maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/CodeGen/X86/widen_shuffle-1.ll

72 lines
3.1 KiB
LLVM
Raw Normal View History

Update tests to use sse4.2 instead of sse42. llvm-svn: 189145
2013-08-24 04:46:35 +08:00
; RUN: llc < %s -march=x86 -mattr=+sse4.2 | FileCheck %s
Added some basic test cases for r61209 llvm-svn: 61210
2008-12-19 04:05:58 +08:00
; widening shuffle v3float and then a add
define void @shuf(<3 x float>* %dst.addr, <3 x float> %src1,<3 x float> %src2) nounwind {
entry:
Mass update to CodeGen tests to use CHECK-LABEL for labels corresponding to function definitions for more informative error messages. No functionality change and all updated tests passed locally. This update was done with the following bash script: find test/CodeGen -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc.*debug" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_-]*\):\( *\)$FUNC: *\$/;\1\2-LABEL:\3$FUNC:/g" $TEMP done sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP mv $TEMP $NAME fi done llvm-svn: 186280
2013-07-14 14:24:09 +08:00
; CHECK-LABEL: shuf:
Change handling of illegal vector types to widen when possible instead of expanding: e.g. <2 x float> -> <4 x float> instead of -> 2 floats. This affects two places in the code: handling cross block values and handling function return and arguments. Since vectors are already widened by legalizetypes, this gives us much better code and unblocks x86-64 abi and SPU abi work. For example, this (which is a silly example of a cross-block value): define <4 x float> @test2(<4 x float> %A) nounwind { %B = shufflevector <4 x float> %A, <4 x float> undef, <2 x i32> <i32 0, i32 1> %C = fadd <2 x float> %B, %B br label %BB BB: %D = fadd <2 x float> %C, %C %E = shufflevector <2 x float> %D, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef> ret <4 x float> %E } Now compiles into: _test2: ## @test2 ## BB#0: addps %xmm0, %xmm0 addps %xmm0, %xmm0 ret previously it compiled into: _test2: ## @test2 ## BB#0: addps %xmm0, %xmm0 pshufd $1, %xmm0, %xmm1 ## kill: XMM0<def> XMM0<kill> XMM0<def> insertps $0, %xmm0, %xmm0 insertps $16, %xmm1, %xmm0 addps %xmm0, %xmm0 ret This implements rdar://8230384 llvm-svn: 112101
2010-08-26 06:49:25 +08:00
; CHECK: extractps
Fixed a bug during widening where we would avoid legalizing a node. When we replace an OpA with a widened OpB, it is possible to get new uses of OpA due to CSE when recursively updating nodes. Since OpA has been processed, the new uses are not examined again. The patch checks if this occurred and it it did, updates the new uses of OpA to use OpB. llvm-svn: 105453
2010-06-04 09:20:10 +08:00
; CHECK: extractps
Added some basic test cases for r61209 llvm-svn: 61210
2008-12-19 04:05:58 +08:00
%x = shufflevector <3 x float> %src1, <3 x float> %src2, <3 x i32> < i32 0, i32 1, i32 2>
Delete useless trailing semicolons. llvm-svn: 92740
2010-01-06 01:55:26 +08:00
%val = fadd <3 x float> %x, %src2
Added some basic test cases for r61209 llvm-svn: 61210
2008-12-19 04:05:58 +08:00
store <3 x float> %val, <3 x float>* %dst.addr
ret void
Transform: (EXTRACT_VECTOR_ELT( VECTOR_SHUFFLE )) -> EXTRACT_VECTOR_ELT. llvm-svn: 148337
2012-01-18 05:44:01 +08:00
; CHECK: ret
Added some basic test cases for r61209 llvm-svn: 61210
2008-12-19 04:05:58 +08:00
}
