[X86][CGP] Reduce memcmp() expansion to 2 load pairs (PR33914)

D35067/rL308322 attempted to support up to 4 load pairs for memcmp inlining which resulted in regressions for some optimized libc memcmp implementations (PR33914).

Until we can match these more optimal cases, this patch reduces the memcmp expansion to a maximum of 2 load pairs (which matches what we do for -Os).

This patch should be considered for the 5.0.0 release branch as well

Differential Revision: https://reviews.llvm.org/D35830

llvm-svn: 308986

1 files changed, 122 insertions, 288 deletions

diff --git a/llvm/test/CodeGen/X86/memcmp.ll b/llvm/test/CodeGen/X86/memcmp.ll
index d15df314507..889f6a74bf7 100644
--- a/llvm/test/CodeGen/X86/memcmp.ll
+++ b/llvm/test/CodeGen/X86/memcmp.ll
@@ -475,25 +475,14 @@ define i1 @length8_eq_const(i8* %X) nounwind {
 
 define i1 @length12_eq(i8* %X, i8* %Y) nounwind {
 ; X86-LABEL: length12_eq:
-; X86:       # BB#0: # %loadbb
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    movl (%ecx), %edx
-; X86-NEXT:    cmpl (%eax), %edx
-; X86-NEXT:    jne .LBB14_1
-; X86-NEXT:  # BB#2: # %loadbb1
-; X86-NEXT:    movl 4(%ecx), %edx
-; X86-NEXT:    cmpl 4(%eax), %edx
-; X86-NEXT:    jne .LBB14_1
-; X86-NEXT:  # BB#3: # %loadbb2
-; X86-NEXT:    movl 8(%ecx), %edx
-; X86-NEXT:    xorl %ecx, %ecx
-; X86-NEXT:    cmpl 8(%eax), %edx
-; X86-NEXT:    je .LBB14_4
-; X86-NEXT:  .LBB14_1: # %res_block
-; X86-NEXT:    movl $1, %ecx
-; X86-NEXT:  .LBB14_4: # %endblock
-; X86-NEXT:    testl %ecx, %ecx
+; X86:       # BB#0:
+; X86-NEXT:    pushl $0
+; X86-NEXT:    pushl $12
+; X86-NEXT:    pushl {{[0-9]+}}(%esp)
+; X86-NEXT:    pushl {{[0-9]+}}(%esp)
+; X86-NEXT:    calll memcmp
+; X86-NEXT:    addl $16, %esp
+; X86-NEXT:    testl %eax, %eax
 ; X86-NEXT:    setne %al
 ; X86-NEXT:    retl
 ;
@@ -520,40 +509,13 @@ define i1 @length12_eq(i8* %X, i8* %Y) nounwind {
 
 define i32 @length12(i8* %X, i8* %Y) nounwind {
 ; X86-LABEL: length12:
-; X86:       # BB#0: # %loadbb
-; X86-NEXT:    pushl %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %esi
-; X86-NEXT:    movl (%esi), %ecx
-; X86-NEXT:    movl (%eax), %edx
-; X86-NEXT:    bswapl %ecx
-; X86-NEXT:    bswapl %edx
-; X86-NEXT:    cmpl %edx, %ecx
-; X86-NEXT:    jne .LBB15_1
-; X86-NEXT:  # BB#2: # %loadbb1
-; X86-NEXT:    movl 4(%esi), %ecx
-; X86-NEXT:    movl 4(%eax), %edx
-; X86-NEXT:    bswapl %ecx
-; X86-NEXT:    bswapl %edx
