16 = extractelement <32 x i8> %15, i32 0
ret i8 %16
}
;
; 512-bit Vectors
;
define i64 @test_reduce_v8i64(<8 x i64> %a0) {
; X86-SSE2-LABEL: test_reduce_v8i64:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: movdqa {{.*#+}} xmm4 = [2147483648,2147483648,2147483648,2147483648]
; X86-SSE2-NEXT: movdqa %xmm1, %xmm5
; X86-SSE2-NEXT: pxor %xmm4, %xmm5
; X86-SSE2-NEXT: movdqa %xmm3, %xmm6
; X86-SSE2-NEXT: pxor %xmm4, %xmm6
; X86-SSE2-NEXT: movdqa %xmm6, %xmm7
; X86-SSE2-NEXT: pcmpgtd %xmm5, %xmm7
; X86-SSE2-NEXT: pcmpeqd %xmm5, %xmm6
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm7[0,0,2,2]
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3]
; X86-SSE2-NEXT: pand %xmm5, %xmm6
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm7[1,1,3,3]
; X86-SSE2-NEXT: por %xmm6, %xmm5
; X86-SSE2-NEXT: pand %xmm5, %xmm1
; X86-SSE2-NEXT: pandn %xmm3, %xmm5
; X86-SSE2-NEXT: por %xmm1, %xmm5
; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
; X86-SSE2-NEXT: pxor %xmm4, %xmm1
; X86-SSE2-NEXT: movdqa %xmm2, %xmm3
; X86-SSE2-NEXT: pxor %xmm4, %xmm3
; X86-SSE2-NEXT: movdqa %xmm3, %xmm6
; X86-SSE2-NEXT: pcmpgtd %xmm1, %xmm6
; X86-SSE2-NEXT: pcmpeqd %xmm1, %xmm3
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm6[0,0,2,2]
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
; X86-SSE2-NEXT: pand %xmm1, %xmm3
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm6[1,1,3,3]
; X86-SSE2-NEXT: por %xmm3, %xmm1
; X86-SSE2-NEXT: pand %xmm1, %xmm0
; X86-SSE2-NEXT: pandn %xmm2, %xmm1
; X86-SSE2-NEXT: por %xmm0, %xmm1
; X86-SSE2-NEXT: movdqa %xmm1, %xmm0
; X86-SSE2-NEXT: pxor %xmm4, %xmm0
; X86-SSE2-NEXT: movdqa %xmm5, %xmm2
; X86-SSE2-NEXT: pxor %xmm4, %xmm2
; X86-SSE2-NEXT: movdqa %xmm2, %xmm3
; X86-SSE2-NEXT: pcmpgtd %xmm0, %xmm3
; X86-SSE2-NEXT: pcmpeqd %xmm0, %xmm2
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm3[0,0,2,2]
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
; X86-SSE2-NEXT: pand %xmm0, %xmm2
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm3[1,1,3,3]
; X86-SSE2-NEXT: por %xmm2, %xmm0
; X86-SSE2-NEXT: pand %xmm0, %xmm1
; X86-SSE2-NEXT: pandn %xmm5, %xmm0
; X86-SSE2-NEXT: por %xmm1, %xmm0
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: pxor %xmm4, %xmm2
; X86-SSE2-NEXT: pxor %xmm1, %xmm4
; X86-SSE2-NEXT: movdqa %xmm4, %xmm3
; X86-SSE2-NEXT: pcmpgtd %xmm2, %xmm3
; X86-SSE2-NEXT: pcmpeqd %xmm2, %xmm4
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm3[0,0,2,2]
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3]
; X86-SSE2-NEXT: pand %xmm2, %xmm4
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm3[1,1,3,3]
; X86-SSE2-NEXT: por %xmm4, %xmm2
; X86-SSE2-NEXT: pand %xmm2, %xmm0
; X86-SSE2-NEXT: pandn %xmm1, %xmm2
; X86-SSE2-NEXT: por %xmm0, %xmm2
; X86-SSE2-NEXT: movd %xmm2, %eax
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm2[1,1,1,1]
; X86-SSE2-NEXT: movd %xmm0, %edx
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v8i64:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: movdqa %xmm0, %xmm4
; X86-SSE42-NEXT: movdqa {{.*#+}} xmm5 = [0,2147483648,0,2147483648]
; X86-SSE42-NEXT: movdqa %xmm1, %xmm6
; X86-SSE42-NEXT: pxor %xmm5, %xmm6
; X86-SSE42-NEXT: movdqa %xmm3, %xmm0
; X86-SSE42-NEXT: pxor %xmm5, %xmm0
; X86-SSE42-NEXT: pcmpgtq %xmm6, %xmm0
; X86-SSE42-NEXT: blendvpd %xmm0, %xmm1, %xmm3
; X86-SSE42-NEXT: movdqa %xmm4, %xmm1
; X86-SSE42-NEXT: pxor %xmm5, %xmm1
; X86-SSE42-NEXT: movdqa %xmm2, %xmm0
; X86-SSE42-NEXT: pxor %xmm5, %xmm0
; X86-SSE42-NEXT: pcmpgtq %xmm1, %xmm0
; X86-SSE42-NEXT: blendvpd %xmm0, %xmm4, %xmm2
; X86-SSE42-NEXT: movapd %xmm2, %xmm1
; X86-SSE42-NEXT: xorpd %xmm5, %xmm1
; X86-SSE42-NEXT: movapd %xmm3, %xmm0
; X86-SSE42-NEXT: xorpd %xmm5, %xmm0
; X86-SSE42-NEXT: pcmpgtq %xmm1, %xmm0
; X86-SSE42-NEXT: blendvpd %xmm0, %xmm2, %xmm3
; X86-SSE42-NEXT: pshufd {{.*#+}} xmm1 = xmm3[2,3,2,3]
; X86-SSE42-NEXT: movdqa %xmm3, %xmm0
; X86-SSE42-NEXT: pxor %xmm5, %xmm0
; X86-SSE42-NEXT: pxor %xmm1, %xmm5
; X86-SSE42-NEXT: pcmpgtq %xmm0, %xmm5
; X86-SSE42-NEXT: movdqa %xmm5, %xmm0
; X86-SSE42-NEXT: blendvpd %xmm0, %xmm3, %xmm1
; X86-SSE42-NEXT: movd %xmm1, %eax
; X86-SSE42-NEXT: pextrd $1, %xmm1, %edx
; X86-SSE42-NEXT: retl
;
; X86-AVX1-LABEL: test_reduce_v8i64:
; X86-AVX1: ## %bb.0:
; X86-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
; X86-AVX1-NEXT: vmovddup {{.*#+}} xmm2 = [9223372036854775808,9223372036854775808]
; X86-AVX1-NEXT: ## xmm2 = mem[0,0]
; X86-AVX1-NEXT: vxorps %xmm2, %xmm3, %xmm4
; X86-AVX1-NEXT: vextractf128 $1, %ymm1, %xmm5
; X86-AVX1-NEXT: vxorps %xmm2, %xmm5, %xmm6
; X86-AVX1-NEXT: vpcmpgtq %xmm4, %xmm6, %xmm4
; X86-AVX1-NEXT: vxorps %xmm2, %xmm0, %xmm6
; X86-AVX1-NEXT: vxorps %xmm2, %xmm1, %xmm7
; X86-AVX1-NEXT: vpcmpgtq %xmm6, %xmm7, %xmm6
; X86-AVX1-NEXT: vblendvpd %xmm6, %xmm0, %xmm1, %xmm0
; X86-AVX1-NEXT: vxorpd %xmm2, %xmm0, %xmm1
; X86-AVX1-NEXT: vblendvpd %xmm4, %xmm3, %xmm5, %xmm3
; X86-AVX1-NEXT: vxorpd %xmm2, %xmm3, %xmm4
; X86-AVX1-NEXT: vpcmpgtq %xmm1, %xmm4, %xmm1
; X86-AVX1-NEXT: vblendvpd %xmm1, %xmm0, %xmm3, %xmm0
; X86-AVX1-NEXT: vpermilps {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-AVX1-NEXT: vxorpd %xmm2, %xmm0, %xmm3
; X86-AVX1-NEXT: vxorpd %xmm2, %xmm1, %xmm2
; X86-AVX1-NEXT: vpcmpgtq %xmm3, %xmm2, %xmm2
; X86-AVX1-NEXT: vblendvpd %xmm2, %xmm0, %xmm1, %xmm0
; X86-AVX1-NEXT: vmovd %xmm0, %eax
; X86-AVX1-NEXT: vpextrd $1, %xmm0, %edx
; X86-AVX1-NEXT: vzeroupper
; X86-AVX1-NEXT: retl
;
; X86-AVX2-LABEL: test_reduce_v8i64:
; X86-AVX2: ## %bb.0:
; X86-AVX2-NEXT: vpbroadcastq {{.*#+}} ymm2 = [9223372036854775808,9223372036854775808,9223372036854775808,9223372036854775808]
