forked from OSchip/llvm-project
413a6cb42b
when the input is a concat_vectors and the insert replaces one of the
concat halves:
Lower half: fold (insert_subvector (concat_vectors X, Y), Z) ->
(concat_vectors Z, Y)
Upper half: fold (insert_subvector (concat_vectors X, Y), Z) ->
(concat_vectors X, Z)
This can be seen with the following IR:
define <8 x float> @lower_half(<4 x float> %v1, <4 x float> %v2, <4 x
float> %v3) {
%1 = shufflevector <4 x float> %v1, <4 x float> %v2, <8 x i32> <i32
0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
%2 = tail call <8 x float> @llvm.x86.avx.vinsertf128.ps.256(<8 x
float> %1, <4 x float> %v3, i8 0)
The vinsertf128 intrinsic is converted into an insert_subvector node
in SelectionDAGBuilder.cpp.
Using AVX, without the patch this generates two vinsertf128 instructions:
vinsertf128 $1, %xmm1, %ymm0, %ymm0
vinsertf128 $0, %xmm2, %ymm0, %ymm0
With the patch this is optimized into:
vinsertf128 $1, %xmm1, %ymm2, %ymm0
Patch by Robert Lougher.
llvm-svn: 200506
|
||
|---|---|---|
| .. | ||
| CMakeLists.txt | ||
| DAGCombiner.cpp | ||
| FastISel.cpp | ||
| FunctionLoweringInfo.cpp | ||
| InstrEmitter.cpp | ||
| InstrEmitter.h | ||
| LLVMBuild.txt | ||
| LegalizeDAG.cpp | ||
| LegalizeFloatTypes.cpp | ||
| LegalizeIntegerTypes.cpp | ||
| LegalizeTypes.cpp | ||
| LegalizeTypes.h | ||
| LegalizeTypesGeneric.cpp | ||
| LegalizeVectorOps.cpp | ||
| LegalizeVectorTypes.cpp | ||
| Makefile | ||
| ResourcePriorityQueue.cpp | ||
| SDNodeDbgValue.h | ||
| ScheduleDAGFast.cpp | ||
| ScheduleDAGRRList.cpp | ||
| ScheduleDAGSDNodes.cpp | ||
| ScheduleDAGSDNodes.h | ||
| ScheduleDAGVLIW.cpp | ||
| SelectionDAG.cpp | ||
| SelectionDAGBuilder.cpp | ||
| SelectionDAGBuilder.h | ||
| SelectionDAGDumper.cpp | ||
| SelectionDAGISel.cpp | ||
| SelectionDAGPrinter.cpp | ||
| TargetLowering.cpp | ||
| TargetSelectionDAGInfo.cpp | ||