-; X86-NEXT:    cmpl %edx, %ecx
-; X86-NEXT:    jne .LBB15_1
-; X86-NEXT:  # BB#3: # %loadbb2
-; X86-NEXT:    movl 8(%esi), %ecx
-; X86-NEXT:    movl 8(%eax), %edx
-; X86-NEXT:    bswapl %ecx
-; X86-NEXT:    bswapl %edx
-; X86-NEXT:    xorl %eax, %eax
-; X86-NEXT:    cmpl %edx, %ecx
-; X86-NEXT:    jne .LBB15_1
-; X86-NEXT:  # BB#4: # %endblock
-; X86-NEXT:    popl %esi
-; X86-NEXT:    retl
-; X86-NEXT:  .LBB15_1: # %res_block
-; X86-NEXT:    cmpl %edx, %ecx
-; X86-NEXT:    movl $-1, %ecx
-; X86-NEXT:    movl $1, %eax
-; X86-NEXT:    cmovbl %ecx, %eax
-; X86-NEXT:    popl %esi
+; X86:       # BB#0:
+; X86-NEXT:    pushl $0
+; X86-NEXT:    pushl $12
+; X86-NEXT:    pushl {{[0-9]+}}(%esp)
+; X86-NEXT:    pushl {{[0-9]+}}(%esp)
+; X86-NEXT:    calll memcmp
+; X86-NEXT:    addl $16, %esp
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: length12:
@@ -588,47 +550,13 @@ define i32 @length12(i8* %X, i8* %Y) nounwind {
 
 define i32 @length16(i8* %X, i8* %Y) nounwind {
 ; X86-LABEL: length16:
-; X86:       # BB#0: # %loadbb
-; X86-NEXT:    pushl %esi
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %esi
-; X86-NEXT:    movl (%esi), %ecx
-; X86-NEXT:    movl (%eax), %edx
-; X86-NEXT:    bswapl %ecx
-; X86-NEXT:    bswapl %edx
-; X86-NEXT:    cmpl %edx, %ecx
-; X86-NEXT:    jne .LBB16_1
-; X86-NEXT:  # BB#2: # %loadbb1
-; X86-NEXT:    movl 4(%esi), %ecx
-; X86-NEXT:    movl 4(%eax), %edx
-; X86-NEXT:    bswapl %ecx
-; X86-NEXT:    bswapl %edx
-; X86-NEXT:    cmpl %edx, %ecx
-; X86-NEXT:    jne .LBB16_1
-; X86-NEXT:  # BB#3: # %loadbb2
-; X86-NEXT:    movl 8(%esi), %ecx
-; X86-NEXT:    movl 8(%eax), %edx
-; X86-NEXT:    bswapl %ecx
-; X86-NEXT:    bswapl %edx
-; X86-NEXT:    cmpl %edx, %ecx
-; X86-NEXT:    jne .LBB16_1
-; X86-NEXT:  # BB#4: # %loadbb3
-; X86-NEXT:    movl 12(%esi), %ecx
-; X86-NEXT:    movl 12(%eax), %edx
-; X86-NEXT:    bswapl %ecx
-; X86-NEXT:    bswapl %edx
-; X86-NEXT:    xorl %eax, %eax
-; X86-NEXT:    cmpl %edx, %ecx
-; X86-NEXT:    jne .LBB16_1
-; X86-NEXT:  # BB#5: # %endblock
-; X86-NEXT:    popl %esi
-; X86-NEXT:    retl
-; X86-NEXT:  .LBB16_1: # %res_block
-; X86-NEXT:    cmpl %edx, %ecx
-; X86-NEXT:    movl $-1, %ecx
-; X86-NEXT:    movl $1, %eax
-; X86-NEXT:    cmovbl %ecx, %eax
-; X86-NEXT:    popl %esi
+; X86:       # BB#0:
+; X86-NEXT:    pushl $0
+; X86-NEXT:    pushl $16
+; X86-NEXT:    pushl {{[0-9]+}}(%esp)
+; X86-NEXT:    pushl {{[0-9]+}}(%esp)