; X86-AVX2-NEXT: vpxor %ymm2, %ymm0, %ymm3
; X86-AVX2-NEXT: vpxor %ymm2, %ymm1, %ymm4
; X86-AVX2-NEXT: vpcmpgtq %ymm3, %ymm4, %ymm3
; X86-AVX2-NEXT: vblendvpd %ymm3, %ymm0, %ymm1, %ymm0
; X86-AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1
; X86-AVX2-NEXT: vxorpd %xmm2, %xmm0, %xmm3
; X86-AVX2-NEXT: vxorpd %xmm2, %xmm1, %xmm4
; X86-AVX2-NEXT: vpcmpgtq %xmm3, %xmm4, %xmm3
; X86-AVX2-NEXT: vblendvpd %xmm3, %xmm0, %xmm1, %xmm0
; X86-AVX2-NEXT: vpermilps {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-AVX2-NEXT: vxorpd %xmm2, %xmm0, %xmm3
; X86-AVX2-NEXT: vxorpd %xmm2, %xmm1, %xmm2
; X86-AVX2-NEXT: vpcmpgtq %xmm3, %xmm2, %xmm2
; X86-AVX2-NEXT: vblendvpd %xmm2, %xmm0, %xmm1, %xmm0
; X86-AVX2-NEXT: vmovd %xmm0, %eax
; X86-AVX2-NEXT: vpextrd $1, %xmm0, %edx
; X86-AVX2-NEXT: vzeroupper
; X86-AVX2-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v8i64:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: movdqa {{.*#+}} xmm4 = [9223372039002259456,9223372039002259456]
; X64-SSE2-NEXT: movdqa %xmm1, %xmm5
; X64-SSE2-NEXT: pxor %xmm4, %xmm5
; X64-SSE2-NEXT: movdqa %xmm3, %xmm6
; X64-SSE2-NEXT: pxor %xmm4, %xmm6
; X64-SSE2-NEXT: movdqa %xmm6, %xmm7
; X64-SSE2-NEXT: pcmpgtd %xmm5, %xmm7
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm8 = xmm7[0,0,2,2]
; X64-SSE2-NEXT: pcmpeqd %xmm5, %xmm6
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm6[1,1,3,3]
; X64-SSE2-NEXT: pand %xmm8, %xmm6
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm7[1,1,3,3]
; X64-SSE2-NEXT: por %xmm6, %xmm5
; X64-SSE2-NEXT: pand %xmm5, %xmm1
; X64-SSE2-NEXT: pandn %xmm3, %xmm5
; X64-SSE2-NEXT: por %xmm1, %xmm5
; X64-SSE2-NEXT: movdqa %xmm0, %xmm1
; X64-SSE2-NEXT: pxor %xmm4, %xmm1
; X64-SSE2-NEXT: movdqa %xmm2, %xmm3
; X64-SSE2-NEXT: pxor %xmm4, %xmm3
; X64-SSE2-NEXT: movdqa %xmm3, %xmm6
; X64-SSE2-NEXT: pcmpgtd %xmm1, %xmm6
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm7 = xmm6[0,0,2,2]
; X64-SSE2-NEXT: pcmpeqd %xmm1, %xmm3
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm3[1,1,3,3]
; X64-SSE2-NEXT: pand %xmm7, %xmm1
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm6[1,1,3,3]
; X64-SSE2-NEXT: por %xmm1, %xmm3
; X64-SSE2-NEXT: pand %xmm3, %xmm0
; X64-SSE2-NEXT: pandn %xmm2, %xmm3
; X64-SSE2-NEXT: por %xmm0, %xmm3
; X64-SSE2-NEXT: movdqa %xmm3, %xmm0
; X64-SSE2-NEXT: pxor %xmm4, %xmm0
; X64-SSE2-NEXT: movdqa %xmm5, %xmm1
; X64-SSE2-NEXT: pxor %xmm4, %xmm1
; X64-SSE2-NEXT: movdqa %xmm1, %xmm2
; X64-SSE2-NEXT: pcmpgtd %xmm0, %xmm2
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm6 = xmm2[0,0,2,2]
; X64-SSE2-NEXT: pcmpeqd %xmm0, %xmm1
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,3,3]
; X64-SSE2-NEXT: pand %xmm6, %xmm0
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm2[1,1,3,3]
; X64-SSE2-NEXT: por %xmm0, %xmm1
; X64-SSE2-NEXT: pand %xmm1, %xmm3
; X64-SSE2-NEXT: pandn %xmm5, %xmm1
; X64-SSE2-NEXT: por %xmm3, %xmm1
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[2,3,2,3]
; X64-SSE2-NEXT: movdqa %xmm1, %xmm2
; X64-SSE2-NEXT: pxor %xmm4, %xmm2
; X64-SSE2-NEXT: pxor %xmm0, %xmm4
; X64-SSE2-NEXT: movdqa %xmm4, %xmm3
; X64-SSE2-NEXT: pcmpgtd %xmm2, %xmm3
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm5 = xmm3[0,0,2,2]
; X64-SSE2-NEXT: pcmpeqd %xmm2, %xmm4
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm4[1,1,3,3]
; X64-SSE2-NEXT: pand %xmm5, %xmm2
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
; X64-SSE2-NEXT: por %xmm2, %xmm3
; X64-SSE2-NEXT: pand %xmm3, %xmm1
; X64-SSE2-NEXT: pandn %xmm0, %xmm3
; X64-SSE2-NEXT: por %xmm1, %xmm3
; X64-SSE2-NEXT: movq %xmm3, %rax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v8i64:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: movdqa %xmm0, %xmm5
; X64-SSE42-NEXT: movdqa {{.*#+}} xmm4 = [9223372036854775808,9223372036854775808]
; X64-SSE42-NEXT: movdqa %xmm1, %xmm6
; X64-SSE42-NEXT: pxor %xmm4, %xmm6
; X64-SSE42-NEXT: movdqa %xmm3, %xmm0
; X64-SSE42-NEXT: pxor %xmm4, %xmm0
; X64-SSE42-NEXT: pcmpgtq %xmm6, %xmm0
; X64-SSE42-NEXT: blendvpd %xmm0, %xmm1, %xmm3
; X64-SSE42-NEXT: movdqa %xmm5, %xmm1
; X64-SSE42-NEXT: pxor %xmm4, %xmm1
; X64-SSE42-NEXT: movdqa %xmm2, %xmm0
; X64-SSE42-NEXT: pxor %xmm4, %xmm0
; X64-SSE42-NEXT: pcmpgtq %xmm1, %xmm0
; X64-SSE42-NEXT: blendvpd %xmm0, %xmm5, %xmm2
; X64-SSE42-NEXT: movapd %xmm2, %xmm1
; X64-SSE42-NEXT: xorpd %xmm4, %xmm1
; X64-SSE42-NEXT: movapd %xmm3, %xmm0
; X64-SSE42-NEXT: xorpd %xmm4, %xmm0
; X64-SSE42-NEXT: pcmpgtq %xmm1, %xmm0
; X64-SSE42-NEXT: blendvpd %xmm0, %xmm2, %xmm3
; X64-SSE42-NEXT: pshufd {{.*#+}} xmm1 = xmm3[2,3,2,3]
; X64-SSE42-NEXT: movdqa %xmm3, %xmm0
; X64-SSE42-NEXT: pxor %xmm4, %xmm0
; X64-SSE42-NEXT: pxor %xmm1, %xmm4
; X64-SSE42-NEXT: pcmpgtq %xmm0, %xmm4
; X64-SSE42-NEXT: movdqa %xmm4, %xmm0
; X64-SSE42-NEXT: blendvpd %xmm0, %xmm3, %xmm1
; X64-SSE42-NEXT: movq %xmm1, %rax
; X64-SSE42-NEXT: retq
;
; X64-AVX1-LABEL: test_reduce_v8i64:
; X64-AVX1: ## %bb.0:
; X64-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm2
; X64-AVX1-NEXT: vmovdqa {{.*#+}} xmm3 = [9223372036854775808,9223372036854775808]
; X64-AVX1-NEXT: vpxor %xmm3, %xmm2, %xmm4
; X64-AVX1-NEXT: vextractf128 $1, %ymm1, %xmm5
; X64-AVX1-NEXT: vpxor %xmm3, %xmm5, %xmm6
; X64-AVX1-NEXT: vpcmpgtq %xmm4, %xmm6, %xmm4
; X64-AVX1-NEXT: vpxor %xmm3, %xmm0, %xmm6
; X64-AVX1-NEXT: vpxor %xmm3, %xmm1, %xmm7
; X64-AVX1-NEXT: vpcmpgtq %xmm6, %xmm7, %xmm6
; X64-AVX1-NEXT: vblendvpd %xmm6, %xmm0, %xmm1, %xmm0
; X64-AVX1-NEXT: vxorpd %xmm3, %xmm0, %xmm1
; X64-AVX1-NEXT: vblendvpd %xmm4, %xmm2, %xmm5, %xmm2
; X64-AVX1-NEXT: vxorpd %xmm3, %xmm2, %xmm4
