2016-04-02 01:40:25 +08:00
|
|
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
2017-05-10 23:52:59 +08:00
|
|
|
; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X86
|
|
|
|
; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X64
|
2013-04-30 06:41:29 +08:00
|
|
|
|
Finish implementing a readme entry: when inserting an i64 variable
into a vector of zeros or undef, and when the top part is obviously
zero, we can just use movd + shuffle. This allows us to compile
vec_set-B.ll into:
_test3:
movl $1234567, %eax
andl 4(%esp), %eax
movd %eax, %xmm0
ret
instead of:
_test3:
subl $28, %esp
movl $1234567, %eax
andl 32(%esp), %eax
movl %eax, (%esp)
movl $0, 4(%esp)
movq (%esp), %xmm0
addl $28, %esp
ret
llvm-svn: 48090
2008-03-09 13:42:06 +08:00
|
|
|
; These should both generate something like this:
|
|
|
|
;_test3:
|
|
|
|
; movl $1234567, %eax
|
|
|
|
; andl 4(%esp), %eax
|
|
|
|
; movd %eax, %xmm0
|
|
|
|
; ret
|
|
|
|
|
2008-05-08 06:59:08 +08:00
|
|
|
define <2 x i64> @test3(i64 %arg) nounwind {
|
2017-05-10 23:52:59 +08:00
|
|
|
; X86-LABEL: test3:
|
|
|
|
; X86: # BB#0:
|
|
|
|
; X86-NEXT: movl $1234567, %eax # imm = 0x12D687
|
|
|
|
; X86-NEXT: andl {{[0-9]+}}(%esp), %eax
|
|
|
|
; X86-NEXT: movd %eax, %xmm0
|
|
|
|
; X86-NEXT: retl
|
|
|
|
;
|
|
|
|
; X64-LABEL: test3:
|
|
|
|
; X64: # BB#0:
|
|
|
|
; X64-NEXT: andl $1234567, %edi # imm = 0x12D687
|
|
|
|
; X64-NEXT: movq %rdi, %xmm0
|
|
|
|
; X64-NEXT: retq
|
2016-04-02 01:40:25 +08:00
|
|
|
%A = and i64 %arg, 1234567
|
|
|
|
%B = insertelement <2 x i64> zeroinitializer, i64 %A, i32 0
|
|
|
|
ret <2 x i64> %B
|
Finish implementing a readme entry: when inserting an i64 variable
into a vector of zeros or undef, and when the top part is obviously
zero, we can just use movd + shuffle. This allows us to compile
vec_set-B.ll into:
_test3:
movl $1234567, %eax
andl 4(%esp), %eax
movd %eax, %xmm0
ret
instead of:
_test3:
subl $28, %esp
movl $1234567, %eax
andl 32(%esp), %eax
movl %eax, (%esp)
movl $0, 4(%esp)
movq (%esp), %xmm0
addl $28, %esp
ret
llvm-svn: 48090
2008-03-09 13:42:06 +08:00
|
|
|
}
|
|
|
|
|
2008-05-08 06:59:08 +08:00
|
|
|
define <2 x i64> @test2(i64 %arg) nounwind {
|
2017-05-10 23:52:59 +08:00
|
|
|
; X86-LABEL: test2:
|
|
|
|
; X86: # BB#0:
|
|
|
|
; X86-NEXT: movl $1234567, %eax # imm = 0x12D687
|
|
|
|
; X86-NEXT: andl {{[0-9]+}}(%esp), %eax
|
|
|
|
; X86-NEXT: movd %eax, %xmm0
|
|
|
|
; X86-NEXT: retl
|
|
|
|
;
|
|
|
|
; X64-LABEL: test2:
|
|
|
|
; X64: # BB#0:
|
|
|
|
; X64-NEXT: andl $1234567, %edi # imm = 0x12D687
|
|
|
|
; X64-NEXT: movq %rdi, %xmm0
|
|
|
|
; X64-NEXT: retq
|
2016-04-02 01:40:25 +08:00
|
|
|
%A = and i64 %arg, 1234567
|
|
|
|
%B = insertelement <2 x i64> undef, i64 %A, i32 0
|
|
|
|
ret <2 x i64> %B
|
Finish implementing a readme entry: when inserting an i64 variable
into a vector of zeros or undef, and when the top part is obviously
zero, we can just use movd + shuffle. This allows us to compile
vec_set-B.ll into:
_test3:
movl $1234567, %eax
andl 4(%esp), %eax
movd %eax, %xmm0
ret
instead of:
_test3:
subl $28, %esp
movl $1234567, %eax
andl 32(%esp), %eax
movl %eax, (%esp)
movl $0, 4(%esp)
movq (%esp), %xmm0
addl $28, %esp
ret
llvm-svn: 48090
2008-03-09 13:42:06 +08:00
|
|
|
}
|