Fixed a bug during widening where we would avoid legalizing a node. When we replace an OpA with a widened OpB, it is possible to get new uses of OpA due to CSE when recursively updating nodes. Since OpA has been processed, the new uses are not examined again. The patch checks if this occurred and it it did, updates the new uses of OpA to use OpB. llvm-svn: 105453
2010-06-04 09:20:10 +08:00
; widening shuffle v3float with a different mask and then a add
define void @shuf2(<3 x float>* %dst.addr, <3 x float> %src1,<3 x float> %src2) nounwind {
entry:
Mass update to CodeGen tests to use CHECK-LABEL for labels corresponding to function definitions for more informative error messages. No functionality change and all updated tests passed locally. This update was done with the following bash script: find test/CodeGen -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc.*debug" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_-]*\):\( *\)$FUNC: *\$/;\1\2-LABEL:\3$FUNC:/g" $TEMP done sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP mv $TEMP $NAME fi done llvm-svn: 186280
2013-07-14 14:24:09 +08:00
; CHECK-LABEL: shuf2:
Change handling of illegal vector types to widen when possible instead of expanding: e.g. <2 x float> -> <4 x float> instead of -> 2 floats. This affects two places in the code: handling cross block values and handling function return and arguments. Since vectors are already widened by legalizetypes, this gives us much better code and unblocks x86-64 abi and SPU abi work. For example, this (which is a silly example of a cross-block value): define <4 x float> @test2(<4 x float> %A) nounwind { %B = shufflevector <4 x float> %A, <4 x float> undef, <2 x i32> <i32 0, i32 1> %C = fadd <2 x float> %B, %B br label %BB BB: %D = fadd <2 x float> %C, %C %E = shufflevector <2 x float> %D, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef> ret <4 x float> %E } Now compiles into: _test2: ## @test2 ## BB#0: addps %xmm0, %xmm0 addps %xmm0, %xmm0 ret previously it compiled into: _test2: ## @test2 ## BB#0: addps %xmm0, %xmm0 pshufd $1, %xmm0, %xmm1 ## kill: XMM0<def> XMM0<kill> XMM0<def> insertps $0, %xmm0, %xmm0 insertps $16, %xmm1, %xmm0 addps %xmm0, %xmm0 ret This implements rdar://8230384 llvm-svn: 112101
2010-08-26 06:49:25 +08:00
; CHECK: extractps
Fixed a bug during widening where we would avoid legalizing a node. When we replace an OpA with a widened OpB, it is possible to get new uses of OpA due to CSE when recursively updating nodes. Since OpA has been processed, the new uses are not examined again. The patch checks if this occurred and it it did, updates the new uses of OpA to use OpB. llvm-svn: 105453
2010-06-04 09:20:10 +08:00
; CHECK: extractps
%x = shufflevector <3 x float> %src1, <3 x float> %src2, <3 x i32> < i32 0, i32 4, i32 2>
%val = fadd <3 x float> %x, %src2
store <3 x float> %val, <3 x float>* %dst.addr
ret void
Transform: (EXTRACT_VECTOR_ELT( VECTOR_SHUFFLE )) -> EXTRACT_VECTOR_ELT. llvm-svn: 148337
2012-01-18 05:44:01 +08:00
; CHECK: ret
Fixed a bug during widening where we would avoid legalizing a node. When we replace an OpA with a widened OpB, it is possible to get new uses of OpA due to CSE when recursively updating nodes. Since OpA has been processed, the new uses are not examined again. The patch checks if this occurred and it it did, updates the new uses of OpA to use OpB. llvm-svn: 105453
2010-06-04 09:20:10 +08:00
}
; Example of when widening a v3float operation causes the DAG to replace a node
; with the operation that we are currently widening, i.e. when replacing
; opA with opB, the DAG will produce new operations with opA.