+; X86-NEXT:    calll memcmp
+; X86-NEXT:    addl $16, %esp
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: length16:
@@ -660,32 +588,29 @@ define i32 @length16(i8* %X, i8* %Y) nounwind {
 }
 
 define i1 @length16_eq(i8* %x, i8* %y) nounwind {
-; X86-LABEL: length16_eq:
-; X86:       # BB#0: # %loadbb
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %ecx
-; X86-NEXT:    movl (%ecx), %edx
-; X86-NEXT:    cmpl (%eax), %edx
-; X86-NEXT:    jne .LBB17_1
-; X86-NEXT:  # BB#2: # %loadbb1
-; X86-NEXT:    movl 4(%ecx), %edx
-; X86-NEXT:    cmpl 4(%eax), %edx
-; X86-NEXT:    jne .LBB17_1
-; X86-NEXT:  # BB#3: # %loadbb2
-; X86-NEXT:    movl 8(%ecx), %edx
-; X86-NEXT:    cmpl 8(%eax), %edx
-; X86-NEXT:    jne .LBB17_1
-; X86-NEXT:  # BB#4: # %loadbb3
-; X86-NEXT:    movl 12(%ecx), %edx
-; X86-NEXT:    xorl %ecx, %ecx
-; X86-NEXT:    cmpl 12(%eax), %edx
-; X86-NEXT:    je .LBB17_5
-; X86-NEXT:  .LBB17_1: # %res_block
-; X86-NEXT:    movl $1, %ecx
-; X86-NEXT:  .LBB17_5: # %endblock
-; X86-NEXT:    testl %ecx, %ecx
-; X86-NEXT:    setne %al
-; X86-NEXT:    retl
+; X86-NOSSE-LABEL: length16_eq:
+; X86-NOSSE:       # BB#0:
+; X86-NOSSE-NEXT:    pushl $0
+; X86-NOSSE-NEXT:    pushl $16
+; X86-NOSSE-NEXT:    pushl {{[0-9]+}}(%esp)
+; X86-NOSSE-NEXT:    pushl {{[0-9]+}}(%esp)
+; X86-NOSSE-NEXT:    calll memcmp
+; X86-NOSSE-NEXT:    addl $16, %esp
+; X86-NOSSE-NEXT:    testl %eax, %eax
+; X86-NOSSE-NEXT:    setne %al
+; X86-NOSSE-NEXT:    retl
+;
+; X86-SSE2-LABEL: length16_eq:
+; X86-SSE2:       # BB#0:
+; X86-SSE2-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X86-SSE2-NEXT:    movl {{[0-9]+}}(%esp), %ecx
+; X86-SSE2-NEXT:    movdqu (%ecx), %xmm0
+; X86-SSE2-NEXT:    movdqu (%eax), %xmm1
+; X86-SSE2-NEXT:    pcmpeqb %xmm0, %xmm1
+; X86-SSE2-NEXT:    pmovmskb %xmm1, %eax
+; X86-SSE2-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
+; X86-SSE2-NEXT:    setne %al
+; X86-SSE2-NEXT:    retl
 ;
 ; X64-LABEL: length16_eq:
 ; X64:       # BB#0: # %loadbb
@@ -709,27 +634,27 @@ define i1 @length16_eq(i8* %x, i8* %y) nounwind {
 }
 
 define i1 @length16_eq_const(i8* %X) nounwind {
-; X86-LABEL: length16_eq_const:
-; X86:       # BB#0: # %loadbb
-; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
-; X86-NEXT:    cmpl $858927408, (%eax) # imm = 0x33323130
-; X86-NEXT:    jne .LBB18_1
-; X86-NEXT:  # BB#2: # %loadbb1
-; X86-NEXT:    cmpl $926299444, 4(%eax) # imm = 0x37363534