; X64-AVX1-NEXT: vpcmpgtq %xmm1, %xmm4, %xmm1
; X64-AVX1-NEXT: vblendvpd %xmm1, %xmm0, %xmm2, %xmm0
; X64-AVX1-NEXT: vpermilps {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-AVX1-NEXT: vxorpd %xmm3, %xmm0, %xmm2
; X64-AVX1-NEXT: vxorpd %xmm3, %xmm1, %xmm3
; X64-AVX1-NEXT: vpcmpgtq %xmm2, %xmm3, %xmm2
; X64-AVX1-NEXT: vblendvpd %xmm2, %xmm0, %xmm1, %xmm0
; X64-AVX1-NEXT: vmovq %xmm0, %rax
; X64-AVX1-NEXT: vzeroupper
; X64-AVX1-NEXT: retq
;
; X64-AVX2-LABEL: test_reduce_v8i64:
; X64-AVX2: ## %bb.0:
; X64-AVX2-NEXT: vpbroadcastq {{.*#+}} ymm2 = [9223372036854775808,9223372036854775808,9223372036854775808,9223372036854775808]
; X64-AVX2-NEXT: vpxor %ymm2, %ymm0, %ymm3
; X64-AVX2-NEXT: vpxor %ymm2, %ymm1, %ymm4
; X64-AVX2-NEXT: vpcmpgtq %ymm3, %ymm4, %ymm3
; X64-AVX2-NEXT: vblendvpd %ymm3, %ymm0, %ymm1, %ymm0
; X64-AVX2-NEXT: vextractf128 $1, %ymm0, %xmm1
; X64-AVX2-NEXT: vxorpd %xmm2, %xmm0, %xmm3
; X64-AVX2-NEXT: vxorpd %xmm2, %xmm1, %xmm4
; X64-AVX2-NEXT: vpcmpgtq %xmm3, %xmm4, %xmm3
; X64-AVX2-NEXT: vblendvpd %xmm3, %xmm0, %xmm1, %xmm0
; X64-AVX2-NEXT: vpermilps {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-AVX2-NEXT: vxorpd %xmm2, %xmm0, %xmm3
; X64-AVX2-NEXT: vxorpd %xmm2, %xmm1, %xmm2
; X64-AVX2-NEXT: vpcmpgtq %xmm3, %xmm2, %xmm2
; X64-AVX2-NEXT: vblendvpd %xmm2, %xmm0, %xmm1, %xmm0
; X64-AVX2-NEXT: vmovq %xmm0, %rax
; X64-AVX2-NEXT: vzeroupper
; X64-AVX2-NEXT: retq
;
; X64-AVX512-LABEL: test_reduce_v8i64:
; X64-AVX512: ## %bb.0:
; X64-AVX512-NEXT: vextracti64x4 $1, %zmm0, %ymm1
; X64-AVX512-NEXT: vpminuq %zmm1, %zmm0, %zmm0
; X64-AVX512-NEXT: vextracti128 $1, %ymm0, %xmm1
; X64-AVX512-NEXT: vpminuq %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-AVX512-NEXT: vpminuq %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vmovq %xmm0, %rax
; X64-AVX512-NEXT: vzeroupper
; X64-AVX512-NEXT: retq
%1 = shufflevector <8 x i64> %a0, <8 x i64> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef>
%2 = icmp ult <8 x i64> %a0, %1
%3 = select <8 x i1> %2, <8 x i64> %a0, <8 x i64> %1
%4 = shufflevector <8 x i64> %3, <8 x i64> undef, <8 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%5 = icmp ult <8 x i64> %3, %4
%6 = select <8 x i1> %5, <8 x i64> %3, <8 x i64> %4
%7 = shufflevector <8 x i64> %6, <8 x i64> undef, <8 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%8 = icmp ult <8 x i64> %6, %7
%9 = select <8 x i1> %8, <8 x i64> %6, <8 x i64> %7
%10 = extractelement <8 x i64> %9, i32 0
ret i64 %10
}
define i32 @test_reduce_v16i32(<16 x i32> %a0) {
; X86-SSE2-LABEL: test_reduce_v16i32:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: movdqa {{.*#+}} xmm4 = [2147483648,2147483648,2147483648,2147483648]
; X86-SSE2-NEXT: movdqa %xmm1, %xmm6
; X86-SSE2-NEXT: pxor %xmm4, %xmm6
; X86-SSE2-NEXT: movdqa %xmm3, %xmm5
; X86-SSE2-NEXT: pxor %xmm4, %xmm5
; X86-SSE2-NEXT: pcmpgtd %xmm6, %xmm5
; X86-SSE2-NEXT: pand %xmm5, %xmm1
; X86-SSE2-NEXT: pandn %xmm3, %xmm5
; X86-SSE2-NEXT: por %xmm1, %xmm5
; X86-SSE2-NEXT: movdqa %xmm0, %xmm3
; X86-SSE2-NEXT: pxor %xmm4, %xmm3
; X86-SSE2-NEXT: movdqa %xmm2, %xmm1
; X86-SSE2-NEXT: pxor %xmm4, %xmm1
; X86-SSE2-NEXT: pcmpgtd %xmm3, %xmm1
; X86-SSE2-NEXT: pand %xmm1, %xmm0
; X86-SSE2-NEXT: pandn %xmm2, %xmm1
; X86-SSE2-NEXT: por %xmm0, %xmm1
; X86-SSE2-NEXT: movdqa %xmm1, %xmm2
; X86-SSE2-NEXT: pxor %xmm4, %xmm2
; X86-SSE2-NEXT: movdqa %xmm5, %xmm0
; X86-SSE2-NEXT: pxor %xmm4, %xmm0
; X86-SSE2-NEXT: pcmpgtd %xmm2, %xmm0
; X86-SSE2-NEXT: pand %xmm0, %xmm1
; X86-SSE2-NEXT: pandn %xmm5, %xmm0
; X86-SSE2-NEXT: por %xmm1, %xmm0
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[2,3,2,3]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm3
; X86-SSE2-NEXT: pxor %xmm4, %xmm3
; X86-SSE2-NEXT: movdqa %xmm2, %xmm1
; X86-SSE2-NEXT: pxor %xmm4, %xmm1
; X86-SSE2-NEXT: pcmpgtd %xmm3, %xmm1
; X86-SSE2-NEXT: pand %xmm1, %xmm0
; X86-SSE2-NEXT: pandn %xmm2, %xmm1
; X86-SSE2-NEXT: por %xmm0, %xmm1
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
; X86-SSE2-NEXT: movdqa %xmm1, %xmm2
; X86-SSE2-NEXT: pxor %xmm4, %xmm2
; X86-SSE2-NEXT: pxor %xmm0, %xmm4
; X86-SSE2-NEXT: pcmpgtd %xmm2, %xmm4
; X86-SSE2-NEXT: pand %xmm4, %xmm1
; X86-SSE2-NEXT: pandn %xmm0, %xmm4
; X86-SSE2-NEXT: por %xmm1, %xmm4
; X86-SSE2-NEXT: movd %xmm4, %eax
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v16i32:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: pminud %xmm3, %xmm1
; X86-SSE42-NEXT: pminud %xmm2, %xmm1
; X86-SSE42-NEXT: pminud %xmm0, %xmm1
; X86-SSE42-NEXT: pshufd {{.*#+}} xmm0 = xmm1[2,3,2,3]
; X86-SSE42-NEXT: pminud %xmm1, %xmm0
; X86-SSE42-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X86-SSE42-NEXT: pminud %xmm0, %xmm1
; X86-SSE42-NEXT: movd %xmm1, %eax
; X86-SSE42-NEXT: retl
;
; X86-AVX1-LABEL: test_reduce_v16i32:
; X86-AVX1: ## %bb.0:
; X86-AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; X86-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
; X86-AVX1-NEXT: vpminud %xmm2, %xmm3, %xmm2
; X86-AVX1-NEXT: vpminud %xmm2, %xmm1, %xmm1
; X86-AVX1-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X86-AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-AVX1-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X86-AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X86-AVX1-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X86-AVX1-NEXT: vmovd %xmm0, %eax
; X86-AVX1-NEXT: vzeroupper
; X86-AVX1-NEXT: retl
;
; X86-AVX2-LABEL: test_reduce_v16i32:
; X86-AVX2: ## %bb.0:
; X86-AVX2-NEXT: vpminud %ymm1, %ymm0, %ymm0