Change handling of illegal vector types to widen when possible instead of expanding: e.g. <2 x float> -> <4 x float> instead of -> 2 floats. This affects two places in the code: handling cross block values and handling function return and arguments. Since vectors are already widened by legalizetypes, this gives us much better code and unblocks x86-64 abi and SPU abi work. For example, this (which is a silly example of a cross-block value): define <4 x float> @test2(<4 x float> %A) nounwind { %B = shufflevector <4 x float> %A, <4 x float> undef, <2 x i32> <i32 0, i32 1> %C = fadd <2 x float> %B, %B br label %BB BB: %D = fadd <2 x float> %C, %C %E = shufflevector <2 x float> %D, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 undef, i32 undef> ret <4 x float> %E } Now compiles into: _test2: ## @test2 ## BB#0: addps %xmm0, %xmm0 addps %xmm0, %xmm0 ret previously it compiled into: _test2: ## @test2 ## BB#0: addps %xmm0, %xmm0 pshufd $1, %xmm0, %xmm1 ## kill: XMM0<def> XMM0<kill> XMM0<def> insertps $0, %xmm0, %xmm0 insertps $16, %xmm1, %xmm0 addps %xmm0, %xmm0 ret This implements rdar://8230384 llvm-svn: 112101
2010-08-26 06:49:25 +08:00
define void @shuf3(<4 x float> %tmp10, <4 x float> %vecinit15, <4 x float>* %dst) nounwind {
Fixed a bug during widening where we would avoid legalizing a node. When we replace an OpA with a widened OpB, it is possible to get new uses of OpA due to CSE when recursively updating nodes. Since OpA has been processed, the new uses are not examined again. The patch checks if this occurred and it it did, updates the new uses of OpA to use OpB. llvm-svn: 105453
2010-06-04 09:20:10 +08:00
entry:
Mass update to CodeGen tests to use CHECK-LABEL for labels corresponding to function definitions for more informative error messages. No functionality change and all updated tests passed locally. This update was done with the following bash script: find test/CodeGen -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc.*debug" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_-]*\):\( *\)$FUNC: *\$/;\1\2-LABEL:\3$FUNC:/g" $TEMP done sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP mv $TEMP $NAME fi done llvm-svn: 186280
2013-07-14 14:24:09 +08:00
; CHECK-LABEL: shuf3:
[DAG] Further improve the logic in DAGCombiner that folds a pair of shuffles into a single shuffle if the resulting mask is legal. This patch teaches the DAGCombiner how to fold shuffles according to the following new rules: 1. shuffle(shuffle(x, y), undef) -> x 2. shuffle(shuffle(x, y), undef) -> y 3. shuffle(shuffle(x, y), undef) -> shuffle(x, undef) 4. shuffle(shuffle(x, y), undef) -> shuffle(y, undef) The backend avoids to combine shuffles according to rules 3. and 4. if the resulting shuffle does not have a legal mask. This is to avoid introducing illegal shuffles that are potentially expanded into a sub-optimal sequence of target specific dag nodes during vector legalization. Added test case combine-vec-shuffle-2.ll to verify that we correctly triggers the new rules when combining shuffles. llvm-svn: 212748
2014-07-11 02:04:55 +08:00
; CHECK-NOT: movlhps
; CHECK-NOT: shufps
; CHECK: pshufd
Fixed a bug during widening where we would avoid legalizing a node. When we replace an OpA with a widened OpB, it is possible to get new uses of OpA due to CSE when recursively updating nodes. Since OpA has been processed, the new uses are not examined again. The patch checks if this occurred and it it did, updates the new uses of OpA to use OpB. llvm-svn: 105453
2010-06-04 09:20:10 +08:00
%shuffle.i.i.i12 = shufflevector <4 x float> %tmp10, <4 x float> %vecinit15, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
%tmp25.i.i = shufflevector <4 x float> %shuffle.i.i.i12, <4 x float> undef, <3 x i32> <i32 0, i32 1, i32 2>
%tmp1.i.i = shufflevector <3 x float> %tmp25.i.i, <3 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%tmp3.i13 = shufflevector <4 x float> %tmp1.i.i, <4 x float> undef, <3 x i32> <i32 0, i32 1, i32 2> ; <<3 x float>>