-; X86-NEXT:    jne .LBB18_1
-; X86-NEXT:  # BB#3: # %loadbb2
-; X86-NEXT:    cmpl $825243960, 8(%eax) # imm = 0x31303938
-; X86-NEXT:    jne .LBB18_1
-; X86-NEXT:  # BB#4: # %loadbb3
-; X86-NEXT:    xorl %ecx, %ecx
-; X86-NEXT:    cmpl $892613426, 12(%eax) # imm = 0x35343332
-; X86-NEXT:    je .LBB18_5
-; X86-NEXT:  .LBB18_1: # %res_block
-; X86-NEXT:    movl $1, %ecx
-; X86-NEXT:  .LBB18_5: # %endblock
-; X86-NEXT:    testl %ecx, %ecx
-; X86-NEXT:    sete %al
-; X86-NEXT:    retl
+; X86-NOSSE-LABEL: length16_eq_const:
+; X86-NOSSE:       # BB#0:
+; X86-NOSSE-NEXT:    pushl $0
+; X86-NOSSE-NEXT:    pushl $16
+; X86-NOSSE-NEXT:    pushl $.L.str
+; X86-NOSSE-NEXT:    pushl {{[0-9]+}}(%esp)
+; X86-NOSSE-NEXT:    calll memcmp
+; X86-NOSSE-NEXT:    addl $16, %esp
+; X86-NOSSE-NEXT:    testl %eax, %eax
+; X86-NOSSE-NEXT:    sete %al
+; X86-NOSSE-NEXT:    retl
+;
+; X86-SSE2-LABEL: length16_eq_const:
+; X86-SSE2:       # BB#0:
+; X86-SSE2-NEXT:    movl {{[0-9]+}}(%esp), %eax
+; X86-SSE2-NEXT:    movdqu (%eax), %xmm0
+; X86-SSE2-NEXT:    pcmpeqb {{\.LCPI.*}}, %xmm0
+; X86-SSE2-NEXT:    pmovmskb %xmm0, %eax
+; X86-SSE2-NEXT:    cmpl $65535, %eax # imm = 0xFFFF
+; X86-SSE2-NEXT:    sete %al
+; X86-SSE2-NEXT:    retl
 ;
 ; X64-LABEL: length16_eq_const:
 ; X64:       # BB#0: # %loadbb
@@ -766,36 +691,9 @@ define i32 @length24(i8* %X, i8* %Y) nounwind {
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: length24:
-; X64:       # BB#0: # %loadbb
-; X64-NEXT:    movq (%rdi), %rcx
-; X64-NEXT:    movq (%rsi), %rdx
-; X64-NEXT:    bswapq %rcx
-; X64-NEXT:    bswapq %rdx
-; X64-NEXT:    cmpq %rdx, %rcx
-; X64-NEXT:    jne .LBB19_1
-; X64-NEXT:  # BB#2: # %loadbb1
-; X64-NEXT:    movq 8(%rdi), %rcx
-; X64-NEXT:    movq 8(%rsi), %rdx
-; X64-NEXT:    bswapq %rcx
-; X64-NEXT:    bswapq %rdx
-; X64-NEXT:    cmpq %rdx, %rcx
-; X64-NEXT:    jne .LBB19_1
-; X64-NEXT:  # BB#3: # %loadbb2
-; X64-NEXT:    movq 16(%rdi), %rcx
-; X64-NEXT:    movq 16(%rsi), %rdx
-; X64-NEXT:    bswapq %rcx
-; X64-NEXT:    bswapq %rdx
-; X64-NEXT:    xorl %eax, %eax
-; X64-NEXT:    cmpq %rdx, %rcx
-; X64-NEXT:    jne .LBB19_1
-; X64-NEXT:  # BB#4: # %endblock
-; X64-NEXT:    retq
-; X64-NEXT:  .LBB19_1: # %res_block
-; X64-NEXT:    cmpq %rdx, %rcx
-; X64-NEXT:    movl $-1, %ecx
-; X64-NEXT:    movl $1, %eax
-; X64-NEXT:    cmovbl %ecx, %eax
-; X64-NEXT:    retq
+; X64:       # BB#0:
+; X64-NEXT:    movl $24, %edx