; X86-AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; X86-AVX2-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X86-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-AVX2-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X86-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X86-AVX2-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X86-AVX2-NEXT: vmovd %xmm0, %eax
; X86-AVX2-NEXT: vzeroupper
; X86-AVX2-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v16i32:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: movdqa {{.*#+}} xmm4 = [2147483648,2147483648,2147483648,2147483648]
; X64-SSE2-NEXT: movdqa %xmm1, %xmm5
; X64-SSE2-NEXT: pxor %xmm4, %xmm5
; X64-SSE2-NEXT: movdqa %xmm3, %xmm6
; X64-SSE2-NEXT: pxor %xmm4, %xmm6
; X64-SSE2-NEXT: pcmpgtd %xmm5, %xmm6
; X64-SSE2-NEXT: pand %xmm6, %xmm1
; X64-SSE2-NEXT: pandn %xmm3, %xmm6
; X64-SSE2-NEXT: por %xmm1, %xmm6
; X64-SSE2-NEXT: movdqa %xmm0, %xmm1
; X64-SSE2-NEXT: pxor %xmm4, %xmm1
; X64-SSE2-NEXT: movdqa %xmm2, %xmm3
; X64-SSE2-NEXT: pxor %xmm4, %xmm3
; X64-SSE2-NEXT: pcmpgtd %xmm1, %xmm3
; X64-SSE2-NEXT: pand %xmm3, %xmm0
; X64-SSE2-NEXT: pandn %xmm2, %xmm3
; X64-SSE2-NEXT: por %xmm0, %xmm3
; X64-SSE2-NEXT: movdqa %xmm3, %xmm0
; X64-SSE2-NEXT: pxor %xmm4, %xmm0
; X64-SSE2-NEXT: movdqa %xmm6, %xmm1
; X64-SSE2-NEXT: pxor %xmm4, %xmm1
; X64-SSE2-NEXT: pcmpgtd %xmm0, %xmm1
; X64-SSE2-NEXT: pand %xmm1, %xmm3
; X64-SSE2-NEXT: pandn %xmm6, %xmm1
; X64-SSE2-NEXT: por %xmm3, %xmm1
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[2,3,2,3]
; X64-SSE2-NEXT: movdqa %xmm1, %xmm2
; X64-SSE2-NEXT: pxor %xmm4, %xmm2
; X64-SSE2-NEXT: movdqa %xmm0, %xmm3
; X64-SSE2-NEXT: pxor %xmm4, %xmm3
; X64-SSE2-NEXT: pcmpgtd %xmm2, %xmm3
; X64-SSE2-NEXT: pand %xmm3, %xmm1
; X64-SSE2-NEXT: pandn %xmm0, %xmm3
; X64-SSE2-NEXT: por %xmm1, %xmm3
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm3[1,1,1,1]
; X64-SSE2-NEXT: movdqa %xmm3, %xmm1
; X64-SSE2-NEXT: pxor %xmm4, %xmm1
; X64-SSE2-NEXT: pxor %xmm0, %xmm4
; X64-SSE2-NEXT: pcmpgtd %xmm1, %xmm4
; X64-SSE2-NEXT: pand %xmm4, %xmm3
; X64-SSE2-NEXT: pandn %xmm0, %xmm4
; X64-SSE2-NEXT: por %xmm3, %xmm4
; X64-SSE2-NEXT: movd %xmm4, %eax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v16i32:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: pminud %xmm3, %xmm1
; X64-SSE42-NEXT: pminud %xmm2, %xmm1
; X64-SSE42-NEXT: pminud %xmm0, %xmm1
; X64-SSE42-NEXT: pshufd {{.*#+}} xmm0 = xmm1[2,3,2,3]
; X64-SSE42-NEXT: pminud %xmm1, %xmm0
; X64-SSE42-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X64-SSE42-NEXT: pminud %xmm0, %xmm1
; X64-SSE42-NEXT: movd %xmm1, %eax
; X64-SSE42-NEXT: retq
;
; X64-AVX1-LABEL: test_reduce_v16i32:
; X64-AVX1: ## %bb.0:
; X64-AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; X64-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
; X64-AVX1-NEXT: vpminud %xmm2, %xmm3, %xmm2
; X64-AVX1-NEXT: vpminud %xmm2, %xmm1, %xmm1
; X64-AVX1-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-AVX1-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX1-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X64-AVX1-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX1-NEXT: vmovd %xmm0, %eax
; X64-AVX1-NEXT: vzeroupper
; X64-AVX1-NEXT: retq
;
; X64-AVX2-LABEL: test_reduce_v16i32:
; X64-AVX2: ## %bb.0:
; X64-AVX2-NEXT: vpminud %ymm1, %ymm0, %ymm0
; X64-AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; X64-AVX2-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-AVX2-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X64-AVX2-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT: vmovd %xmm0, %eax
; X64-AVX2-NEXT: vzeroupper
; X64-AVX2-NEXT: retq
;
; X64-AVX512-LABEL: test_reduce_v16i32:
; X64-AVX512: ## %bb.0:
; X64-AVX512-NEXT: vextracti64x4 $1, %zmm0, %ymm1
; X64-AVX512-NEXT: vpminud %zmm1, %zmm0, %zmm0
; X64-AVX512-NEXT: vextracti128 $1, %ymm0, %xmm1
; X64-AVX512-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-AVX512-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X64-AVX512-NEXT: vpminud %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vmovd %xmm0, %eax
; X64-AVX512-NEXT: vzeroupper
; X64-AVX512-NEXT: retq
%1 = shufflevector <16 x i32> %a0, <16 x i32> undef, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%2 = icmp ult <16 x i32> %a0, %1
%3 = select <16 x i1> %2, <16 x i32> %a0, <16 x i32> %1
%4 = shufflevector <16 x i32> %3, <16 x i32> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%5 = icmp ult <16 x i32> %3, %4
%6 = select <16 x i1> %5, <16 x i32> %3, <16 x i32> %4
%7 = shufflevector <16 x i32> %6, <16 x i32> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%8 = icmp ult <16 x i32> %6, %7
%9 = select <16 x i1> %8, <16 x i32> %6, <16 x i32> %7
%10 = shufflevector <16 x i32> %9, <16 x i32> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%11 = icmp ult <16 x i32> %9, %10
%12 = select <16 x i1> %11, <16 x i32> %9, <16 x i32> %10
%13 = extractelement <16 x i32> %12, i32 0
ret i32 %13
}
define i16 @test_reduce_v32i16(<32 x i16> %a0) {
; X86-SSE2-LABEL: test_reduce_v32i16:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: movdqa %xmm1, %xmm4
; X86-SSE2-NEXT: psubusw %xmm3, %xmm4
; X86-SSE2-NEXT: psubw %xmm4, %xmm1
; X86-SSE2-NEXT: movdqa %xmm0, %xmm3
; X86-SSE2-NEXT: psubusw %xmm2, %xmm3
; X86-SSE2-NEXT: psubw %xmm3, %xmm0
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
; X86-SSE2-NEXT: psrld $16, %xmm1
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: movd %xmm0, %eax
; X86-SSE2-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v32i16:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: pminuw %xmm3, %xmm1
; X86-SSE42-NEXT: pminuw %xmm2, %xmm1