%tmp6.i14 = shufflevector <3 x float> %tmp3.i13, <3 x float> zeroinitializer, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%tmp97.i = shufflevector <4 x float> %tmp6.i14, <4 x float> undef, <3 x i32> <i32 0, i32 1, i32 2>
%tmp2.i18 = shufflevector <3 x float> %tmp97.i, <3 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 2>
%t5 = bitcast <4 x float> %tmp2.i18 to <4 x i32>
%shr.i.i19 = lshr <4 x i32> %t5, <i32 19, i32 19, i32 19, i32 19>
%and.i.i20 = and <4 x i32> %shr.i.i19, <i32 4080, i32 4080, i32 4080, i32 4080>
%shuffle.i.i.i21 = shufflevector <4 x float> %tmp2.i18, <4 x float> undef, <4 x i32> <i32 2, i32 3, i32 2, i32 3>
store <4 x float> %shuffle.i.i.i21, <4 x float>* %dst
ret void
Transform: (EXTRACT_VECTOR_ELT( VECTOR_SHUFFLE )) -> EXTRACT_VECTOR_ELT. llvm-svn: 148337
2012-01-18 05:44:01 +08:00
; CHECK: ret
Fixed a bug during widening where we would avoid legalizing a node. When we replace an OpA with a widened OpB, it is possible to get new uses of OpA due to CSE when recursively updating nodes. Since OpA has been processed, the new uses are not examined again. The patch checks if this occurred and it it did, updates the new uses of OpA to use OpB. llvm-svn: 105453
2010-06-04 09:20:10 +08:00
}
PR10421: Fix a straightforward bug in the widening logic for CONCAT_VECTORS. llvm-svn: 135595
2011-07-21 02:14:33 +08:00
; PR10421: make sure we correctly handle extreme widening with CONCAT_VECTORS
define <8 x i8> @shuf4(<4 x i8> %a, <4 x i8> %b) nounwind readnone {
Mass update to CodeGen tests to use CHECK-LABEL for labels corresponding to function definitions for more informative error messages. No functionality change and all updated tests passed locally. This update was done with the following bash script: find test/CodeGen -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc.*debug" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_-]*\):\( *\)$FUNC: *\$/;\1\2-LABEL:\3$FUNC:/g" $TEMP done sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP mv $TEMP $NAME fi done llvm-svn: 186280
2013-07-14 14:24:09 +08:00
; CHECK-LABEL: shuf4:
Enable element promotion type legalization by deafault. Changed tests which assumed that vectors are legalized by widening them. llvm-svn: 142152
2011-10-17 04:31:33 +08:00
; CHECK-NOT: punpckldq
PR10421: Fix a straightforward bug in the widening logic for CONCAT_VECTORS. llvm-svn: 135595
2011-07-21 02:14:33 +08:00
%vshuf = shufflevector <4 x i8> %a, <4 x i8> %b, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
ret <8 x i8> %vshuf
Transform: (EXTRACT_VECTOR_ELT( VECTOR_SHUFFLE )) -> EXTRACT_VECTOR_ELT. llvm-svn: 148337
2012-01-18 05:44:01 +08:00
; CHECK: ret
PR10421: Fix a straightforward bug in the widening logic for CONCAT_VECTORS. llvm-svn: 135595
2011-07-21 02:14:33 +08:00
}
CONCAT_VECTORS can have more than two operands. PR11389. llvm-svn: 144768
2011-11-16 10:52:39 +08:00
; PR11389: another CONCAT_VECTORS case
define void @shuf5(<8 x i8>* %p) nounwind {
Mass update to CodeGen tests to use CHECK-LABEL for labels corresponding to function definitions for more informative error messages. No functionality change and all updated tests passed locally. This update was done with the following bash script: find test/CodeGen -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc.*debug" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_-]*\):\( *\)$FUNC: *\$/;\1\2-LABEL:\3$FUNC:/g" $TEMP done sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP mv $TEMP $NAME fi done llvm-svn: 186280
2013-07-14 14:24:09 +08:00
; CHECK-LABEL: shuf5:
CONCAT_VECTORS can have more than two operands. PR11389. llvm-svn: 144768
2011-11-16 10:52:39 +08:00
%v = shufflevector <2 x i8> <i8 4, i8 33>, <2 x i8> undef, <8 x i32> <i32 1, i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
store <8 x i8> %v, <8 x i8>* %p, align 8
ret void
Transform: (EXTRACT_VECTOR_ELT( VECTOR_SHUFFLE )) -> EXTRACT_VECTOR_ELT. llvm-svn: 148337
2012-01-18 05:44:01 +08:00
; CHECK: ret
CONCAT_VECTORS can have more than two operands. PR11389. llvm-svn: 144768
2011-11-16 10:52:39 +08:00
}