+; X64-NEXT:    jmp memcmp # TAILCALL
   %m = tail call i32 @memcmp(i8* %X, i8* %Y, i64 24) nounwind
   ret i32 %m
 }
@@ -814,24 +712,13 @@ define i1 @length24_eq(i8* %x, i8* %y) nounwind {
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: length24_eq:
-; X64:       # BB#0: # %loadbb
-; X64-NEXT:    movq (%rdi), %rax
-; X64-NEXT:    cmpq (%rsi), %rax
-; X64-NEXT:    jne .LBB20_1
-; X64-NEXT:  # BB#2: # %loadbb1
-; X64-NEXT:    movq 8(%rdi), %rax
-; X64-NEXT:    cmpq 8(%rsi), %rax
-; X64-NEXT:    jne .LBB20_1
-; X64-NEXT:  # BB#3: # %loadbb2
-; X64-NEXT:    movq 16(%rdi), %rcx
-; X64-NEXT:    xorl %eax, %eax
-; X64-NEXT:    cmpq 16(%rsi), %rcx
-; X64-NEXT:    je .LBB20_4
-; X64-NEXT:  .LBB20_1: # %res_block
-; X64-NEXT:    movl $1, %eax
-; X64-NEXT:  .LBB20_4: # %endblock
+; X64:       # BB#0:
+; X64-NEXT:    pushq %rax
+; X64-NEXT:    movl $24, %edx
+; X64-NEXT:    callq memcmp
 ; X64-NEXT:    testl %eax, %eax
 ; X64-NEXT:    sete %al
+; X64-NEXT:    popq %rcx
 ; X64-NEXT:    retq
   %call = tail call i32 @memcmp(i8* %x, i8* %y, i64 24) nounwind
   %cmp = icmp eq i32 %call, 0
@@ -852,24 +739,14 @@ define i1 @length24_eq_const(i8* %X) nounwind {
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: length24_eq_const:
-; X64:       # BB#0: # %loadbb
-; X64-NEXT:    movabsq $3978425819141910832, %rax # imm = 0x3736353433323130
-; X64-NEXT:    cmpq %rax, (%rdi)
-; X64-NEXT:    jne .LBB21_1
-; X64-NEXT:  # BB#2: # %loadbb1
-; X64-NEXT:    movabsq $3833745473465760056, %rax # imm = 0x3534333231303938
-; X64-NEXT:    cmpq %rax, 8(%rdi)
-; X64-NEXT:    jne .LBB21_1
-; X64-NEXT:  # BB#3: # %loadbb2
-; X64-NEXT:    xorl %eax, %eax
-; X64-NEXT:    movabsq $3689065127958034230, %rcx # imm = 0x3332313039383736
-; X64-NEXT:    cmpq %rcx, 16(%rdi)
-; X64-NEXT:    je .LBB21_4
-; X64-NEXT:  .LBB21_1: # %res_block
-; X64-NEXT:    movl $1, %eax
-; X64-NEXT:  .LBB21_4: # %endblock
+; X64:       # BB#0:
+; X64-NEXT:    pushq %rax
+; X64-NEXT:    movl $.L.str, %esi
+; X64-NEXT:    movl $24, %edx
+; X64-NEXT:    callq memcmp
 ; X64-NEXT:    testl %eax, %eax
 ; X64-NEXT:    setne %al
+; X64-NEXT:    popq %rcx
 ; X64-NEXT:    retq
   %m = tail call i32 @memcmp(i8* %X, i8* getelementptr inbounds ([65 x i8], [65 x i8]* @.str, i32 0, i32 0), i64 24) nounwind
   %c = icmp ne i32 %m, 0
@@ -888,43 +765,9 @@ define i32 @length32(i8* %X, i8* %Y) nounwind {
 ; X86-NEXT:    retl
 ;
 ; X64-LABEL: length32:
-; X64:       # BB#0: # %loadbb
-; X64-NEXT:    movq (%rdi), %rcx
-; X64-NEXT:    movq (%rsi), %rdx
-; X64-NEXT:    bswapq %rcx
-; X64-NEXT:    bswapq %rdx
-; X64-NEXT:    cmpq %rdx, %rcx
-; X64-NEXT:    jne .LBB22_1
-; X64-NEXT:  # BB#2: # %loadbb1
-; X64-NEXT:    movq 8(%rdi), %rcx
-; X64-NEXT:    movq 8(%rsi), %rdx
-; X64-NEXT:    bswapq %rcx
-; X64-NEXT:    bswapq %rdx
-; X64-NEXT:    cmpq %rdx, %rcx
-; X64-NEXT:    jne .LBB22_1
-; X64-NEXT:  # BB#3: # %loadbb2
-; X64-NEXT:    movq 16(%rdi), %rcx
-; X64-NEXT:    movq 16(%rsi), %rdx
-; X64-NEXT:    bswapq %rcx
-; X64-NEXT:    bswapq %rdx
-; X64-NEXT:    cmpq %rdx, %rcx
-; X64-NEXT:    jne .LBB22_1
-; X64-NEXT:  # BB#4: # %loadbb3
-; X64-NEXT:    movq 24(%rdi), %rcx
-; X64-NEXT:    movq 24(%rsi), %rdx
-; X64-NEXT:    bswapq %rcx
-; X64-NEXT:    bswapq %rdx
-; X64-NEXT:    xorl %eax, %eax
-; X64-NEXT:    cmpq %rdx, %rcx
-; X64-NEXT:    jne .LBB22_1
-; X64-NEXT:  # BB#5: # %endblock
-; X64-NEXT:    retq
-; X64-NEXT:  .LBB22_1: # %res_block
-; X64-NEXT:    cmpq %rdx, %rcx
-; X64-NEXT:    movl $-1, %ecx
-; X64-NEXT:    movl $1, %eax
-; X64-NEXT:    cmovbl %ecx, %eax
-; X64-NEXT:    retq
+; X64:       # BB#0:
+; X64-NEXT:    movl $32, %edx
+; X64-NEXT:    jmp memcmp # TAILCALL
   %m = tail call i32 @memcmp(i8* %X, i8* %Y, i64 32) nounwind
   ret i32 %m
 }
@@ -944,30 +787,25 @@ define i1 @length32_eq(i8* %x, i8* %y) nounwind {
 ; X86-NEXT:    sete %al
 ; X86-NEXT:    retl
 ;
-; X64-LABEL: length32_eq:
-; X64:       # BB#0: # %loadbb
-; X64-NEXT:    movq (%rdi), %rax
-; X64-NEXT:    cmpq (%rsi), %rax
-; X64-NEXT:    jne .LBB23_1
-; X64-NEXT:  # BB#2: # %loadbb1
-; X64-NEXT:    movq 8(%rdi), %rax
-; X64-NEXT:    cmpq 8(%rsi), %rax
-; X64-NEXT:    jne .LBB23_1
-; X64-NEXT:  # BB#3: # %loadbb2
-; X64-NEXT:    movq 16(%rdi), %rax
-; X64-NEXT:    cmpq 16(%rsi), %rax
-; X64-NEXT:    jne .LBB23_1
-; X64-NEXT:  # BB#4: # %loadbb3
-; X64-NEXT:    movq 24(%rdi), %rcx
-; X64-NEXT:    xorl %eax, %eax
-; X64-NEXT:    cmpq 24(%rsi), %rcx
-; X64-NEXT:    je .LBB23_5
-; X64-NEXT:  .LBB23_1: # %res_block
-; X64-NEXT:    movl $1, %eax
-; X64-NEXT:  .LBB23_5: # %endblock
-; X64-NEXT:    testl %eax, %eax
-; X64-NEXT:    sete %al
-; X64-NEXT:    retq
+; X64-SSE2-LABEL: length32_eq:
+; X64-SSE2:       # BB#0:
+; X64-SSE2-NEXT:    pushq %rax
+; X64-SSE2-NEXT:    movl $32, %edx
+; X64-SSE2-NEXT:    callq memcmp
+; X64-SSE2-NEXT:    testl %eax, %eax
