diff options
Diffstat (limited to 'llvm/test/CodeGen/X86')
| -rw-r--r-- | llvm/test/CodeGen/X86/bittest-big-integer.ll | 1005 | ||||
| -rw-r--r-- | llvm/test/CodeGen/X86/isel-llvm.sincos.ll | 133 | ||||
| -rw-r--r-- | llvm/test/CodeGen/X86/llvm.sincos.vec.ll | 404 |
3 files changed, 1349 insertions, 193 deletions
diff --git a/llvm/test/CodeGen/X86/bittest-big-integer.ll b/llvm/test/CodeGen/X86/bittest-big-integer.ll index dffe900..c311ab8 100644 --- a/llvm/test/CodeGen/X86/bittest-big-integer.ll +++ b/llvm/test/CodeGen/X86/bittest-big-integer.ll @@ -356,20 +356,41 @@ define i1 @init_eq_i64(ptr %word, i32 %position, i1 zeroext %value) nounwind { ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl %ecx, %esi -; X86-NEXT: andl $32, %esi -; X86-NEXT: shrl $3, %esi -; X86-NEXT: movl (%edx,%esi), %edi -; X86-NEXT: btl %ecx, %edi +; X86-NEXT: movl $1, %edx +; X86-NEXT: xorl %esi, %esi +; X86-NEXT: shldl %cl, %edx, %esi +; X86-NEXT: shll %cl, %edx +; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax +; X86-NEXT: xorl %edi, %edi +; X86-NEXT: shldl %cl, %eax, %edi +; X86-NEXT: shll %cl, %eax +; X86-NEXT: testb $32, %cl +; X86-NEXT: je .LBB9_2 +; X86-NEXT: # %bb.1: +; X86-NEXT: movl %edx, %esi +; X86-NEXT: movl $0, %edx +; X86-NEXT: .LBB9_2: +; X86-NEXT: movl {{[0-9]+}}(%esp), %ebx +; X86-NEXT: notl %esi +; X86-NEXT: notl %edx +; X86-NEXT: je .LBB9_4 +; X86-NEXT: # %bb.3: +; X86-NEXT: movl %eax, %edi +; X86-NEXT: xorl %eax, %eax +; X86-NEXT: .LBB9_4: +; X86-NEXT: andl 4(%ebx), %esi +; X86-NEXT: orl %edi, %esi +; X86-NEXT: andl (%ebx), %edx +; X86-NEXT: orl %eax, %edx +; X86-NEXT: movl %ecx, %eax +; X86-NEXT: andl $32, %eax +; X86-NEXT: shrl $3, %eax +; X86-NEXT: movl (%ebx,%eax), %eax +; X86-NEXT: btl %ecx, %eax ; X86-NEXT: setae %al -; X86-NEXT: btrl %ecx, %edi -; X86-NEXT: movzbl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: # kill: def $cl killed $cl killed $ecx -; X86-NEXT: shll %cl, %ebx -; X86-NEXT: orl %edi, %ebx -; X86-NEXT: movl %ebx, (%edx,%esi) +; X86-NEXT: movl %edx, (%ebx) +; X86-NEXT: movl %esi, 4(%ebx) ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx @@ -579,55 +600,208 @@ define i1 @set_ne_i128(ptr %word, i32 %position) nounwind { define i1 @init_eq_i128(ptr %word, i32 %position, i1 zeroext %value) nounwind { ; X86-LABEL: init_eq_i128: ; X86: # %bb.0: +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl %ecx, %esi -; X86-NEXT: andl $96, %esi -; X86-NEXT: shrl $3, %esi -; X86-NEXT: movl (%edx,%esi), %edi -; X86-NEXT: btl %ecx, %edi -; X86-NEXT: setae %al -; X86-NEXT: btrl %ecx, %edi -; X86-NEXT: movzbl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: # kill: def $cl killed $cl killed $ecx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $96, %esp +; X86-NEXT: movl 12(%ebp), %ecx +; X86-NEXT: movzbl 16(%ebp), %ebx +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $1, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ecx, %eax +; X86-NEXT: shrb $3, %al +; X86-NEXT: andb $12, %al +; X86-NEXT: negb %al +; X86-NEXT: movsbl %al, %eax +; X86-NEXT: movl 64(%esp,%eax), %edx +; X86-NEXT: movl 68(%esp,%eax), %esi +; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl %eax, %esi +; X86-NEXT: movl %eax, (%esp) # 4-byte Spill +; X86-NEXT: movzbl %bl, %eax +; X86-NEXT: movl 72(%esp,%esi), %ebx +; X86-NEXT: movl 76(%esp,%esi), %esi +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: movl %ebx, %edi +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: shldl %cl, %eax, %edi +; X86-NEXT: shldl %cl, %ebx, %esi +; X86-NEXT: movl %edx, %ebx ; X86-NEXT: shll %cl, %ebx -; X86-NEXT: orl %edi, %ebx -; X86-NEXT: movl %ebx, (%edx,%esi) +; X86-NEXT: # kill: def $cl killed $cl killed $ecx +; X86-NEXT: shldl %cl, %edx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: notl %edi +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl (%esp), %ecx # 4-byte Reload +; X86-NEXT: movl 36(%esp,%ecx), %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 40(%esp,%ecx), %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 12(%ebp), %ecx +; X86-NEXT: # kill: def $cl killed $cl killed $ecx +; X86-NEXT: shldl %cl, %eax, %edx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: andl 8(%eax), %edi +; X86-NEXT: orl %edx, %edi +; X86-NEXT: notl %esi +; X86-NEXT: movl (%esp), %eax # 4-byte Reload +; X86-NEXT: movl 44(%esp,%eax), %eax +; X86-NEXT: movl 12(%ebp), %ecx +; X86-NEXT: # kill: def $cl killed $cl killed $ecx +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload +; X86-NEXT: shldl %cl, %edx, %eax +; X86-NEXT: movl 8(%ebp), %ecx +; X86-NEXT: andl 12(%ecx), %esi +; X86-NEXT: orl %eax, %esi +; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: notl %ebx +; X86-NEXT: movl (%esp), %eax # 4-byte Reload +; X86-NEXT: movl 32(%esp,%eax), %edx +; X86-NEXT: movl %edx, (%esp) # 4-byte Spill +; X86-NEXT: movl 12(%ebp), %ecx +; X86-NEXT: shll %cl, %edx +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: andl (%eax), %ebx +; X86-NEXT: orl %edx, %ebx +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload +; X86-NEXT: notl %edx +; X86-NEXT: # kill: def $cl killed $cl killed $ecx +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl (%esp), %esi # 4-byte Reload +; X86-NEXT: shldl %cl, %esi, %eax +; X86-NEXT: movl 8(%ebp), %ecx +; X86-NEXT: andl 4(%ecx), %edx +; X86-NEXT: orl %eax, %edx +; X86-NEXT: movl 12(%ebp), %esi +; X86-NEXT: movl %esi, %eax +; X86-NEXT: andl $96, %eax +; X86-NEXT: shrl $3, %eax +; X86-NEXT: movl (%ecx,%eax), %eax +; X86-NEXT: btl %esi, %eax +; X86-NEXT: movl %ecx, %eax +; X86-NEXT: movl %edi, 8(%ecx) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ecx # 4-byte Reload +; X86-NEXT: movl %ecx, 12(%eax) +; X86-NEXT: movl %ebx, (%eax) +; X86-NEXT: movl %edx, 4(%eax) +; X86-NEXT: setae %al +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; SSE-LABEL: init_eq_i128: ; SSE: # %bb.0: ; SSE-NEXT: movl %esi, %ecx -; SSE-NEXT: andl $96, %esi -; SSE-NEXT: shrl $3, %esi -; SSE-NEXT: movl (%rdi,%rsi), %r8d -; SSE-NEXT: btl %ecx, %r8d +; SSE-NEXT: movl $1, %esi +; SSE-NEXT: xorl %r8d, %r8d +; SSE-NEXT: shldq %cl, %rsi, %r8 +; SSE-NEXT: shlq %cl, %rsi +; SSE-NEXT: movl %edx, %eax +; SSE-NEXT: xorl %edx, %edx +; SSE-NEXT: shldq %cl, %rax, %rdx +; SSE-NEXT: shlq %cl, %rax +; SSE-NEXT: xorl %r9d, %r9d +; SSE-NEXT: testb $64, %cl +; SSE-NEXT: cmovneq %rsi, %r8 +; SSE-NEXT: cmovneq %r9, %rsi +; SSE-NEXT: notq %r8 +; SSE-NEXT: cmovneq %rax, %rdx +; SSE-NEXT: cmovneq %r9, %rax +; SSE-NEXT: notq %rsi +; SSE-NEXT: andq 8(%rdi), %r8 +; SSE-NEXT: orq %rdx, %r8 +; SSE-NEXT: andq (%rdi), %rsi +; SSE-NEXT: orq %rax, %rsi +; SSE-NEXT: movl %ecx, %eax +; SSE-NEXT: andl $96, %eax +; SSE-NEXT: shrl $3, %eax +; SSE-NEXT: movl (%rdi,%rax), %eax +; SSE-NEXT: btl %ecx, %eax ; SSE-NEXT: setae %al -; SSE-NEXT: shll %cl, %edx -; SSE-NEXT: btrl %ecx, %r8d -; SSE-NEXT: orl %r8d, %edx -; SSE-NEXT: movl %edx, (%rdi,%rsi) +; SSE-NEXT: movq %rsi, (%rdi) +; SSE-NEXT: movq %r8, 8(%rdi) ; SSE-NEXT: retq ; -; AVX-LABEL: init_eq_i128: -; AVX: # %bb.0: -; AVX-NEXT: movl %esi, %ecx -; AVX-NEXT: andl $96, %ecx -; AVX-NEXT: shrl $3, %ecx -; AVX-NEXT: movl (%rdi,%rcx), %r8d -; AVX-NEXT: btl %esi, %r8d -; AVX-NEXT: setae %al -; AVX-NEXT: btrl %esi, %r8d -; AVX-NEXT: shlxl %esi, %edx, %edx -; AVX-NEXT: orl %r8d, %edx -; AVX-NEXT: movl %edx, (%rdi,%rcx) -; AVX-NEXT: retq +; AVX2-LABEL: init_eq_i128: +; AVX2: # %bb.0: +; AVX2-NEXT: movl %esi, %ecx +; AVX2-NEXT: movl $1, %eax +; AVX2-NEXT: xorl %esi, %esi +; AVX2-NEXT: shldq %cl, %rax, %rsi +; AVX2-NEXT: movl %edx, %edx +; AVX2-NEXT: xorl %r8d, %r8d +; AVX2-NEXT: shldq %cl, %rdx, %r8 +; AVX2-NEXT: xorl %r9d, %r9d +; AVX2-NEXT: shlxq %rcx, %rax, %rax +; AVX2-NEXT: testb $64, %cl +; AVX2-NEXT: cmovneq %rax, %rsi +; AVX2-NEXT: cmovneq %r9, %rax +; AVX2-NEXT: shlxq %rcx, %rdx, %rdx +; AVX2-NEXT: cmovneq %rdx, %r8 +; AVX2-NEXT: cmovneq %r9, %rdx +; AVX2-NEXT: andnq 8(%rdi), %rsi, %rsi +; AVX2-NEXT: orq %r8, %rsi +; AVX2-NEXT: andnq (%rdi), %rax, %r8 +; AVX2-NEXT: orq %rdx, %r8 +; AVX2-NEXT: movl %ecx, %eax +; AVX2-NEXT: andl $96, %eax +; AVX2-NEXT: shrl $3, %eax +; AVX2-NEXT: movl (%rdi,%rax), %eax +; AVX2-NEXT: btl %ecx, %eax +; AVX2-NEXT: setae %al +; AVX2-NEXT: movq %r8, (%rdi) +; AVX2-NEXT: movq %rsi, 8(%rdi) +; AVX2-NEXT: retq +; +; AVX512-LABEL: init_eq_i128: +; AVX512: # %bb.0: +; AVX512-NEXT: movl %esi, %ecx +; AVX512-NEXT: movl $1, %eax +; AVX512-NEXT: xorl %esi, %esi +; AVX512-NEXT: shldq %cl, %rax, %rsi +; AVX512-NEXT: xorl %r8d, %r8d +; AVX512-NEXT: shlxq %rcx, %rax, %rax +; AVX512-NEXT: movl %edx, %edx +; AVX512-NEXT: xorl %r9d, %r9d +; AVX512-NEXT: shldq %cl, %rdx, %r9 +; AVX512-NEXT: testb $64, %cl +; AVX512-NEXT: cmovneq %rax, %rsi +; AVX512-NEXT: cmovneq %r8, %rax +; AVX512-NEXT: shlxq %rcx, %rdx, %rdx +; AVX512-NEXT: cmovneq %rdx, %r9 +; AVX512-NEXT: cmovneq %r8, %rdx +; AVX512-NEXT: andnq 8(%rdi), %rsi, %rsi +; AVX512-NEXT: orq %r9, %rsi +; AVX512-NEXT: andnq (%rdi), %rax, %r8 +; AVX512-NEXT: orq %rdx, %r8 +; AVX512-NEXT: movl %ecx, %eax +; AVX512-NEXT: andl $96, %eax +; AVX512-NEXT: shrl $3, %eax +; AVX512-NEXT: movl (%rdi,%rax), %eax +; AVX512-NEXT: btl %ecx, %eax +; AVX512-NEXT: setae %al +; AVX512-NEXT: movq %r8, (%rdi) +; AVX512-NEXT: movq %rsi, 8(%rdi) +; AVX512-NEXT: retq %rem = and i32 %position, 127 %ofs = zext nneg i32 %rem to i128 %bit = shl nuw i128 1, %ofs @@ -803,55 +977,673 @@ define i1 @set_ne_i512(ptr %word, i32 %position) nounwind { define i1 @init_eq_i512(ptr %word, i32 %position, i1 zeroext %value) nounwind { ; X86-LABEL: init_eq_i512: ; X86: # %bb.0: +; X86-NEXT: pushl %ebp +; X86-NEXT: movl %esp, %ebp ; X86-NEXT: pushl %ebx ; X86-NEXT: pushl %edi ; X86-NEXT: pushl %esi -; X86-NEXT: movl {{[0-9]+}}(%esp), %edx -; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X86-NEXT: movl %ecx, %esi -; X86-NEXT: shrl $3, %esi -; X86-NEXT: andl $60, %esi -; X86-NEXT: movl (%edx,%esi), %edi -; X86-NEXT: btl %ecx, %edi -; X86-NEXT: setae %al -; X86-NEXT: btrl %ecx, %edi -; X86-NEXT: movzbl {{[0-9]+}}(%esp), %ebx -; X86-NEXT: # kill: def $cl killed $cl killed $ecx -; X86-NEXT: shll %cl, %ebx +; X86-NEXT: andl $-16, %esp +; X86-NEXT: subl $352, %esp # imm = 0x160 +; X86-NEXT: movl 12(%ebp), %ecx +; X86-NEXT: movl %ecx, %edx +; X86-NEXT: shrl $3, %edx +; X86-NEXT: andl $60, %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; X86-NEXT: subl %edx, %eax +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $1, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl 56(%eax), %esi +; X86-NEXT: movl 60(%eax), %ebx +; X86-NEXT: movl 52(%eax), %edi +; X86-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 48(%eax), %edi +; X86-NEXT: movl 44(%eax), %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 40(%eax), %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 36(%eax), %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 32(%eax), %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 28(%eax), %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 24(%eax), %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 20(%eax), %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 16(%eax), %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 12(%eax), %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 8(%eax), %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl (%eax), %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl 4(%eax), %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movzbl 16(%ebp), %eax +; X86-NEXT: movzbl %al, %eax +; X86-NEXT: movl %eax, {{[0-9]+}}(%esp) +; X86-NEXT: andl $31, %ecx +; X86-NEXT: shldl %cl, %esi, %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: shldl %cl, %eax, %esi +; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: shldl %cl, %edi, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: shldl %cl, %ebx, %edi +; X86-NEXT: movl %edi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload +; X86-NEXT: shldl %cl, %edx, %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: shldl %cl, %eax, %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload +; X86-NEXT: shldl %cl, %edx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: shldl %cl, %eax, %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload +; X86-NEXT: shldl %cl, %edx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: shldl %cl, %eax, %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload +; X86-NEXT: shldl %cl, %edx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: shldl %cl, %eax, %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %edx # 4-byte Reload +; X86-NEXT: shldl %cl, %edx, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload +; X86-NEXT: shldl %cl, %esi, %edx +; X86-NEXT: movl %edx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: shldl %cl, %eax, %esi +; X86-NEXT: movl %esi, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: shll %cl, %eax +; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: notl %ebx +; X86-NEXT: leal {{[0-9]+}}(%esp), %eax +; X86-NEXT: subl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Folded Reload +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl $0, {{[0-9]+}}(%esp) +; X86-NEXT: movl 56(%eax), %esi +; X86-NEXT: movl 60(%eax), %edi +; X86-NEXT: shldl %cl, %esi, %edi +; X86-NEXT: movl 8(%ebp), %edx +; X86-NEXT: andl 60(%edx), %ebx ; X86-NEXT: orl %edi, %ebx -; X86-NEXT: movl %ebx, (%edx,%esi) +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: notl %ebx +; X86-NEXT: movl 52(%eax), %edi +; X86-NEXT: shldl %cl, %edi, %esi +; X86-NEXT: andl 56(%edx), %ebx +; X86-NEXT: orl %esi, %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: notl %ebx +; X86-NEXT: movl 48(%eax), %esi +; X86-NEXT: shldl %cl, %esi, %edi +; X86-NEXT: andl 52(%edx), %ebx +; X86-NEXT: orl %edi, %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: notl %ebx +; X86-NEXT: movl 44(%eax), %edi +; X86-NEXT: shldl %cl, %edi, %esi +; X86-NEXT: andl 48(%edx), %ebx +; X86-NEXT: orl %esi, %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: notl %ebx +; X86-NEXT: movl 40(%eax), %esi +; X86-NEXT: shldl %cl, %esi, %edi +; X86-NEXT: andl 44(%edx), %ebx +; X86-NEXT: orl %edi, %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: notl %ebx +; X86-NEXT: movl 36(%eax), %edi +; X86-NEXT: shldl %cl, %edi, %esi +; X86-NEXT: andl 40(%edx), %ebx +; X86-NEXT: orl %esi, %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: notl %ebx +; X86-NEXT: movl 32(%eax), %esi +; X86-NEXT: shldl %cl, %esi, %edi +; X86-NEXT: andl 36(%edx), %ebx +; X86-NEXT: orl %edi, %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: notl %ebx +; X86-NEXT: movl 28(%eax), %edi +; X86-NEXT: shldl %cl, %edi, %esi +; X86-NEXT: andl 32(%edx), %ebx +; X86-NEXT: orl %esi, %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: notl %ebx +; X86-NEXT: movl 24(%eax), %esi +; X86-NEXT: shldl %cl, %esi, %edi +; X86-NEXT: andl 28(%edx), %ebx +; X86-NEXT: orl %edi, %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: notl %ebx +; X86-NEXT: movl 20(%eax), %edi +; X86-NEXT: shldl %cl, %edi, %esi +; X86-NEXT: andl 24(%edx), %ebx +; X86-NEXT: orl %esi, %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: notl %ebx +; X86-NEXT: movl 16(%eax), %esi +; X86-NEXT: shldl %cl, %esi, %edi +; X86-NEXT: andl 20(%edx), %ebx +; X86-NEXT: orl %edi, %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: notl %ebx +; X86-NEXT: movl 12(%eax), %edi +; X86-NEXT: shldl %cl, %edi, %esi +; X86-NEXT: andl 16(%edx), %ebx +; X86-NEXT: orl %esi, %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: notl %ebx +; X86-NEXT: movl 8(%eax), %esi +; X86-NEXT: shldl %cl, %esi, %edi +; X86-NEXT: andl 12(%edx), %ebx +; X86-NEXT: orl %edi, %ebx +; X86-NEXT: movl %ebx, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload +; X86-NEXT: notl %ebx +; X86-NEXT: movl 4(%eax), %edi +; X86-NEXT: shldl %cl, %edi, %esi +; X86-NEXT: andl 8(%edx), %ebx +; X86-NEXT: orl %esi, %ebx +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload +; X86-NEXT: notl %esi +; X86-NEXT: movl (%eax), %eax +; X86-NEXT: shldl %cl, %eax, %edi +; X86-NEXT: andl 4(%edx), %esi +; X86-NEXT: orl %edi, %esi +; X86-NEXT: movl %esi, %edi +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %esi # 4-byte Reload +; X86-NEXT: notl %esi +; X86-NEXT: shll %cl, %eax +; X86-NEXT: andl (%edx), %esi +; X86-NEXT: orl %eax, %esi +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl (%edx,%eax), %eax +; X86-NEXT: btl %ecx, %eax +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, 60(%edx) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, 56(%edx) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, 52(%edx) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, 48(%edx) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, 44(%edx) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, 40(%edx) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, 36(%edx) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, 32(%edx) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, 28(%edx) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, 24(%edx) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, 20(%edx) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, 16(%edx) +; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: movl %eax, 12(%edx) +; X86-NEXT: movl %ebx, 8(%edx) +; X86-NEXT: movl %edi, 4(%edx) +; X86-NEXT: movl %esi, (%edx) +; X86-NEXT: setae %al +; X86-NEXT: leal -12(%ebp), %esp ; X86-NEXT: popl %esi ; X86-NEXT: popl %edi ; X86-NEXT: popl %ebx +; X86-NEXT: popl %ebp ; X86-NEXT: retl ; ; SSE-LABEL: init_eq_i512: ; SSE: # %bb.0: +; SSE-NEXT: pushq %rbp +; SSE-NEXT: pushq %r15 +; SSE-NEXT: pushq %r14 +; SSE-NEXT: pushq %r13 +; SSE-NEXT: pushq %r12 +; SSE-NEXT: pushq %rbx +; SSE-NEXT: subq $168, %rsp +; SSE-NEXT: movl %esi, {{[-0-9]+}}(%r{{[sb]}}p) # 4-byte Spill +; SSE-NEXT: xorps %xmm0, %xmm0 +; SSE-NEXT: movups %xmm0, {{[0-9]+}}(%rsp) +; SSE-NEXT: movups %xmm0, {{[0-9]+}}(%rsp) +; SSE-NEXT: movups %xmm0, {{[0-9]+}}(%rsp) +; SSE-NEXT: movaps %xmm0, {{[0-9]+}}(%rsp) +; SSE-NEXT: movaps %xmm0, {{[0-9]+}}(%rsp) +; SSE-NEXT: movaps %xmm0, {{[0-9]+}}(%rsp) +; SSE-NEXT: movaps %xmm0, {{[0-9]+}}(%rsp) +; SSE-NEXT: movq $0, {{[0-9]+}}(%rsp) +; SSE-NEXT: movq $1, {{[0-9]+}}(%rsp) ; SSE-NEXT: movl %esi, %ecx -; SSE-NEXT: shrl $3, %esi -; SSE-NEXT: andl $60, %esi -; SSE-NEXT: movl (%rdi,%rsi), %r8d -; SSE-NEXT: btl %ecx, %r8d +; SSE-NEXT: andl $63, %ecx +; SSE-NEXT: movl %esi, %eax +; SSE-NEXT: shrl $3, %eax +; SSE-NEXT: movq %rax, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill +; SSE-NEXT: # kill: def $eax killed $eax killed $rax +; SSE-NEXT: andl $56, %eax +; SSE-NEXT: negl %eax +; SSE-NEXT: movslq %eax, %r12 +; SSE-NEXT: movq 136(%rsp,%r12), %r9 +; SSE-NEXT: movq 144(%rsp,%r12), %rax +; SSE-NEXT: movq %rax, %rsi +; SSE-NEXT: shldq %cl, %r9, %rsi +; SSE-NEXT: movq %rsi, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill +; SSE-NEXT: movq 152(%rsp,%r12), %r11 +; SSE-NEXT: shldq %cl, %rax, %r11 +; SSE-NEXT: movq 120(%rsp,%r12), %r10 +; SSE-NEXT: movq 128(%rsp,%r12), %rax +; SSE-NEXT: movq %rax, %rbx +; SSE-NEXT: shldq %cl, %r10, %rbx +; SSE-NEXT: shldq %cl, %rax, %r9 +; SSE-NEXT: movq 104(%rsp,%r12), %r14 +; SSE-NEXT: movq 112(%rsp,%r12), %rax +; SSE-NEXT: movq %rax, %r15 +; SSE-NEXT: shldq %cl, %r14, %r15 +; SSE-NEXT: shldq %cl, %rax, %r10 +; SSE-NEXT: movq 96(%rsp,%r12), %rax +; SSE-NEXT: movq %rax, %r13 +; SSE-NEXT: shlq %cl, %r13 +; SSE-NEXT: shldq %cl, %rax, %r14 +; SSE-NEXT: movl %edx, %eax +; SSE-NEXT: movups %xmm0, {{[0-9]+}}(%rsp) +; SSE-NEXT: movups %xmm0, -{{[0-9]+}}(%rsp) +; SSE-NEXT: movups %xmm0, -{{[0-9]+}}(%rsp) +; SSE-NEXT: movaps %xmm0, -{{[0-9]+}}(%rsp) +; SSE-NEXT: movaps %xmm0, -{{[0-9]+}}(%rsp) +; SSE-NEXT: movaps %xmm0, -{{[0-9]+}}(%rsp) +; SSE-NEXT: movaps %xmm0, -{{[0-9]+}}(%rsp) +; SSE-NEXT: movq %rax, -{{[0-9]+}}(%rsp) +; SSE-NEXT: movq $0, {{[0-9]+}}(%rsp) +; SSE-NEXT: movq 8(%rsp,%r12), %r8 +; SSE-NEXT: movq 16(%rsp,%r12), %rsi +; SSE-NEXT: movq %rsi, %rbp +; SSE-NEXT: shldq %cl, %r8, %rbp +; SSE-NEXT: movq {{[-0-9]+}}(%r{{[sb]}}p), %rax # 8-byte Reload +; SSE-NEXT: notq %rax +; SSE-NEXT: andq 48(%rdi), %rax +; SSE-NEXT: orq %rbp, %rax +; SSE-NEXT: movq %rax, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill +; SSE-NEXT: notq %rbx +; SSE-NEXT: notq %r11 +; SSE-NEXT: movq 24(%rsp,%r12), %rax +; SSE-NEXT: shldq %cl, %rsi, %rax +; SSE-NEXT: movq -8(%rsp,%r12), %rbp +; SSE-NEXT: movq (%rsp,%r12), %rdx +; SSE-NEXT: movq %rdx, %rsi +; SSE-NEXT: shldq %cl, %rbp, %rsi +; SSE-NEXT: andq 56(%rdi), %r11 +; SSE-NEXT: andq 32(%rdi), %rbx +; SSE-NEXT: orq %rax, %r11 +; SSE-NEXT: orq %rsi, %rbx +; SSE-NEXT: notq %r15 +; SSE-NEXT: shldq %cl, %rdx, %r8 +; SSE-NEXT: notq %r9 +; SSE-NEXT: andq 40(%rdi), %r9 +; SSE-NEXT: orq %r8, %r9 +; SSE-NEXT: movq -24(%rsp,%r12), %rax +; SSE-NEXT: movq -16(%rsp,%r12), %rdx +; SSE-NEXT: movq %rdx, %rsi +; SSE-NEXT: shldq %cl, %rax, %rsi +; SSE-NEXT: andq 16(%rdi), %r15 +; SSE-NEXT: orq %rsi, %r15 +; SSE-NEXT: shldq %cl, %rdx, %rbp +; SSE-NEXT: notq %r10 +; SSE-NEXT: notq %r13 +; SSE-NEXT: movq -32(%rsp,%r12), %rdx +; SSE-NEXT: movq %rdx, %rsi +; SSE-NEXT: shlq %cl, %rsi +; SSE-NEXT: andq 24(%rdi), %r10 +; SSE-NEXT: andq (%rdi), %r13 +; SSE-NEXT: orq %rbp, %r10 +; SSE-NEXT: orq %rsi, %r13 +; SSE-NEXT: notq %r14 +; SSE-NEXT: # kill: def $cl killed $cl killed $ecx +; SSE-NEXT: shldq %cl, %rdx, %rax +; SSE-NEXT: andq 8(%rdi), %r14 +; SSE-NEXT: orq %rax, %r14 +; SSE-NEXT: movq {{[-0-9]+}}(%r{{[sb]}}p), %rax # 8-byte Reload +; SSE-NEXT: andl $60, %eax +; SSE-NEXT: movl (%rdi,%rax), %eax +; SSE-NEXT: movl {{[-0-9]+}}(%r{{[sb]}}p), %ecx # 4-byte Reload +; SSE-NEXT: btl %ecx, %eax +; SSE-NEXT: movq {{[-0-9]+}}(%r{{[sb]}}p), %rax # 8-byte Reload +; SSE-NEXT: movq %rax, 48(%rdi) +; SSE-NEXT: movq %r11, 56(%rdi) +; SSE-NEXT: movq %rbx, 32(%rdi) +; SSE-NEXT: movq %r9, 40(%rdi) +; SSE-NEXT: movq %r15, 16(%rdi) +; SSE-NEXT: movq %r10, 24(%rdi) +; SSE-NEXT: movq %r13, (%rdi) +; SSE-NEXT: movq %r14, 8(%rdi) ; SSE-NEXT: setae %al -; SSE-NEXT: shll %cl, %edx -; SSE-NEXT: btrl %ecx, %r8d -; SSE-NEXT: orl %r8d, %edx -; SSE-NEXT: movl %edx, (%rdi,%rsi) +; SSE-NEXT: addq $168, %rsp +; SSE-NEXT: popq %rbx +; SSE-NEXT: popq %r12 +; SSE-NEXT: popq %r13 +; SSE-NEXT: popq %r14 +; SSE-NEXT: popq %r15 +; SSE-NEXT: popq %rbp ; SSE-NEXT: retq ; -; AVX-LABEL: init_eq_i512: -; AVX: # %bb.0: -; AVX-NEXT: movl %esi, %ecx -; AVX-NEXT: shrl $3, %ecx -; AVX-NEXT: andl $60, %ecx -; AVX-NEXT: movl (%rdi,%rcx), %r8d -; AVX-NEXT: btl %esi, %r8d -; AVX-NEXT: setae %al -; AVX-NEXT: btrl %esi, %r8d -; AVX-NEXT: shlxl %esi, %edx, %edx -; AVX-NEXT: orl %r8d, %edx -; AVX-NEXT: movl %edx, (%rdi,%rcx) -; AVX-NEXT: retq +; AVX2-LABEL: init_eq_i512: +; AVX2: # %bb.0: +; AVX2-NEXT: pushq %rbp +; AVX2-NEXT: pushq %r15 +; AVX2-NEXT: pushq %r14 +; AVX2-NEXT: pushq %r13 +; AVX2-NEXT: pushq %r12 +; AVX2-NEXT: pushq %rbx +; AVX2-NEXT: subq $184, %rsp +; AVX2-NEXT: vxorps %xmm0, %xmm0, %xmm0 +; AVX2-NEXT: vmovups %ymm0, {{[0-9]+}}(%rsp) +; AVX2-NEXT: vmovups %ymm0, {{[0-9]+}}(%rsp) +; AVX2-NEXT: vmovups %ymm0, {{[0-9]+}}(%rsp) +; AVX2-NEXT: vmovss {{.