; X86-SSE42-NEXT: pminuw %xmm0, %xmm1
; X86-SSE42-NEXT: phminposuw %xmm1, %xmm0
; X86-SSE42-NEXT: movd %xmm0, %eax
; X86-SSE42-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-SSE42-NEXT: retl
;
; X86-AVX1-LABEL: test_reduce_v32i16:
; X86-AVX1: ## %bb.0:
; X86-AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; X86-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
; X86-AVX1-NEXT: vpminuw %xmm2, %xmm3, %xmm2
; X86-AVX1-NEXT: vpminuw %xmm2, %xmm1, %xmm1
; X86-AVX1-NEXT: vpminuw %xmm1, %xmm0, %xmm0
; X86-AVX1-NEXT: vphminposuw %xmm0, %xmm0
; X86-AVX1-NEXT: vmovd %xmm0, %eax
; X86-AVX1-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-AVX1-NEXT: vzeroupper
; X86-AVX1-NEXT: retl
;
; X86-AVX2-LABEL: test_reduce_v32i16:
; X86-AVX2: ## %bb.0:
; X86-AVX2-NEXT: vpminuw %ymm1, %ymm0, %ymm0
; X86-AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; X86-AVX2-NEXT: vpminuw %xmm1, %xmm0, %xmm0
; X86-AVX2-NEXT: vphminposuw %xmm0, %xmm0
; X86-AVX2-NEXT: vmovd %xmm0, %eax
; X86-AVX2-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-AVX2-NEXT: vzeroupper
; X86-AVX2-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v32i16:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: movdqa %xmm1, %xmm4
; X64-SSE2-NEXT: psubusw %xmm3, %xmm4
; X64-SSE2-NEXT: psubw %xmm4, %xmm1
; X64-SSE2-NEXT: movdqa %xmm0, %xmm3
; X64-SSE2-NEXT: psubusw %xmm2, %xmm3
; X64-SSE2-NEXT: psubw %xmm3, %xmm0
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: movdqa %xmm0, %xmm1
; X64-SSE2-NEXT: psrld $16, %xmm1
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: movd %xmm0, %eax
; X64-SSE2-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v32i16:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: pminuw %xmm3, %xmm1
; X64-SSE42-NEXT: pminuw %xmm2, %xmm1
; X64-SSE42-NEXT: pminuw %xmm0, %xmm1
; X64-SSE42-NEXT: phminposuw %xmm1, %xmm0
; X64-SSE42-NEXT: movd %xmm0, %eax
; X64-SSE42-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-SSE42-NEXT: retq
;
; X64-AVX1-LABEL: test_reduce_v32i16:
; X64-AVX1: ## %bb.0:
; X64-AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; X64-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
; X64-AVX1-NEXT: vpminuw %xmm2, %xmm3, %xmm2
; X64-AVX1-NEXT: vpminuw %xmm2, %xmm1, %xmm1
; X64-AVX1-NEXT: vpminuw %xmm1, %xmm0, %xmm0
; X64-AVX1-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX1-NEXT: vmovd %xmm0, %eax
; X64-AVX1-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-AVX1-NEXT: vzeroupper
; X64-AVX1-NEXT: retq
;
; X64-AVX2-LABEL: test_reduce_v32i16:
; X64-AVX2: ## %bb.0:
; X64-AVX2-NEXT: vpminuw %ymm1, %ymm0, %ymm0
; X64-AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; X64-AVX2-NEXT: vpminuw %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX2-NEXT: vmovd %xmm0, %eax
; X64-AVX2-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-AVX2-NEXT: vzeroupper
; X64-AVX2-NEXT: retq
;
; X64-AVX512-LABEL: test_reduce_v32i16:
; X64-AVX512: ## %bb.0:
; X64-AVX512-NEXT: vextracti64x4 $1, %zmm0, %ymm1
; X64-AVX512-NEXT: vpminuw %ymm1, %ymm0, %ymm0
; X64-AVX512-NEXT: vextracti128 $1, %ymm0, %xmm1
; X64-AVX512-NEXT: vpminuw %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX512-NEXT: vmovd %xmm0, %eax
; X64-AVX512-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-AVX512-NEXT: vzeroupper
; X64-AVX512-NEXT: retq
%1 = shufflevector <32 x i16> %a0, <32 x i16> undef, <32 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%2 = icmp ult <32 x i16> %a0, %1
%3 = select <32 x i1> %2, <32 x i16> %a0, <32 x i16> %1
%4 = shufflevector <32 x i16> %3, <32 x i16> undef, <32 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%5 = icmp ult <32 x i16> %3, %4
%6 = select <32 x i1> %5, <32 x i16> %3, <32 x i16> %4
%7 = shufflevector <32 x i16> %6, <32 x i16> undef, <32 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%8 = icmp ult <32 x i16> %6, %7
%9 = select <32 x i1> %8, <32 x i16> %6, <32 x i16> %7
%10 = shufflevector <32 x i16> %9, <32 x i16> undef, <32 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%11 = icmp ult <32 x i16> %9, %10
%12 = select <32 x i1> %11, <32 x i16> %9, <32 x i16> %10
%13 = shufflevector <32 x i16> %12, <32 x i16> undef, <32 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%14 = icmp ult <32 x i16> %12, %13
%15 = select <32 x i1> %14, <32 x i16> %12, <32 x i16> %13
%16 = extractelement <32 x i16> %15, i32 0
ret i16 %16
}
define i8 @test_reduce_v64i8(<64 x i8> %a0) {
; X86-SSE2-LABEL: test_reduce_v64i8:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: pminub %xmm3, %xmm1
; X86-SSE2-NEXT: pminub %xmm2, %xmm1
; X86-SSE2-NEXT: pminub %xmm0, %xmm1
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[2,3,2,3]
; X86-SSE2-NEXT: pminub %xmm1, %xmm0
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X86-SSE2-NEXT: pminub %xmm0, %xmm1
; X86-SSE2-NEXT: movdqa %xmm1, %xmm0
; X86-SSE2-NEXT: psrld $16, %xmm0
; X86-SSE2-NEXT: pminub %xmm1, %xmm0
; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
; X86-SSE2-NEXT: psrlw $8, %xmm1
; X86-SSE2-NEXT: pminub %xmm0, %xmm1
; X86-SSE2-NEXT: movd %xmm1, %eax
; X86-SSE2-NEXT: ## kill: def $al killed $al killed $eax
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v64i8:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: pminub %xmm3, %xmm1
; X86-SSE42-NEXT: pminub %xmm2, %xmm1
; X86-SSE42-NEXT: pminub %xmm0, %xmm1
; X86-SSE42-NEXT: movdqa %xmm1, %xmm0
; X86-SSE42-NEXT: psrlw $8, %xmm0
; X86-SSE42-NEXT: pminub %xmm1, %xmm0
; X86-SSE42-NEXT: phminposuw %xmm0, %xmm0
; X86-SSE42-NEXT: movd %xmm0, %eax
; X86-SSE42-NEXT: ## kill: def $al killed $al killed $eax
