6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2025-11-03 14:21:30 +00:00
Code Issues Projects Releases Wiki Activity

[AArch64] Prefer DUP/MOV ("CPY") to INS for vector_extract.

Browse Source 
This avoids a partial false dependency on the previous content of
the upper lanes of the destination vector register.

Differential Revision: http://reviews.llvm.org/D7307


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227820 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Ahmed Bougacha
2015-02-02 17:55:57 +00:00
parent 5fc4d409c6
commit 0f1a21bcb8
5 changed files with 16 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
test/CodeGen/AArch64/neon-scalar-copy.ll
View File

@@ -2,7 +2,7 @@
define float @test_dup_sv2S(<2 x float> %v) #0 {
; CHECK-LABEL: test_dup_sv2S:
; CHECK-NEXT: ins {{v[0-9]+}}.s[0], {{v[0-9]+}}.s[1]
; CHECK-NEXT: mov s{{[0-9]+}}, {{v[0-9]+}}.s[1]
; CHECK-NEXT: ret
%tmp1 = extractelement <2 x float> %v, i32 1
ret float %tmp1
@@ -19,7 +19,7 @@ define float @test_dup_sv2S_0(<2 x float> %v) #0 {
define float @test_dup_sv4S(<4 x float> %v) #0 {
; CHECK-LABEL: test_dup_sv4S:
; CHECK-NEXT: ins {{v[0-9]+}}.s[0], {{v[0-9]+}}.s[1]
; CHECK-NEXT: mov s{{[0-9]+}}, {{v[0-9]+}}.s[1]
; CHECK-NEXT: ret
%tmp1 = extractelement <4 x float> %v, i32 1
ret float %tmp1
@@ -45,7 +45,7 @@ define double @test_dup_dvD(<1 x double> %v) #0 {
define double @test_dup_dv2D(<2 x double> %v) #0 {
; CHECK-LABEL: test_dup_dv2D:
; CHECK-NEXT: ins {{v[0-9]+}}.d[0], {{v[0-9]+}}.d[1]
; CHECK-NEXT: mov d{{[0-9]+}}, {{v[0-9]+}}.d[1]
; CHECK-NEXT: ret
%tmp1 = extractelement <2 x double> %v, i32 1
ret double %tmp1
@@ -62,7 +62,7 @@ define double @test_dup_dv2D_0(<2 x double> %v) #0 {
define half @test_dup_hv8H(<8 x half> %v) #0 {
; CHECK-LABEL: test_dup_hv8H:
; CHECK-NEXT: ins {{v[0-9]+}}.h[0], {{v[0-9]+}}.h[1]
; CHECK-NEXT: mov h{{[0-9]+}}, {{v[0-9]+}}.h[1]
; CHECK-NEXT: ret
%tmp1 = extractelement <8 x half> %v, i32 1
ret half %tmp1