+; X64-SSE2-NEXT:    sete %al
+; X64-SSE2-NEXT:    popq %rcx
+; X64-SSE2-NEXT:    retq
+;
+; X64-AVX2-LABEL: length32_eq:
+; X64-AVX2:       # BB#0:
+; X64-AVX2-NEXT:    vmovdqu (%rdi), %ymm0
+; X64-AVX2-NEXT:    vpcmpeqb (%rsi), %ymm0, %ymm0
+; X64-AVX2-NEXT:    vpmovmskb %ymm0, %eax
+; X64-AVX2-NEXT:    cmpl $-1, %eax
+; X64-AVX2-NEXT:    sete %al
+; X64-AVX2-NEXT:    vzeroupper
+; X64-AVX2-NEXT:    retq
   %call = tail call i32 @memcmp(i8* %x, i8* %y, i64 32) nounwind
   %cmp = icmp eq i32 %call, 0
   ret i1 %cmp
@@ -986,30 +824,26 @@ define i1 @length32_eq_const(i8* %X) nounwind {
 ; X86-NEXT:    setne %al
 ; X86-NEXT:    retl
 ;
-; X64-LABEL: length32_eq_const:
-; X64:       # BB#0: # %loadbb
-; X64-NEXT:    movabsq $3978425819141910832, %rax # imm = 0x3736353433323130
-; X64-NEXT:    cmpq %rax, (%rdi)
-; X64-NEXT:    jne .LBB24_1
-; X64-NEXT:  # BB#2: # %loadbb1
-; X64-NEXT:    movabsq $3833745473465760056, %rax # imm = 0x3534333231303938
-; X64-NEXT:    cmpq %rax, 8(%rdi)
-; X64-NEXT:    jne .LBB24_1
-; X64-NEXT:  # BB#3: # %loadbb2
-; X64-NEXT:    movabsq $3689065127958034230, %rax # imm = 0x3332313039383736
-; X64-NEXT:    cmpq %rax, 16(%rdi)
-; X64-NEXT:    jne .LBB24_1
-; X64-NEXT:  # BB#4: # %loadbb3
-; X64-NEXT:    xorl %eax, %eax
-; X64-NEXT:    movabsq $3544395820347831604, %rcx # imm = 0x3130393837363534
-; X64-NEXT:    cmpq %rcx, 24(%rdi)
-; X64-NEXT:    je .LBB24_5
-; X64-NEXT:  .LBB24_1: # %res_block
-; X64-NEXT:    movl $1, %eax
-; X64-NEXT:  .LBB24_5: # %endblock
-; X64-NEXT:    testl %eax, %eax
-; X64-NEXT:    setne %al
-; X64-NEXT:    retq
+; X64-SSE2-LABEL: length32_eq_const:
+; X64-SSE2:       # BB#0:
+; X64-SSE2-NEXT:    pushq %rax
+; X64-SSE2-NEXT:    movl $.L.str, %esi
+; X64-SSE2-NEXT:    movl $32, %edx
+; X64-SSE2-NEXT:    callq memcmp
+; X64-SSE2-NEXT:    testl %eax, %eax
+; X64-SSE2-NEXT:    setne %al
+; X64-SSE2-NEXT:    popq %rcx
+; X64-SSE2-NEXT:    retq
+;
+; X64-AVX2-LABEL: length32_eq_const:
+; X64-AVX2:       # BB#0:
+; X64-AVX2-NEXT:    vmovdqu (%rdi), %ymm0
+; X64-AVX2-NEXT:    vpcmpeqb {{.*}}(%rip), %ymm0, %ymm0
+; X64-AVX2-NEXT:    vpmovmskb %ymm0, %eax
+; X64-AVX2-NEXT:    cmpl $-1, %eax
+; X64-AVX2-NEXT:    setne %al
+; X64-AVX2-NEXT:    vzeroupper
+; X64-AVX2-NEXT:    retq
   %m = tail call i32 @memcmp(i8* %X, i8* getelementptr inbounds ([65 x i8], [65 x i8]* @.str, i32 0, i32 0), i64 32) nounwind
   %c = icmp ne i32 %m, 0
   ret i1 %c