; X86-SSE42-NEXT: retl
;
; X86-AVX1-LABEL: test_reduce_v64i8:
; X86-AVX1: ## %bb.0:
; X86-AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; X86-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
; X86-AVX1-NEXT: vpminub %xmm2, %xmm3, %xmm2
; X86-AVX1-NEXT: vpminub %xmm2, %xmm1, %xmm1
; X86-AVX1-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X86-AVX1-NEXT: vpsrlw $8, %xmm0, %xmm1
; X86-AVX1-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X86-AVX1-NEXT: vphminposuw %xmm0, %xmm0
; X86-AVX1-NEXT: vmovd %xmm0, %eax
; X86-AVX1-NEXT: ## kill: def $al killed $al killed $eax
; X86-AVX1-NEXT: vzeroupper
; X86-AVX1-NEXT: retl
;
; X86-AVX2-LABEL: test_reduce_v64i8:
; X86-AVX2: ## %bb.0:
; X86-AVX2-NEXT: vpminub %ymm1, %ymm0, %ymm0
; X86-AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; X86-AVX2-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X86-AVX2-NEXT: vpsrlw $8, %xmm0, %xmm1
; X86-AVX2-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X86-AVX2-NEXT: vphminposuw %xmm0, %xmm0
; X86-AVX2-NEXT: vmovd %xmm0, %eax
; X86-AVX2-NEXT: ## kill: def $al killed $al killed $eax
; X86-AVX2-NEXT: vzeroupper
; X86-AVX2-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v64i8:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: pminub %xmm3, %xmm1
; X64-SSE2-NEXT: pminub %xmm2, %xmm1
; X64-SSE2-NEXT: pminub %xmm0, %xmm1
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[2,3,2,3]
; X64-SSE2-NEXT: pminub %xmm1, %xmm0
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X64-SSE2-NEXT: pminub %xmm0, %xmm1
; X64-SSE2-NEXT: movdqa %xmm1, %xmm0
; X64-SSE2-NEXT: psrld $16, %xmm0
; X64-SSE2-NEXT: pminub %xmm1, %xmm0
; X64-SSE2-NEXT: movdqa %xmm0, %xmm1
; X64-SSE2-NEXT: psrlw $8, %xmm1
; X64-SSE2-NEXT: pminub %xmm0, %xmm1
; X64-SSE2-NEXT: movd %xmm1, %eax
; X64-SSE2-NEXT: ## kill: def $al killed $al killed $eax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v64i8:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: pminub %xmm3, %xmm1
; X64-SSE42-NEXT: pminub %xmm2, %xmm1
; X64-SSE42-NEXT: pminub %xmm0, %xmm1
; X64-SSE42-NEXT: movdqa %xmm1, %xmm0
; X64-SSE42-NEXT: psrlw $8, %xmm0
; X64-SSE42-NEXT: pminub %xmm1, %xmm0
; X64-SSE42-NEXT: phminposuw %xmm0, %xmm0
; X64-SSE42-NEXT: movd %xmm0, %eax
; X64-SSE42-NEXT: ## kill: def $al killed $al killed $eax
; X64-SSE42-NEXT: retq
;
; X64-AVX1-LABEL: test_reduce_v64i8:
; X64-AVX1: ## %bb.0:
; X64-AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2
; X64-AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3
; X64-AVX1-NEXT: vpminub %xmm2, %xmm3, %xmm2
; X64-AVX1-NEXT: vpminub %xmm2, %xmm1, %xmm1
; X64-AVX1-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X64-AVX1-NEXT: vpsrlw $8, %xmm0, %xmm1
; X64-AVX1-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X64-AVX1-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX1-NEXT: vmovd %xmm0, %eax
; X64-AVX1-NEXT: ## kill: def $al killed $al killed $eax
; X64-AVX1-NEXT: vzeroupper
; X64-AVX1-NEXT: retq
;
; X64-AVX2-LABEL: test_reduce_v64i8:
; X64-AVX2: ## %bb.0:
; X64-AVX2-NEXT: vpminub %ymm1, %ymm0, %ymm0
; X64-AVX2-NEXT: vextracti128 $1, %ymm0, %xmm1
; X64-AVX2-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT: vpsrlw $8, %xmm0, %xmm1
; X64-AVX2-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X64-AVX2-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX2-NEXT: vmovd %xmm0, %eax
; X64-AVX2-NEXT: ## kill: def $al killed $al killed $eax
; X64-AVX2-NEXT: vzeroupper
; X64-AVX2-NEXT: retq
;
; X64-AVX512-LABEL: test_reduce_v64i8:
; X64-AVX512: ## %bb.0:
; X64-AVX512-NEXT: vextracti64x4 $1, %zmm0, %ymm1
; X64-AVX512-NEXT: vpminub %ymm1, %ymm0, %ymm0
; X64-AVX512-NEXT: vextracti128 $1, %ymm0, %xmm1
; X64-AVX512-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vpsrlw $8, %xmm0, %xmm1
; X64-AVX512-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X64-AVX512-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX512-NEXT: vmovd %xmm0, %eax
; X64-AVX512-NEXT: ## kill: def $al killed $al killed $eax
; X64-AVX512-NEXT: vzeroupper
; X64-AVX512-NEXT: retq
%1 = shufflevector <64 x i8> %a0, <64 x i8> undef, <64 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%2 = icmp ult <64 x i8> %a0, %1
%3 = select <64 x i1> %2, <64 x i8> %a0, <64 x i8> %1
%4 = shufflevector <64 x i8> %3, <64 x i8> undef, <64 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%5 = icmp ult <64 x i8> %3, %4
%6 = select <64 x i1> %5, <64 x i8> %3, <64 x i8> %4
%7 = shufflevector <64 x i8> %6, <64 x i8> undef, <64 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%8 = icmp ult <64 x i8> %6, %7
%9 = select <64 x i1> %8, <64 x i8> %6, <64 x i8> %7
%10 = shufflevector <64 x i8> %9, <64 x i8> undef, <64 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%11 = icmp ult <64 x i8> %9, %10
%12 = select <64 x i1> %11, <64 x i8> %9, <64 x i8> %10
%13 = shufflevector <64 x i8> %12, <64 x i8> undef, <64 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%14 = icmp ult <64 x i8> %12, %13
%15 = select <64 x i1> %14, <64 x i8> %12, <64 x i8> %13
%16 = shufflevector <64 x i8> %15, <64 x i8> undef, <64 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%17 = icmp ult <64 x i8> %15, %16
%18 = select <64 x i1> %17, <64 x i8> %15, <64 x i8> %16
%19 = extractelement <64 x i8> %18, i32 0
ret i8 %19
}
;
; Partial Vector Reductions
;
define i16 @test_reduce_v16i16_v8i16(<16 x i16> %a0) {
; X86-SSE2-LABEL: test_reduce_v16i16_v8i16:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
; X86-SSE2-NEXT: psrld $16, %xmm1
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: movd %xmm0, %eax