*#+}} xmm1 = [1,0,0,0] +; AVX2-NEXT: vmovups %ymm1, {{[0-9]+}}(%rsp) +; AVX2-NEXT: movl %esi, {{[-0-9]+}}(%r{{[sb]}}p) # 4-byte Spill +; AVX2-NEXT: movl %esi, %ecx +; AVX2-NEXT: andl $63, %ecx +; AVX2-NEXT: movl %esi, %ebx +; AVX2-NEXT: shrl $3, %ebx +; AVX2-NEXT: movl %ebx, %eax +; AVX2-NEXT: andl $56, %eax +; AVX2-NEXT: negl %eax +; AVX2-NEXT: movslq %eax, %r11 +; AVX2-NEXT: movq 128(%rsp,%r11), %r15 +; AVX2-NEXT: movq 136(%rsp,%r11), %rax +; AVX2-NEXT: movq %rax, %rsi +; AVX2-NEXT: shldq %cl, %r15, %rsi +; AVX2-NEXT: movq %rsi, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill +; AVX2-NEXT: movq 120(%rsp,%r11), %r8 +; AVX2-NEXT: shldq %cl, %r8, %r15 +; AVX2-NEXT: movq 144(%rsp,%r11), %r14 +; AVX2-NEXT: movq 152(%rsp,%r11), %rsi +; AVX2-NEXT: movq %rsi, %r9 +; AVX2-NEXT: shldq %cl, %r14, %r9 +; AVX2-NEXT: movq %r9, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill +; AVX2-NEXT: shldq %cl, %rax, %r14 +; AVX2-NEXT: movq 112(%rsp,%r11), %rax +; AVX2-NEXT: movq %rax, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill +; AVX2-NEXT: movq 160(%rsp,%r11), %r13 +; AVX2-NEXT: movq 168(%rsp,%r11), %r12 +; AVX2-NEXT: shldq %cl, %r13, %r12 +; AVX2-NEXT: shldq %cl, %rsi, %r13 +; AVX2-NEXT: shldq %cl, %rax, %r8 +; AVX2-NEXT: movq %r8, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill +; AVX2-NEXT: movl %edx, %eax +; AVX2-NEXT: vxorps %xmm1, %xmm1, %xmm1 +; AVX2-NEXT: vmovups %xmm1, {{[0-9]+}}(%rsp) +; AVX2-NEXT: vmovups %ymm0, -{{[0-9]+}}(%rsp) +; AVX2-NEXT: vmovups %ymm0, -{{[0-9]+}}(%rsp) +; AVX2-NEXT: vmovups %ymm0, -{{[0-9]+}}(%rsp) +; AVX2-NEXT: movq %rax, -{{[0-9]+}}(%rsp) +; AVX2-NEXT: movq $0, {{[0-9]+}}(%rsp) +; AVX2-NEXT: movq 24(%rsp,%r11), %rbp +; AVX2-NEXT: movq 32(%rsp,%r11), %rdx +; AVX2-NEXT: movq %rdx, %rax +; AVX2-NEXT: shldq %cl, %rbp, %rax +; AVX2-NEXT: movq 40(%rsp,%r11), %r10 +; AVX2-NEXT: shldq %cl, %rdx, %r10 +; AVX2-NEXT: movq 8(%rsp,%r11), %r9 +; AVX2-NEXT: movq 16(%rsp,%r11), %rdx +; AVX2-NEXT: movq %rdx, %r8 +; AVX2-NEXT: shldq %cl, %r9, %r8 +; AVX2-NEXT: shldq %cl, %rdx, %rbp +; AVX2-NEXT: andnq 48(%rdi), %r13, %r13 +; AVX2-NEXT: orq %rax, %r13 +; AVX2-NEXT: movq -8(%rsp,%r11), %rax +; AVX2-NEXT: movq (%rsp,%r11), %rdx +; AVX2-NEXT: movq %rdx, %rsi +; AVX2-NEXT: shldq %cl, %rax, %rsi +; AVX2-NEXT: shldq %cl, %rdx, %r9 +; AVX2-NEXT: andnq 56(%rdi), %r12, %r12 +; AVX2-NEXT: andnq 32(%rdi), %r14, %r14 +; AVX2-NEXT: orq %r10, %r12 +; AVX2-NEXT: orq %r8, %r14 +; AVX2-NEXT: movq {{[-0-9]+}}(%r{{[sb]}}p), %rdx # 8-byte Reload +; AVX2-NEXT: andnq 40(%rdi), %rdx, %rdx +; AVX2-NEXT: orq %rbp, %rdx +; AVX2-NEXT: shlxq %rcx, {{[-0-9]+}}(%r{{[sb]}}p), %r8 # 8-byte Folded Reload +; AVX2-NEXT: movq -16(%rsp,%r11), %r10 +; AVX2-NEXT: shlxq %rcx, %r10, %r11 +; AVX2-NEXT: # kill: def $cl killed $cl killed $rcx +; AVX2-NEXT: shldq %cl, %r10, %rax +; AVX2-NEXT: andnq 16(%rdi), %r15, %rcx +; AVX2-NEXT: movq {{[-0-9]+}}(%r{{[sb]}}p), %r10 # 8-byte Reload +; AVX2-NEXT: andnq 24(%rdi), %r10, %r10 +; AVX2-NEXT: orq %rsi, %rcx +; AVX2-NEXT: orq %r9, %r10 +; AVX2-NEXT: andnq (%rdi), %r8, %rsi +; AVX2-NEXT: orq %r11, %rsi +; AVX2-NEXT: movq {{[-0-9]+}}(%r{{[sb]}}p), %r8 # 8-byte Reload +; AVX2-NEXT: andnq 8(%rdi), %r8, %r8 +; AVX2-NEXT: orq %rax, %r8 +; AVX2-NEXT: andl $60, %ebx +; AVX2-NEXT: movl (%rdi,%rbx), %eax +; AVX2-NEXT: movl {{[-0-9]+}}(%r{{[sb]}}p), %r9d # 4-byte Reload +; AVX2-NEXT: btl %r9d, %eax +; AVX2-NEXT: movq %r13, 48(%rdi) +; AVX2-NEXT: movq %r12, 56(%rdi) +; AVX2-NEXT: movq %r14, 32(%rdi) +; AVX2-NEXT: movq %rdx, 40(%rdi) +; AVX2-NEXT: movq %rcx, 16(%rdi) +; AVX2-NEXT: movq %r10, 24(%rdi) +; AVX2-NEXT: movq %rsi, (%rdi) +; AVX2-NEXT: movq %r8, 8(%rdi) +; AVX2-NEXT: setae %al +; AVX2-NEXT: addq $184, %rsp +; AVX2-NEXT: popq %rbx +; AVX2-NEXT: popq %r12 +; AVX2-NEXT: popq %r13 +; AVX2-NEXT: popq %r14 +; AVX2-NEXT: popq %r15 +; AVX2-NEXT: popq %rbp +; AVX2-NEXT: vzeroupper +; AVX2-NEXT: retq +; +; AVX512-LABEL: init_eq_i512: +; AVX512: # %bb.0: +; AVX512-NEXT: pushq %rbp +; AVX512-NEXT: pushq %r15 +; AVX512-NEXT: pushq %r14 +; AVX512-NEXT: pushq %r13 +; AVX512-NEXT: pushq %r12 +; AVX512-NEXT: pushq %rbx +; AVX512-NEXT: subq $168, %rsp +; AVX512-NEXT: vxorps %xmm0, %xmm0, %xmm0 +; AVX512-NEXT: vmovups %ymm0, {{[0-9]+}}(%rsp) +; AVX512-NEXT: vmovups %ymm0, {{[0-9]+}}(%rsp) +; AVX512-NEXT: vmovups %ymm0, {{[0-9]+}}(%rsp) +; AVX512-NEXT: vmovaps {{.*#+}} xmm1 = [1,0,0,0] +; AVX512-NEXT: vmovups %ymm1, {{[0-9]+}}(%rsp) +; AVX512-NEXT: movl %esi, {{[-0-9]+}}(%r{{[sb]}}p) # 4-byte Spill +; AVX512-NEXT: movl %esi, %ecx +; AVX512-NEXT: andl $63, %ecx +; AVX512-NEXT: movl %esi, %r10d +; AVX512-NEXT: shrl $3, %r10d +; AVX512-NEXT: movl %r10d, %r8d +; AVX512-NEXT: andl $56, %r8d +; AVX512-NEXT: negl %r8d +; AVX512-NEXT: movslq %r8d, %r9 +; AVX512-NEXT: movq 112(%rsp,%r9), %r11 +; AVX512-NEXT: movq 120(%rsp,%r9), %r14 +; AVX512-NEXT: movq %r14, %rax +; AVX512-NEXT: shldq %cl, %r11, %rax +; AVX512-NEXT: movq %rax, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill +; AVX512-NEXT: movq 104(%rsp,%r9), %rax +; AVX512-NEXT: shldq %cl, %rax, %r11 +; AVX512-NEXT: movq 128(%rsp,%r9), %r15 +; AVX512-NEXT: movq 136(%rsp,%r9), %rbp +; AVX512-NEXT: movq %rbp, %rbx +; AVX512-NEXT: shldq %cl, %r15, %rbx +; AVX512-NEXT: shldq %cl, %r14, %r15 +; AVX512-NEXT: movq 144(%rsp,%r9), %r13 +; AVX512-NEXT: movq 152(%rsp,%r9), %r12 +; AVX512-NEXT: shldq %cl, %r13, %r12 +; AVX512-NEXT: movq 96(%rsp,%r9), %r14 +; AVX512-NEXT: shldq %cl, %rbp, %r13 +; AVX512-NEXT: shldq %cl, %r14, %rax +; AVX512-NEXT: movq %rax, {{[-0-9]+}}(%r{{[sb]}}p) # 8-byte Spill +; AVX512-NEXT: movl %edx, %edx +; AVX512-NEXT: vxorps %xmm1, %xmm1, %xmm1 +; AVX512-NEXT: vmovups %xmm1, {{[0-9]+}}(%rsp) +; AVX512-NEXT: vmovups %ymm0, -{{[0-9]+}}(%rsp) +; AVX512-NEXT: vmovups %ymm0, -{{[0-9]+}}(%rsp) +; AVX512-NEXT: vmovups %ymm0, -{{[0-9]+}}(%rsp) +; AVX512-NEXT: movq %rdx, -{{[0-9]+}}(%rsp) +; AVX512-NEXT: movq $0, {{[0-9]+}}(%rsp) +; AVX512-NEXT: movq 8(%rsp,%r9), %r8 +; AVX512-NEXT: movq 16(%rsp,%r9), %rax +; AVX512-NEXT: movq %rax, %rbp +; AVX512-NEXT: shldq %cl, %r8, %rbp +; AVX512-NEXT: andnq 48(%rdi), %r13, %r13 +; AVX512-NEXT: orq %rbp, %r13 +; AVX512-NEXT: movq 24(%rsp,%r9), %rbp +; AVX512-NEXT: shldq %cl, %rax, %rbp +; AVX512-NEXT: movq -8(%rsp,%r9), %rax +; AVX512-NEXT: movq (%rsp,%r9), %rsi +; AVX512-NEXT: movq %rsi, %rdx +; AVX512-NEXT: shldq %cl, %rax, %rdx +; AVX512-NEXT: andnq 56(%rdi), %r12, %r12 +; AVX512-NEXT: orq %rbp, %r12 +; AVX512-NEXT: andnq 32(%rdi), %r15, %r15 +; AVX512-NEXT: orq %rdx, %r15 +; AVX512-NEXT: shldq %cl, %rsi, %r8 +; AVX512-NEXT: movq -24(%rsp,%r9), %rdx +; AVX512-NEXT: movq -16(%rsp,%r9), %rsi +; AVX512-NEXT: movq %rsi, %rbp +; AVX512-NEXT: shldq %cl, %rdx, %rbp +; AVX512-NEXT: andnq 40(%rdi), %rbx, %rbx +; AVX512-NEXT: orq %r8, %rbx +; AVX512-NEXT: andnq 16(%rdi), %r11, %r8 +; AVX512-NEXT: orq %rbp, %r8 +; AVX512-NEXT: shlxq %rcx, %r14, %r11 +; AVX512-NEXT: movq -32(%rsp,%r9), %r9 +; AVX512-NEXT: shldq %cl, %rsi, %rax +; AVX512-NEXT: movq {{[-0-9]+}}(%r{{[sb]}}p), %rsi # 8-byte Reload +; AVX512-NEXT: andnq 24(%rdi), %rsi, %rsi +; AVX512-NEXT: orq %rax, %rsi +; AVX512-NEXT: shlxq %rcx, %r9, %rax +; AVX512-NEXT: # kill: def $cl killed $cl killed $rcx +; AVX512-NEXT: shldq %cl, %r9, %rdx +; AVX512-NEXT: andnq (%rdi), %r11, %rcx +; AVX512-NEXT: orq %rax, %rcx +; AVX512-NEXT: movq {{[-0-9]+}}(%r{{[sb]}}p), %rax # 8-byte Reload +; AVX512-NEXT: andnq 8(%rdi), %rax, %rax +; AVX512-NEXT: orq %rdx, %rax +; AVX512-NEXT: andl $60, %r10d +; AVX512-NEXT: movl (%rdi,%r10), %edx +; AVX512-NEXT: movl {{[-0-9]+}}(%r{{[sb]}}p), %r9d # 4-byte Reload +; AVX512-NEXT: btl %r9d, %edx +; AVX512-NEXT: movq %r13, 48(%rdi) +; AVX512-NEXT: movq %r12, 56(%rdi) +; AVX512-NEXT: movq %r15, 32(%rdi) +; AVX512-NEXT: movq %rbx, 40(%rdi) +; AVX512-NEXT: movq %r8, 16(%rdi) +; AVX512-NEXT: movq %rsi, 24(%rdi) +; AVX512-NEXT: movq %rcx, (%rdi) +; AVX512-NEXT: movq %rax, 8(%rdi) +; AVX512-NEXT: setae %al +; AVX512-NEXT: addq $168, %rsp +; AVX512-NEXT: popq %rbx +; AVX512-NEXT: popq %r12 +; AVX512-NEXT: popq %r13 +; AVX512-NEXT: popq %r14 +; AVX512-NEXT: popq %r15 +; AVX512-NEXT: popq %rbp +; AVX512-NEXT: vzeroupper +; AVX512-NEXT: retq %rem = and i32 %position, 511 %ofs = zext nneg i32 %rem to i512 %bit = shl nuw i512 1, %ofs @@ -1049,40 +1841,33 @@ define i32 @reset_multiload_i128(ptr %word, i32 %position, ptr %p) nounwind { ; X86-NEXT: shrb $3, %al ; X86-NEXT: andb $12, %al ; X86-NEXT: negb %al -; X86-NEXT: movsbl %al, %edi -; X86-NEXT: movl 36(%esp,%edi), %edx -; X86-NEXT: movl 40(%esp,%edi), %ebx -; X86-NEXT: movl %ebx, %esi +; X86-NEXT: movsbl %al, %eax +; X86-NEXT: movl 40(%esp,%eax), %edx +; X86-NEXT: movl 44(%esp,%eax), %esi ; X86-NEXT: shldl %cl, %edx, %esi -; X86-NEXT: movl 32(%esp,%edi), %eax -; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl 44(%esp,%edi), %edi -; X86-NEXT: shldl %cl, %ebx, %edi -; X86-NEXT: movl %eax, %ebx -; X86-NEXT: # kill: def $cl killed $cl killed $ecx -; X86-NEXT: shll %cl, %ebx +; X86-NEXT: movl 32(%esp,%eax), %edi +; X86-NEXT: movl 36(%esp,%eax), %ebx +; X86-NEXT: shldl %cl, %ebx, %edx +; X86-NEXT: shldl %cl, %edi, %ebx ; X86-NEXT: notl %ebx ; X86-NEXT: movl 16(%ebp), %eax ; X86-NEXT: movl (%eax), %eax ; X86-NEXT: movl %eax, {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Spill -; X86-NEXT: movl 12(%ebp), %eax -; X86-NEXT: andl $96, %eax -; X86-NEXT: shrl $3, %eax -; X86-NEXT: movl 8(%ebp), %ecx -; X86-NEXT: movl (%ecx,%eax), %eax -; X86-NEXT: andl %ebx, (%ecx) -; X86-NEXT: movl 12(%ebp), %ecx -; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %ebx # 4-byte Reload -; X86-NEXT: shldl %cl, %ebx, %edx -; X86-NEXT: notl %edx -; X86-NEXT: movl 8(%ebp), %ebx -; X86-NEXT: andl %edx, 4(%ebx) -; X86-NEXT: notl %esi -; X86-NEXT: andl %esi, 8(%ebx) +; X86-NEXT: movl 8(%ebp), %eax +; X86-NEXT: andl %ebx, 4(%eax) +; X86-NEXT: shll %cl, %edi ; X86-NEXT: notl %edi -; X86-NEXT: andl %edi, 12(%ebx) -; X86-NEXT: btl %ecx, %eax +; X86-NEXT: movl %ecx, %ebx +; X86-NEXT: andl $96, %ebx +; X86-NEXT: shrl $3, %ebx +; X86-NEXT: movl (%eax,%ebx), %ebx +; X86-NEXT: andl %edi, (%eax) +; X86-NEXT: notl %esi +; X86-NEXT: andl %esi, 12(%eax) +; X86-NEXT: notl %edx +; X86-NEXT: andl %edx, 8(%eax) ; X86-NEXT: movl {{[-0-9]+}}(%e{{[sb]}}p), %eax # 4-byte Reload +; X86-NEXT: btl %ecx, %ebx ; X86-NEXT: jae .LBB22_2 ; X86-NEXT: # %bb.1: ; X86-NEXT: xorl %eax, %eax @@ -1116,8 +1901,8 @@ define i32 @reset_multiload_i128(ptr %word, i32 %position, ptr %p) nounwind { ; SSE-NEXT: # %bb.1: ; SSE-NEXT: movl (%rdx), %eax ; SSE-NEXT: .LBB22_2: -; SSE-NEXT: andq %r8, 8(%rdi) ; SSE-NEXT: andq %rsi, (%rdi) +; SSE-NEXT: andq %r8, 8(%rdi) ; SSE-NEXT: # kill: def $eax killed $eax killed $rax ; SSE-NEXT: retq ; @@ -1143,8 +1928,8 @@ define i32 @reset_multiload_i128(ptr %word, i32 %position, ptr %p) nounwind { ; AVX2-NEXT: # %bb.1: ; AVX2-NEXT: movl (%rdx), %eax ; AVX2-NEXT: .LBB22_2: -; AVX2-NEXT: andq %rsi, 8(%rdi) ; AVX2-NEXT: andq %r8, (%rdi) +; AVX2-NEXT: andq %rsi, 8(%rdi) ; AVX2-NEXT: # kill: def $eax killed $eax killed $rax ; AVX2-NEXT: retq ; @@ -1170,8 +1955,8 @@ define i32 @reset_multiload_i128(ptr %word, i32 %position, ptr %p) nounwind { ; AVX512-NEXT: # %bb.1: ; AVX512-NEXT: movl (%rdx), %eax ; AVX512-NEXT: .LBB22_2: -; AVX512-NEXT: andq %rsi, 8(%rdi) ; AVX512-NEXT: andq %r8, (%rdi) +; AVX512-NEXT: andq %rsi, 8(%rdi) ; AVX512-NEXT: # kill: def $eax killed $eax killed $rax ; AVX512-NEXT: retq %rem = and i32 %position, 127 diff --git a/llvm/test/CodeGen/X86/isel-llvm.sincos.ll b/llvm/test/CodeGen/X86/isel-llvm.sincos.ll index 065710f..8576f8f 100644 --- a/llvm/test/CodeGen/X86/isel-llvm.sincos.ll +++ b/llvm/test/CodeGen/X86/isel-llvm.sincos.ll @@ -3,6 +3,9 @@ ; RUN: llc < %s -mtriple=x86_64-linux-gnu -fast-isel | FileCheck %s --check-prefixes=X64,FASTISEL-X64 ; RUN: llc < %s -mtriple=i686-linux-gnu -global-isel=0 -fast-isel=0 | FileCheck %s --check-prefixes=X86,SDAG-X86 ; RUN: llc < %s -mtriple=x86_64-linux-gnu -global-isel=0 -fast-isel=0 | FileCheck %s --check-prefixes=X64,SDAG-X64 +; RUN: llc < %s -mtriple=x86_64-apple-macosx10.9.0 -mcpu=core2 | FileCheck %s --check-prefix=MACOS-SINCOS-STRET +; RUN: llc < %s -mtriple=x86_64-apple-macosx10.8.0 -mcpu=core2 | FileCheck %s --check-prefix=MACOS-NOSINCOS-STRET + ; TODO: The below RUN line will fails GISEL selection and will fallback to DAG selection due to lack of support for loads/stores in i686 mode, support is expected soon enough, for this reason the llvm/test/CodeGen/X86/GlobalISel/llvm.sincos.mir test is added for now because of the lack of support for i686 in GlobalISel. ; RUN: llc < %s -mtriple=i686-linux-gnu -global-isel=1 -global-isel-abort=2 | FileCheck %s --check-prefixes=GISEL-X86 ; RUN: llc < %s -mtriple=x86_64-linux-gnu -global-isel=1 -global-isel-abort=1 | FileCheck %s --check-prefixes=GISEL-X64 @@ -34,6 +37,29 @@ define { float, float } @test_sincos_f32(float %Val) nounwind { ; X64-NEXT: popq %rax ; X64-NEXT: retq ; +; MACOS-SINCOS-STRET-LABEL: test_sincos_f32: +; MACOS-SINCOS-STRET: ## %bb.0: +; MACOS-SINCOS-STRET-NEXT: pushq %rax +; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret +; MACOS-SINCOS-STRET-NEXT: movshdup {{.