6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2025-11-23 16:19:52 +00:00
Code Issues Projects Releases Wiki Activity

TargetLowering: Fix getTypeConversion handling of extended vector types

Browse Source 
The code in getTypeConversion attempts to promote the element vector type
before it trys to split or widen the vector.
After it failed finding a legal vector type by promoting it would continue using
the promoted vector element type. Thereby missing legal splitted vector types.
For example the type v32i32 that has a legal split of 4 x v3i32 on x86/sse2
would be transformed to: v32i256 and from there on successively split to:
v16i256, v8i256, v1i256 and then finally ends up as an i64 type.
By resetting the vector element type to the original vector element type that
existed before the promotion the code will attempt to split the vector type to
smaller vector widths of the same type.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@178999 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Arnold Schwaighofer
2013-04-07 20:22:56 +00:00
parent 8764c8979c
commit cd3d60c450
4 changed files with 17 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
test/Analysis/CostModel/X86/testshiftashr.ll
View File

@@ -113,7 +113,7 @@ entry:
define %shifttype32i32 @shift32i32(%shifttype32i32 %a, %shifttype32i32 %b) {
entry:
; SSE2: shift32i32
; SSE2: cost of 256 {{.*}} ashr
; SSE2: cost of 320 {{.*}} ashr
; SSE2-CODEGEN: shift32i32
; SSE2-CODEGEN: sarl %cl
@@ -173,7 +173,7 @@ entry:
define %shifttype32i64 @shift32i64(%shifttype32i64 %a, %shifttype32i64 %b) {
entry:
; SSE2: shift32i64
; SSE2: cost of 256 {{.*}} ashr
; SSE2: cost of 320 {{.*}} ashr
; SSE2-CODEGEN: shift32i64
; SSE2-CODEGEN: sarq %cl
@@ -373,7 +373,7 @@ define %shifttypec32i32 @shift32i32c(%shifttypec32i32 %a, %shifttypec32i32 %b) {
entry:
; SSE2: shift32i32c
; getTypeConversion fails here and promotes this to a i64.
; SSE2: cost of 256 {{.*}} ashr
; SSE2: cost of 8 {{.*}} ashr
; SSE2-CODEGEN: shift32i32c
; SSE2-CODEGEN: psrad $3
%0 = ashr %shifttypec32i32 %a , <i32 3, i32 3, i32 3, i32 3,
@@ -443,7 +443,7 @@ entry:
define %shifttypec32i64 @shift32i64c(%shifttypec32i64 %a, %shifttypec32i64 %b) {
entry:
; SSE2: shift32i64c
; SSE2: cost of 256 {{.*}} ashr
; SSE2: cost of 320 {{.*}} ashr
; SSE2-CODEGEN: shift32i64c
; SSE2-CODEGEN: sarq $3