; X86-SSE2-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v16i16_v8i16:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: phminposuw %xmm0, %xmm0
; X86-SSE42-NEXT: movd %xmm0, %eax
; X86-SSE42-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-SSE42-NEXT: retl
;
; X86-AVX-LABEL: test_reduce_v16i16_v8i16:
; X86-AVX: ## %bb.0:
; X86-AVX-NEXT: vphminposuw %xmm0, %xmm0
; X86-AVX-NEXT: vmovd %xmm0, %eax
; X86-AVX-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-AVX-NEXT: vzeroupper
; X86-AVX-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v16i16_v8i16:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: movdqa %xmm0, %xmm1
; X64-SSE2-NEXT: psrld $16, %xmm1
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: movd %xmm0, %eax
; X64-SSE2-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v16i16_v8i16:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: phminposuw %xmm0, %xmm0
; X64-SSE42-NEXT: movd %xmm0, %eax
; X64-SSE42-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-SSE42-NEXT: retq
;
; X64-AVX-LABEL: test_reduce_v16i16_v8i16:
; X64-AVX: ## %bb.0:
; X64-AVX-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX-NEXT: vmovd %xmm0, %eax
; X64-AVX-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-AVX-NEXT: vzeroupper
; X64-AVX-NEXT: retq
%1 = shufflevector <16 x i16> %a0, <16 x i16> undef, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%2 = icmp ult <16 x i16> %a0, %1
%3 = select <16 x i1> %2, <16 x i16> %a0, <16 x i16> %1
%4 = shufflevector <16 x i16> %3, <16 x i16> undef, <16 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%5 = icmp ult <16 x i16> %3, %4
%6 = select <16 x i1> %5, <16 x i16> %3, <16 x i16> %4
%7 = shufflevector <16 x i16> %6, <16 x i16> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%8 = icmp ult <16 x i16> %6, %7
%9 = select <16 x i1> %8, <16 x i16> %6, <16 x i16> %7
%10 = extractelement <16 x i16> %9, i32 0
ret i16 %10
}
define i16 @test_reduce_v32i16_v8i16(<32 x i16> %a0) {
; X86-SSE2-LABEL: test_reduce_v32i16_v8i16:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
; X86-SSE2-NEXT: psrld $16, %xmm1
; X86-SSE2-NEXT: movdqa %xmm0, %xmm2
; X86-SSE2-NEXT: psubusw %xmm1, %xmm2
; X86-SSE2-NEXT: psubw %xmm2, %xmm0
; X86-SSE2-NEXT: movd %xmm0, %eax
; X86-SSE2-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v32i16_v8i16:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: phminposuw %xmm0, %xmm0
; X86-SSE42-NEXT: movd %xmm0, %eax
; X86-SSE42-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-SSE42-NEXT: retl
;
; X86-AVX-LABEL: test_reduce_v32i16_v8i16:
; X86-AVX: ## %bb.0:
; X86-AVX-NEXT: vphminposuw %xmm0, %xmm0
; X86-AVX-NEXT: vmovd %xmm0, %eax
; X86-AVX-NEXT: ## kill: def $ax killed $ax killed $eax
; X86-AVX-NEXT: vzeroupper
; X86-AVX-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v32i16_v8i16:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: movdqa %xmm0, %xmm1
; X64-SSE2-NEXT: psrld $16, %xmm1
; X64-SSE2-NEXT: movdqa %xmm0, %xmm2
; X64-SSE2-NEXT: psubusw %xmm1, %xmm2
; X64-SSE2-NEXT: psubw %xmm2, %xmm0
; X64-SSE2-NEXT: movd %xmm0, %eax
; X64-SSE2-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v32i16_v8i16:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: phminposuw %xmm0, %xmm0
; X64-SSE42-NEXT: movd %xmm0, %eax
; X64-SSE42-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-SSE42-NEXT: retq
;
; X64-AVX-LABEL: test_reduce_v32i16_v8i16:
; X64-AVX: ## %bb.0:
; X64-AVX-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX-NEXT: vmovd %xmm0, %eax
; X64-AVX-NEXT: ## kill: def $ax killed $ax killed $eax
; X64-AVX-NEXT: vzeroupper
; X64-AVX-NEXT: retq
%1 = shufflevector <32 x i16> %a0, <32 x i16> undef, <32 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%2 = icmp ult <32 x i16> %a0, %1
%3 = select <32 x i1> %2, <32 x i16> %a0, <32 x i16> %1
%4 = shufflevector <32 x i16> %3, <32 x i16> undef, <32 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%5 = icmp ult <32 x i16> %3, %4
%6 = select <32 x i1> %5, <32 x i16> %3, <32 x i16> %4
%7 = shufflevector <32 x i16> %6, <32 x i16> undef, <32 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%8 = icmp ult <32 x i16> %6, %7
%9 = select <32 x i1> %8, <32 x i16> %6, <32 x i16> %7
%10 = extractelement <32 x i16> %9, i32 0
ret i16 %10
}
define i8 @test_reduce_v32i8_v16i8(<32 x i8> %a0) {
; X86-SSE2-LABEL: test_reduce_v32i8_v16i8:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-SSE2-NEXT: pminub %xmm0, %xmm1
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
; X86-SSE2-NEXT: pminub %xmm1, %xmm0
; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
; X86-SSE2-NEXT: psrld $16, %xmm1
; X86-SSE2-NEXT: pminub %xmm0, %xmm1
; X86-SSE2-NEXT: movdqa %xmm1, %xmm0
; X86-SSE2-NEXT: psrlw $8, %xmm0
; X86-SSE2-NEXT: pminub %xmm1, %xmm0
; X86-SSE2-NEXT: movd %xmm0, %eax
; X86-SSE2-NEXT: ## kill: def $al killed $al killed $eax
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v32i8_v16i8:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: movdqa %xmm0, %xmm1
; X86-SSE42-NEXT: psrlw $8, %xmm1
; X86-SSE42-NEXT: pminub %xmm0, %xmm1
; X86-SSE42-NEXT: phminposuw %xmm1, %xmm0
; X86-SSE42-NEXT: movd %xmm0, %eax
; X86-SSE42-NEXT: ## kill: def $al killed $al killed $eax
; X86-SSE42-NEXT: retl
;
; X86-AVX-LABEL: test_reduce_v32i8_v16i8:
; X86-AVX: ## %bb.0:
; X86-AVX-NEXT: vpsrlw $8, %xmm0, %xmm1
; X86-AVX-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X86-AVX-NEXT: vphminposuw %xmm0, %xmm0
; X86-AVX-NEXT: vmovd %xmm0, %eax
; X86-AVX-NEXT: ## kill: def $al killed $al killed $eax
; X86-AVX-NEXT: vzeroupper