*#+}} xmm1 = xmm0[1,1,3,3] +; MACOS-SINCOS-STRET-NEXT: popq %rax +; MACOS-SINCOS-STRET-NEXT: retq +; +; MACOS-NOSINCOS-STRET-LABEL: test_sincos_f32: +; MACOS-NOSINCOS-STRET: ## %bb.0: +; MACOS-NOSINCOS-STRET-NEXT: pushq %rax +; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, (%rsp) ## 4-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _sinf +; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 4-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movss (%rsp), %xmm0 ## 4-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero,zero,zero +; MACOS-NOSINCOS-STRET-NEXT: callq _cosf +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, %xmm1 +; MACOS-NOSINCOS-STRET-NEXT: movss {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 4-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero,zero,zero +; MACOS-NOSINCOS-STRET-NEXT: popq %rax +; MACOS-NOSINCOS-STRET-NEXT: retq +; ; GISEL-X86-LABEL: test_sincos_f32: ; GISEL-X86: # %bb.0: ; GISEL-X86-NEXT: subl $28, %esp @@ -93,6 +119,28 @@ define { double, double } @test_sincos_f64(double %Val) nounwind { ; X64-NEXT: addq $24, %rsp ; X64-NEXT: retq ; +; MACOS-SINCOS-STRET-LABEL: test_sincos_f64: +; MACOS-SINCOS-STRET: ## %bb.0: +; MACOS-SINCOS-STRET-NEXT: pushq %rax +; MACOS-SINCOS-STRET-NEXT: callq ___sincos_stret +; MACOS-SINCOS-STRET-NEXT: popq %rax +; MACOS-SINCOS-STRET-NEXT: retq +; +; MACOS-NOSINCOS-STRET-LABEL: test_sincos_f64: +; MACOS-NOSINCOS-STRET: ## %bb.0: +; MACOS-NOSINCOS-STRET-NEXT: subq $24, %rsp +; MACOS-NOSINCOS-STRET-NEXT: movsd %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 8-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _sin +; MACOS-NOSINCOS-STRET-NEXT: movsd %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 8-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movsd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 8-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero +; MACOS-NOSINCOS-STRET-NEXT: callq _cos +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, %xmm1 +; MACOS-NOSINCOS-STRET-NEXT: movsd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 8-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero +; MACOS-NOSINCOS-STRET-NEXT: addq $24, %rsp +; MACOS-NOSINCOS-STRET-NEXT: retq +; ; GISEL-X86-LABEL: test_sincos_f64: ; GISEL-X86: # %bb.0: ; GISEL-X86-NEXT: subl $44, %esp @@ -153,6 +201,40 @@ define { x86_fp80, x86_fp80 } @test_sincos_f80(x86_fp80 %Val) nounwind { ; X64-NEXT: addq $56, %rsp ; X64-NEXT: retq ; +; MACOS-SINCOS-STRET-LABEL: test_sincos_f80: +; MACOS-SINCOS-STRET: ## %bb.0: +; MACOS-SINCOS-STRET-NEXT: subq $40, %rsp +; MACOS-SINCOS-STRET-NEXT: fldt {{[0-9]+}}(%rsp) +; MACOS-SINCOS-STRET-NEXT: fld %st(0) +; MACOS-SINCOS-STRET-NEXT: fstpt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Spill +; MACOS-SINCOS-STRET-NEXT: fstpt (%rsp) +; MACOS-SINCOS-STRET-NEXT: callq _cosl +; MACOS-SINCOS-STRET-NEXT: fstpt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Spill +; MACOS-SINCOS-STRET-NEXT: fldt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Reload +; MACOS-SINCOS-STRET-NEXT: fstpt (%rsp) +; MACOS-SINCOS-STRET-NEXT: callq _sinl +; MACOS-SINCOS-STRET-NEXT: fldt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Reload +; MACOS-SINCOS-STRET-NEXT: fxch %st(1) +; MACOS-SINCOS-STRET-NEXT: addq $40, %rsp +; MACOS-SINCOS-STRET-NEXT: retq +; +; MACOS-NOSINCOS-STRET-LABEL: test_sincos_f80: +; MACOS-NOSINCOS-STRET: ## %bb.0: +; MACOS-NOSINCOS-STRET-NEXT: subq $40, %rsp +; MACOS-NOSINCOS-STRET-NEXT: fldt {{[0-9]+}}(%rsp) +; MACOS-NOSINCOS-STRET-NEXT: fld %st(0) +; MACOS-NOSINCOS-STRET-NEXT: fstpt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Spill +; MACOS-NOSINCOS-STRET-NEXT: fstpt (%rsp) +; MACOS-NOSINCOS-STRET-NEXT: callq _cosl +; MACOS-NOSINCOS-STRET-NEXT: fstpt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Spill +; MACOS-NOSINCOS-STRET-NEXT: fldt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Reload +; MACOS-NOSINCOS-STRET-NEXT: fstpt (%rsp) +; MACOS-NOSINCOS-STRET-NEXT: callq _sinl +; MACOS-NOSINCOS-STRET-NEXT: fldt {{[-0-9]+}}(%r{{[sb]}}p) ## 10-byte Folded Reload +; MACOS-NOSINCOS-STRET-NEXT: fxch %st(1) +; MACOS-NOSINCOS-STRET-NEXT: addq $40, %rsp +; MACOS-NOSINCOS-STRET-NEXT: retq +; ; GISEL-X86-LABEL: test_sincos_f80: ; GISEL-X86: # %bb.0: ; GISEL-X86-NEXT: subl $60, %esp @@ -288,6 +370,57 @@ define void @can_fold_with_call_in_chain(float %x, ptr noalias %a, ptr noalias % ; SDAG-X64-NEXT: popq %r14 ; SDAG-X64-NEXT: retq ; +; MACOS-SINCOS-STRET-LABEL: can_fold_with_call_in_chain: +; MACOS-SINCOS-STRET: ## %bb.0: ## %entry +; MACOS-SINCOS-STRET-NEXT: pushq %r14 +; MACOS-SINCOS-STRET-NEXT: pushq %rbx +; MACOS-SINCOS-STRET-NEXT: subq $40, %rsp +; MACOS-SINCOS-STRET-NEXT: movq %rsi, %rbx +; MACOS-SINCOS-STRET-NEXT: movq %rdi, %r14 +; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movshdup {{.*#+}} xmm0 = xmm0[1,1,3,3] +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movq %r14, %rdi +; MACOS-SINCOS-STRET-NEXT: movq %rbx, %rsi +; MACOS-SINCOS-STRET-NEXT: callq _foo +; MACOS-SINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: movss %xmm0, (%r14) +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: movss %xmm0, (%rbx) +; MACOS-SINCOS-STRET-NEXT: addq $40, %rsp +; MACOS-SINCOS-STRET-NEXT: popq %rbx +; MACOS-SINCOS-STRET-NEXT: popq %r14 +; MACOS-SINCOS-STRET-NEXT: retq +; +; MACOS-NOSINCOS-STRET-LABEL: can_fold_with_call_in_chain: +; MACOS-NOSINCOS-STRET: ## %bb.0: ## %entry +; MACOS-NOSINCOS-STRET-NEXT: pushq %r14 +; MACOS-NOSINCOS-STRET-NEXT: pushq %rbx +; MACOS-NOSINCOS-STRET-NEXT: pushq %rax +; MACOS-NOSINCOS-STRET-NEXT: movq %rsi, %rbx +; MACOS-NOSINCOS-STRET-NEXT: movq %rdi, %r14 +; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, (%rsp) ## 4-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _sinf +; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 4-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movss (%rsp), %xmm0 ## 4-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero,zero,zero +; MACOS-NOSINCOS-STRET-NEXT: callq _cosf +; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, (%rsp) ## 4-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movq %r14, %rdi +; MACOS-NOSINCOS-STRET-NEXT: movq %rbx, %rsi +; MACOS-NOSINCOS-STRET-NEXT: callq _foo +; MACOS-NOSINCOS-STRET-NEXT: movss {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 4-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero,zero,zero +; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, (%r14) +; MACOS-NOSINCOS-STRET-NEXT: movss (%rsp), %xmm0 ## 4-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = mem[0],zero,zero,zero +; MACOS-NOSINCOS-STRET-NEXT: movss %xmm0, (%rbx) +; MACOS-NOSINCOS-STRET-NEXT: addq $8, %rsp +; MACOS-NOSINCOS-STRET-NEXT: popq %rbx +; MACOS-NOSINCOS-STRET-NEXT: popq %r14 +; MACOS-NOSINCOS-STRET-NEXT: retq +; ; GISEL-X86-LABEL: can_fold_with_call_in_chain: ; GISEL-X86: # %bb.0: # %entry ; GISEL-X86-NEXT: pushl %ebx diff --git a/llvm/test/CodeGen/X86/llvm.sincos.vec.ll b/llvm/test/CodeGen/X86/llvm.sincos.vec.ll index 834dd78..9b02438 100644 --- a/llvm/test/CodeGen/X86/llvm.sincos.vec.ll +++ b/llvm/test/CodeGen/X86/llvm.sincos.vec.ll @@ -1,59 +1,213 @@ ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --no_x86_scrub_sp --version 5 -; RUN: llc < %s -mtriple=i386-unknown-linux-gnu | FileCheck %s +; RUN: llc < %s -mtriple=i386-unknown-linux-gnu | FileCheck -check-prefix=X86 %s +; RUN: llc < %s -mtriple=x86_64-unknown-linux-gnu | FileCheck -check-prefix=X64 %s +; RUN: llc < %s -mtriple=x86_64-apple-macosx10.9.0 | FileCheck --check-prefix=MACOS-SINCOS-STRET %s +; RUN: llc < %s -mtriple=x86_64-apple-macosx10.8.0 | FileCheck --check-prefix=MACOS-NOSINCOS-STRET %s define void @test_sincos_v4f32(<4 x float> %x, ptr noalias %out_sin, ptr noalias %out_cos) nounwind { -; CHECK-LABEL: test_sincos_v4f32: -; CHECK: # %bb.0: -; CHECK-NEXT: pushl %edi -; CHECK-NEXT: pushl %esi -; CHECK-NEXT: subl $52, %esp -; CHECK-NEXT: movl 84(%esp), %esi -; CHECK-NEXT: flds 76(%esp) -; CHECK-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill -; CHECK-NEXT: flds 64(%esp) -; CHECK-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill -; CHECK-NEXT: flds 72(%esp) -; CHECK-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill -; CHECK-NEXT: flds 68(%esp) -; CHECK-NEXT: movl 80(%esp), %edi -; CHECK-NEXT: leal 40(%esp), %eax -; CHECK-NEXT: movl %eax, 8(%esp) -; CHECK-NEXT: leal 4(%edi), %eax -; CHECK-NEXT: movl %eax, 4(%esp) -; CHECK-NEXT: fstps (%esp) -; CHECK-NEXT: calll sincosf -; CHECK-NEXT: leal 44(%esp), %eax -; CHECK-NEXT: movl %eax, 8(%esp) -; CHECK-NEXT: leal 8(%edi), %eax -; CHECK-NEXT: movl %eax, 4(%esp) -; CHECK-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload -; CHECK-NEXT: fstps (%esp) -; CHECK-NEXT: calll sincosf -; CHECK-NEXT: leal 36(%esp), %eax -; CHECK-NEXT: movl %eax, 8(%esp) -; CHECK-NEXT: movl %edi, 4(%esp) -; CHECK-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload -; CHECK-NEXT: fstps (%esp) -; CHECK-NEXT: calll sincosf -; CHECK-NEXT: leal 48(%esp), %eax -; CHECK-NEXT: movl %eax, 8(%esp) -; CHECK-NEXT: addl $12, %edi -; CHECK-NEXT: movl %edi, 4(%esp) -; CHECK-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload -; CHECK-NEXT: fstps (%esp) -; CHECK-NEXT: calll sincosf -; CHECK-NEXT: flds 36(%esp) -; CHECK-NEXT: flds 40(%esp) -; CHECK-NEXT: flds 44(%esp) -; CHECK-NEXT: flds 48(%esp) -; CHECK-NEXT: fstps 12(%esi) -; CHECK-NEXT: fstps 8(%esi) -; CHECK-NEXT: fstps 4(%esi) -; CHECK-NEXT: fstps (%esi) -; CHECK-NEXT: addl $52, %esp -; CHECK-NEXT: popl %esi -; CHECK-NEXT: popl %edi -; CHECK-NEXT: retl +; X86-LABEL: test_sincos_v4f32: +; X86: # %bb.0: +; X86-NEXT: pushl %edi +; X86-NEXT: pushl %esi +; X86-NEXT: subl $52, %esp +; X86-NEXT: movl 84(%esp), %esi +; X86-NEXT: flds 76(%esp) +; X86-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill +; X86-NEXT: flds 64(%esp) +; X86-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill +; X86-NEXT: flds 72(%esp) +; X86-NEXT: fstps {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Spill +; X86-NEXT: flds 68(%esp) +; X86-NEXT: movl 80(%esp), %edi +; X86-NEXT: leal 40(%esp), %eax +; X86-NEXT: movl %eax, 8(%esp) +; X86-NEXT: leal 4(%edi), %eax +; X86-NEXT: movl %eax, 4(%esp) +; X86-NEXT: fstps (%esp) +; X86-NEXT: calll sincosf +; X86-NEXT: leal 44(%esp), %eax +; X86-NEXT: movl %eax, 8(%esp) +; X86-NEXT: leal 8(%edi), %eax +; X86-NEXT: movl %eax, 4(%esp) +; X86-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload +; X86-NEXT: fstps (%esp) +; X86-NEXT: calll sincosf +; X86-NEXT: leal 36(%esp), %eax +; X86-NEXT: movl %eax, 8(%esp) +; X86-NEXT: movl %edi, 4(%esp) +; X86-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload +; X86-NEXT: fstps (%esp) +; X86-NEXT: calll sincosf +; X86-NEXT: leal 48(%esp), %eax +; X86-NEXT: movl %eax, 8(%esp) +; X86-NEXT: addl $12, %edi +; X86-NEXT: movl %edi, 4(%esp) +; X86-NEXT: flds {{[-0-9]+}}(%e{{[sb]}}p) # 4-byte Folded Reload +; X86-NEXT: fstps (%esp) +; X86-NEXT: calll sincosf +; X86-NEXT: flds 36(%esp) +; X86-NEXT: flds 40(%esp) +; X86-NEXT: flds 44(%esp) +; X86-NEXT: flds 48(%esp) +; X86-NEXT: fstps 12(%esi) +; X86-NEXT: fstps 8(%esi) +; X86-NEXT: fstps 4(%esi) +; X86-NEXT: fstps (%esi) +; X86-NEXT: addl $52, %esp +; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: retl +; +; X64-LABEL: test_sincos_v4f32: +; X64: # %bb.0: +; X64-NEXT: pushq %r14 +; X64-NEXT: pushq %rbx +; X64-NEXT: subq $56, %rsp +; X64-NEXT: movq %rsi, %rbx +; X64-NEXT: movq %rdi, %r14 +; X64-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill +; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,3,3,3] +; X64-NEXT: leaq 4(%rsp), %rdi +; X64-NEXT: movq %rsp, %rsi +; X64-NEXT: callq sincosf@PLT +; X64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload +; X64-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1] +; X64-NEXT: leaq 12(%rsp), %rdi +; X64-NEXT: leaq 8(%rsp), %rsi +; X64-NEXT: callq sincosf@PLT +; X64-NEXT: leaq 28(%rsp), %rdi +; X64-NEXT: leaq 24(%rsp), %rsi +; X64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload +; X64-NEXT: callq sincosf@PLT +; X64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload +; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1] +; X64-NEXT: leaq 20(%rsp), %rdi +; X64-NEXT: leaq 16(%rsp), %rsi +; X64-NEXT: callq sincosf@PLT +; X64-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero +; X64-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero +; X64-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1] +; X64-NEXT: movss {{.*#+}} xmm0 = mem[0],zero,zero,zero +; X64-NEXT: movss {{.*#+}} xmm2 = mem[0],zero,zero,zero +; X64-NEXT: unpcklps {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1] +; X64-NEXT: movlhps {{.*#+}} xmm0 = xmm0[0],xmm1[0] +; X64-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero +; X64-NEXT: movss {{.*#+}} xmm2 = mem[0],zero,zero,zero +; X64-NEXT: unpcklps {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1] +; X64-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero +; X64-NEXT: movss {{.*#+}} xmm3 = mem[0],zero,zero,zero +; X64-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1] +; X64-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm2[0] +; X64-NEXT: movups %xmm1, (%r14) +; X64-NEXT: movups %xmm0, (%rbx) +; X64-NEXT: addq $56, %rsp +; X64-NEXT: popq %rbx +; X64-NEXT: popq %r14 +; X64-NEXT: retq +; +; MACOS-SINCOS-STRET-LABEL: test_sincos_v4f32: +; MACOS-SINCOS-STRET: ## %bb.0: +; MACOS-SINCOS-STRET-NEXT: pushq %r14 +; MACOS-SINCOS-STRET-NEXT: pushq %rbx +; MACOS-SINCOS-STRET-NEXT: subq $104, %rsp +; MACOS-SINCOS-STRET-NEXT: movq %rsi, %rbx +; MACOS-SINCOS-STRET-NEXT: movq %rdi, %r14 +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,3,3,3] +; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1] +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1] +; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1] +; MACOS-SINCOS-STRET-NEXT: unpcklps (%rsp), %xmm0 ## 16-byte Folded Reload +; MACOS-SINCOS-STRET-NEXT: ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1] +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, %xmm1 +; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm1 = xmm1[1,1],xmm0[1,1] +; MACOS-SINCOS-STRET-NEXT: movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1] +; MACOS-SINCOS-STRET-NEXT: callq ___sincosf_stret +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1] +; MACOS-SINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1] +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm2 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: unpcklps {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1] +; MACOS-SINCOS-STRET-NEXT: unpcklpd (%rsp), %xmm2 ## 16-byte Folded Reload +; MACOS-SINCOS-STRET-NEXT: ## xmm2 = xmm2[0],mem[0] +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: unpcklps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Folded Reload +; MACOS-SINCOS-STRET-NEXT: ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1] +; MACOS-SINCOS-STRET-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0] +; MACOS-SINCOS-STRET-NEXT: movups %xmm1, (%r14) +; MACOS-SINCOS-STRET-NEXT: movups %xmm2, (%rbx) +; MACOS-SINCOS-STRET-NEXT: addq $104, %rsp +; MACOS-SINCOS-STRET-NEXT: popq %rbx +; MACOS-SINCOS-STRET-NEXT: popq %r14 +; MACOS-SINCOS-STRET-NEXT: retq +; +; MACOS-NOSINCOS-STRET-LABEL: test_sincos_v4f32: +; MACOS-NOSINCOS-STRET: ## %bb.0: +; MACOS-NOSINCOS-STRET-NEXT: pushq %r14 +; MACOS-NOSINCOS-STRET-NEXT: pushq %rbx +; MACOS-NOSINCOS-STRET-NEXT: subq $104, %rsp +; MACOS-NOSINCOS-STRET-NEXT: movq %rsi, %rbx +; MACOS-NOSINCOS-STRET-NEXT: movq %rdi, %r14 +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,3,3,3] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _cosf +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _cosf +; MACOS-NOSINCOS-STRET-NEXT: unpcklps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Folded Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: callq _cosf +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,1,1,1] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _cosf +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1] +; MACOS-NOSINCOS-STRET-NEXT: unpcklpd {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Folded Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm1 = xmm1[0],mem[0] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: callq _sinf +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: callq _sinf +; MACOS-NOSINCOS-STRET-NEXT: unpcklps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Folded Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm0 = xmm0[0],mem[0],xmm0[1],mem[1] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: callq _sinf +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: callq _sinf +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm1 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: unpcklps {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1] +; MACOS-NOSINCOS-STRET-NEXT: unpcklpd {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Folded Reload +; MACOS-NOSINCOS-STRET-NEXT: ## xmm1 = xmm1[0],mem[0] +; MACOS-NOSINCOS-STRET-NEXT: movups %xmm1, (%r14) +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: movups %xmm0, (%rbx) +; MACOS-NOSINCOS-STRET-NEXT: addq $104, %rsp +; MACOS-NOSINCOS-STRET-NEXT: popq %rbx +; MACOS-NOSINCOS-STRET-NEXT: popq %r14 +; MACOS-NOSINCOS-STRET-NEXT: retq %result = call { <4 x float>, <4 x float> } @llvm.sincos.v4f32(<4 x float> %x) %result.0 = extractvalue { <4 x float>, <4 x float> } %result, 0 %result.1 = extractvalue { <4 x float>, <4 x float> } %result, 1 @@ -63,36 +217,120 @@ define void @test_sincos_v4f32(<4 x float> %x, ptr noalias %out_sin, ptr noalias } define void @test_sincos_v2f64(<2 x double> %x, ptr noalias %out_sin, ptr noalias %out_cos) nounwind { -; CHECK-LABEL: test_sincos_v2f64: -; CHECK: # %bb.0: -; CHECK-NEXT: pushl %edi -; CHECK-NEXT: pushl %esi -; CHECK-NEXT: subl $52, %esp -; CHECK-NEXT: movl 84(%esp), %esi -; CHECK-NEXT: fldl 72(%esp) -; CHECK-NEXT: fstpl {{[-0-9]+}}(%e{{[sb]}}p) # 8-byte Folded Spill -; CHECK-NEXT: fldl 64(%esp) -; CHECK-NEXT: movl 80(%esp), %edi -; CHECK-NEXT: leal 24(%esp), %eax -; CHECK-NEXT: movl %eax, 12(%esp) -; CHECK-NEXT: movl %edi, 8(%esp) -; CHECK-NEXT: fstpl (%esp) -; CHECK-NEXT: calll sincos -; CHECK-NEXT: leal 32(%esp), %eax -; CHECK-NEXT: movl %eax, 12(%esp) -; CHECK-NEXT: addl $8, %edi -; CHECK-NEXT: movl %edi, 8(%esp) -; CHECK-NEXT: fldl {{[-0-9]+}}(%e{{[sb]}}p) # 8-byte Folded Reload -; CHECK-NEXT: fstpl (%esp) -; CHECK-NEXT: calll sincos -; CHECK-NEXT: fldl 24(%esp) -; CHECK-NEXT: fldl 32(%esp) -; CHECK-NEXT: fstpl 8(%esi) -; CHECK-NEXT: fstpl (%esi) -; CHECK-NEXT: addl $52, %esp -; CHECK-NEXT: popl %esi -; CHECK-NEXT: popl %edi -; CHECK-NEXT: retl +; X86-LABEL: test_sincos_v2f64: +; X86: # %bb.0: +; X86-NEXT: pushl %edi +; X86-NEXT: pushl %esi +; X86-NEXT: subl $52, %esp +; X86-NEXT: movl 84(%esp), %esi +; X86-NEXT: fldl 72(%esp) +; X86-NEXT: fstpl {{[-0-9]+}}(%e{{[sb]}}p) # 8-byte Folded Spill +; X86-NEXT: fldl 64(%esp) +; X86-NEXT: movl 80(%esp), %edi +; X86-NEXT: leal 24(%esp), %eax +; X86-NEXT: movl %eax, 12(%esp) +; X86-NEXT: movl %edi, 8(%esp) +; X86-NEXT: fstpl (%esp) +; X86-NEXT: calll sincos +; X86-NEXT: leal 32(%esp), %eax +; X86-NEXT: movl %eax, 12(%esp) +; X86-NEXT: addl $8, %edi +; X86-NEXT: movl %edi, 8(%esp) +; X86-NEXT: fldl {{[-0-9]+}}(%e{{[sb]}}p) # 8-byte Folded Reload +; X86-NEXT: fstpl (%esp) +; X86-NEXT: calll sincos +; X86-NEXT: fldl 24(%esp) +; X86-NEXT: fldl 32(%esp) +; X86-NEXT: fstpl 8(%esi) +; X86-NEXT: fstpl (%esi) +; X86-NEXT: addl $52, %esp +; X86-NEXT: popl %esi +; X86-NEXT: popl %edi +; X86-NEXT: retl +; +; X64-LABEL: test_sincos_v2f64: +; X64: # %bb.0: +; X64-NEXT: pushq %r14 +; X64-NEXT: pushq %rbx +; X64-NEXT: subq $56, %rsp +; X64-NEXT: movq %rsi, %rbx +; X64-NEXT: movq %rdi, %r14 +; X64-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill +; X64-NEXT: leaq 24(%rsp), %rdi +; X64-NEXT: leaq 16(%rsp), %rsi +; X64-NEXT: callq sincos@PLT +; X64-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload +; X64-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1] +; X64-NEXT: leaq 8(%rsp), %rdi +; X64-NEXT: movq %rsp, %rsi +; X64-NEXT: callq sincos@PLT +; X64-NEXT: movsd {{.*#+}} xmm0 = mem[0],zero +; X64-NEXT: movhps {{.*#+}} xmm0 = xmm0[0,1],mem[0,1] +; X64-NEXT: movsd {{.*#+}} xmm1 = mem[0],zero +; X64-NEXT: movhps {{.*#+}} xmm1 = xmm1[0,1],mem[0,1] +; X64-NEXT: movups %xmm1, (%r14) +; X64-NEXT: movups %xmm0, (%rbx) +; X64-NEXT: addq $56, %rsp +; X64-NEXT: popq %rbx +; X64-NEXT: popq %r14 +; X64-NEXT: retq +; +; MACOS-SINCOS-STRET-LABEL: test_sincos_v2f64: +; MACOS-SINCOS-STRET: ## %bb.0: +; MACOS-SINCOS-STRET-NEXT: pushq %r14 +; MACOS-SINCOS-STRET-NEXT: pushq %rbx +; MACOS-SINCOS-STRET-NEXT: subq $56, %rsp +; MACOS-SINCOS-STRET-NEXT: movq %rsi, %rbx +; MACOS-SINCOS-STRET-NEXT: movq %rdi, %r14 +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: callq ___sincos_stret +; MACOS-SINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-SINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1] +; MACOS-SINCOS-STRET-NEXT: callq ___sincos_stret +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm2 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: movlhps {{.*#+}} xmm2 = xmm2[0],xmm1[0] +; MACOS-SINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload +; MACOS-SINCOS-STRET-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0] +; MACOS-SINCOS-STRET-NEXT: movups %xmm1, (%r14) +; MACOS-SINCOS-STRET-NEXT: movups %xmm2, (%rbx) +; MACOS-SINCOS-STRET-NEXT: addq $56, %rsp +; MACOS-SINCOS-STRET-NEXT: popq %rbx +; MACOS-SINCOS-STRET-NEXT: popq %r14 +; MACOS-SINCOS-STRET-NEXT: retq +; +; MACOS-NOSINCOS-STRET-LABEL: test_sincos_v2f64: +; MACOS-NOSINCOS-STRET: ## %bb.0: +; MACOS-NOSINCOS-STRET-NEXT: pushq %r14 +; MACOS-NOSINCOS-STRET-NEXT: pushq %rbx +; MACOS-NOSINCOS-STRET-NEXT: subq $56, %rsp +; MACOS-NOSINCOS-STRET-NEXT: movq %rsi, %rbx +; MACOS-NOSINCOS-STRET-NEXT: movq %rdi, %r14 +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _cos +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: movhlps {{.*#+}} xmm0 = xmm0[1,1] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: callq _cos +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0] +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: callq _sin +; MACOS-NOSINCOS-STRET-NEXT: movaps %xmm0, (%rsp) ## 16-byte Spill +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: callq _sin +; MACOS-NOSINCOS-STRET-NEXT: movaps (%rsp), %xmm1 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: movlhps {{.*#+}} xmm1 = xmm1[0],xmm0[0] +; MACOS-NOSINCOS-STRET-NEXT: movups %xmm1, (%r14) +; MACOS-NOSINCOS-STRET-NEXT: movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload +; MACOS-NOSINCOS-STRET-NEXT: movups %xmm0, (%rbx) +; MACOS-NOSINCOS-STRET-NEXT: addq $56, %rsp +; MACOS-NOSINCOS-STRET-NEXT: popq %rbx +; MACOS-NOSINCOS-STRET-NEXT: popq %r14 +; MACOS-NOSINCOS-STRET-NEXT: retq %result = call { <2 x double>, <2 x double> } @llvm.sincos.v2f64(<2 x double> %x) %result.0 = extractvalue { <2 x double>, <2 x double> } %result, 0 %result.1 = extractvalue { <2 x double>, <2 x double> } %result, 1 |