; X86-AVX-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v32i8_v16i8:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-SSE2-NEXT: pminub %xmm0, %xmm1
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
; X64-SSE2-NEXT: pminub %xmm1, %xmm0
; X64-SSE2-NEXT: movdqa %xmm0, %xmm1
; X64-SSE2-NEXT: psrld $16, %xmm1
; X64-SSE2-NEXT: pminub %xmm0, %xmm1
; X64-SSE2-NEXT: movdqa %xmm1, %xmm0
; X64-SSE2-NEXT: psrlw $8, %xmm0
; X64-SSE2-NEXT: pminub %xmm1, %xmm0
; X64-SSE2-NEXT: movd %xmm0, %eax
; X64-SSE2-NEXT: ## kill: def $al killed $al killed $eax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v32i8_v16i8:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: movdqa %xmm0, %xmm1
; X64-SSE42-NEXT: psrlw $8, %xmm1
; X64-SSE42-NEXT: pminub %xmm0, %xmm1
; X64-SSE42-NEXT: phminposuw %xmm1, %xmm0
; X64-SSE42-NEXT: movd %xmm0, %eax
; X64-SSE42-NEXT: ## kill: def $al killed $al killed $eax
; X64-SSE42-NEXT: retq
;
; X64-AVX-LABEL: test_reduce_v32i8_v16i8:
; X64-AVX: ## %bb.0:
; X64-AVX-NEXT: vpsrlw $8, %xmm0, %xmm1
; X64-AVX-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X64-AVX-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX-NEXT: vmovd %xmm0, %eax
; X64-AVX-NEXT: ## kill: def $al killed $al killed $eax
; X64-AVX-NEXT: vzeroupper
; X64-AVX-NEXT: retq
%1 = shufflevector <32 x i8> %a0, <32 x i8> undef, <32 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%2 = icmp ult <32 x i8> %a0, %1
%3 = select <32 x i1> %2, <32 x i8> %a0, <32 x i8> %1
%4 = shufflevector <32 x i8> %3, <32 x i8> undef, <32 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%5 = icmp ult <32 x i8> %3, %4
%6 = select <32 x i1> %5, <32 x i8> %3, <32 x i8> %4
%7 = shufflevector <32 x i8> %6, <32 x i8> undef, <32 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%8 = icmp ult <32 x i8> %6, %7
%9 = select <32 x i1> %8, <32 x i8> %6, <32 x i8> %7
%10 = shufflevector <32 x i8> %9, <32 x i8> undef, <32 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%11 = icmp ult <32 x i8> %9, %10
%12 = select <32 x i1> %11, <32 x i8> %9, <32 x i8> %10
%13 = extractelement <32 x i8> %12, i32 0
ret i8 %13
}
define i8 @test_reduce_v64i8_v16i8(<64 x i8> %a0) {
; X86-SSE2-LABEL: test_reduce_v64i8_v16i8:
; X86-SSE2: ## %bb.0:
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X86-SSE2-NEXT: pminub %xmm0, %xmm1
; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
; X86-SSE2-NEXT: pminub %xmm1, %xmm0
; X86-SSE2-NEXT: movdqa %xmm0, %xmm1
; X86-SSE2-NEXT: psrld $16, %xmm1
; X86-SSE2-NEXT: pminub %xmm0, %xmm1
; X86-SSE2-NEXT: movdqa %xmm1, %xmm0
; X86-SSE2-NEXT: psrlw $8, %xmm0
; X86-SSE2-NEXT: pminub %xmm1, %xmm0
; X86-SSE2-NEXT: movd %xmm0, %eax
; X86-SSE2-NEXT: ## kill: def $al killed $al killed $eax
; X86-SSE2-NEXT: retl
;
; X86-SSE42-LABEL: test_reduce_v64i8_v16i8:
; X86-SSE42: ## %bb.0:
; X86-SSE42-NEXT: movdqa %xmm0, %xmm1
; X86-SSE42-NEXT: psrlw $8, %xmm1
; X86-SSE42-NEXT: pminub %xmm0, %xmm1
; X86-SSE42-NEXT: phminposuw %xmm1, %xmm0
; X86-SSE42-NEXT: movd %xmm0, %eax
; X86-SSE42-NEXT: ## kill: def $al killed $al killed $eax
; X86-SSE42-NEXT: retl
;
; X86-AVX-LABEL: test_reduce_v64i8_v16i8:
; X86-AVX: ## %bb.0:
; X86-AVX-NEXT: vpsrlw $8, %xmm0, %xmm1
; X86-AVX-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X86-AVX-NEXT: vphminposuw %xmm0, %xmm0
; X86-AVX-NEXT: vmovd %xmm0, %eax
; X86-AVX-NEXT: ## kill: def $al killed $al killed $eax
; X86-AVX-NEXT: vzeroupper
; X86-AVX-NEXT: retl
;
; X64-SSE2-LABEL: test_reduce_v64i8_v16i8:
; X64-SSE2: ## %bb.0:
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3]
; X64-SSE2-NEXT: pminub %xmm0, %xmm1
; X64-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,1,1]
; X64-SSE2-NEXT: pminub %xmm1, %xmm0
; X64-SSE2-NEXT: movdqa %xmm0, %xmm1
; X64-SSE2-NEXT: psrld $16, %xmm1
; X64-SSE2-NEXT: pminub %xmm0, %xmm1
; X64-SSE2-NEXT: movdqa %xmm1, %xmm0
; X64-SSE2-NEXT: psrlw $8, %xmm0
; X64-SSE2-NEXT: pminub %xmm1, %xmm0
; X64-SSE2-NEXT: movd %xmm0, %eax
; X64-SSE2-NEXT: ## kill: def $al killed $al killed $eax
; X64-SSE2-NEXT: retq
;
; X64-SSE42-LABEL: test_reduce_v64i8_v16i8:
; X64-SSE42: ## %bb.0:
; X64-SSE42-NEXT: movdqa %xmm0, %xmm1
; X64-SSE42-NEXT: psrlw $8, %xmm1
; X64-SSE42-NEXT: pminub %xmm0, %xmm1
; X64-SSE42-NEXT: phminposuw %xmm1, %xmm0
; X64-SSE42-NEXT: movd %xmm0, %eax
; X64-SSE42-NEXT: ## kill: def $al killed $al killed $eax
; X64-SSE42-NEXT: retq
;
; X64-AVX-LABEL: test_reduce_v64i8_v16i8:
; X64-AVX: ## %bb.0:
; X64-AVX-NEXT: vpsrlw $8, %xmm0, %xmm1
; X64-AVX-NEXT: vpminub %xmm1, %xmm0, %xmm0
; X64-AVX-NEXT: vphminposuw %xmm0, %xmm0
; X64-AVX-NEXT: vmovd %xmm0, %eax
; X64-AVX-NEXT: ## kill: def $al killed $al killed $eax
; X64-AVX-NEXT: vzeroupper
; X64-AVX-NEXT: retq
%1 = shufflevector <64 x i8> %a0, <64 x i8> undef, <64 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%2 = icmp ult <64 x i8> %a0, %1
%3 = select <64 x i1> %2, <64 x i8> %a0, <64 x i8> %1
%4 = shufflevector <64 x i8> %3, <64 x i8> undef, <64 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%5 = icmp ult <64 x i8> %3, %4
%6 = select <64 x i1> %5, <64 x i8> %3, <64 x i8> %4
%7 = shufflevector <64 x i8> %6, <64 x i8> undef, <64 x i32> <i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%8 = icmp ult <64 x i8> %6, %7
%9 = select <64 x i1> %8, <64 x i8> %6, <64 x i8> %7
%10 = shufflevector <64 x i8> %9, <64 x i8> undef, <64 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
%11 = icmp ult <64 x i8> %9, %10
%12 = select <64 x i1> %11, <64 x i8> %9, <64 x i8> %10
%13 = extractelement <64 x i8> %12, i32 0
ret i8 %13
}
Reference in New Issue View Git Blame Copy Permalink