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/vec_insert-5.ll

166 lines
6.0 KiB
LLVM
Raw Normal View History

Make utils/update_llc_test_checks.py note that the assertions are autogenerated. Also update existing test cases which appear to be generated by it and weren't modified (other than addition of the header) by rerunning it. llvm-svn: 253917
2015-11-24 05:33:58 +08:00
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
[X86][SSE] Regenerated the vec_insert tests. llvm-svn: 265179
2016-04-02 03:42:23 +08:00
; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse2,+ssse3 | FileCheck %s --check-prefix=X32
; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse2,+ssse3 | FileCheck %s --check-prefix=X64
Massive rewrite of MMX: The x86_mmx type is used for MMX intrinsics, parameters and return values where these use MMX registers, and is also supported in load, store, and bitcast. Only the above operations generate MMX instructions, and optimizations do not operate on or produce MMX intrinsics. MMX-sized vectors <2 x i32> etc. are lowered to XMM or split into smaller pieces. Optimizations may occur on these forms and the result casted back to x86_mmx, provided the result feeds into a previous existing x86_mmx operation. The point of all this is prevent optimizations from introducing MMX operations, which is unsafe due to the EMMS problem. llvm-svn: 115243
2010-10-01 07:57:10 +08:00
; There are no MMX operations in @t1
Implement vector shift up / down and insert zero with ps{rl}lq / ps{rl}ldq. llvm-svn: 51667
2008-05-29 16:22:04 +08:00
Massive rewrite of MMX: The x86_mmx type is used for MMX intrinsics, parameters and return values where these use MMX registers, and is also supported in load, store, and bitcast. Only the above operations generate MMX instructions, and optimizations do not operate on or produce MMX intrinsics. MMX-sized vectors <2 x i32> etc. are lowered to XMM or split into smaller pieces. Optimizations may occur on these forms and the result casted back to x86_mmx, provided the result feeds into a previous existing x86_mmx operation. The point of all this is prevent optimizations from introducing MMX operations, which is unsafe due to the EMMS problem. llvm-svn: 115243
2010-10-01 07:57:10 +08:00
define void @t1(i32 %a, x86_mmx* %P) nounwind {
[X86][SSE] Regenerated the vec_insert tests. llvm-svn: 265179
2016-04-02 03:42:23 +08:00
; X32-LABEL: t1:
; X32: # BB#0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movl {{[0-9]+}}(%esp), %ecx
; X32-NEXT: shll $12, %ecx
; X32-NEXT: movd %ecx, %xmm0
; X32-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,0,1,1]
; X32-NEXT: movq %xmm0, (%eax)
; X32-NEXT: retl
;
; X64-LABEL: t1:
; X64: # BB#0:
VirtRegMap: Replace some identity copies with KILL instructions. An identity COPY like this: %AL = COPY %AL, %EAX<imp-def> has no semantic effect, but encodes liveness information: Further users of %EAX only depend on this instruction even though it does not define the full register. Replace the COPY with a KILL instruction in those cases to maintain this liveness information. (This reverts a small part of r238588 but this time adds a comment explaining why a KILL instruction is useful). llvm-svn: 274952
2016-07-09 08:19:07 +08:00
; X64-NEXT: # kill: %EDI<def> %EDI<kill> %RDI<def>
[X86][SSE] Regenerated the vec_insert tests. llvm-svn: 265179
2016-04-02 03:42:23 +08:00
; X64-NEXT: shll $12, %edi
[X86][SSE2] Fix asm string for movq (Move Quadword) instruction. Replace "mov{d|q}" with "movq". Differential Revision: https://reviews.llvm.org/D32220 llvm-svn: 301386
2017-04-26 15:08:44 +08:00
; X64-NEXT: movq %rdi, %xmm0
[X86][SSE] Regenerated the vec_insert tests. llvm-svn: 265179
2016-04-02 03:42:23 +08:00
; X64-NEXT: pslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2,3,4,5,6,7]
; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; X64-NEXT: movq %xmm0, (%rsi)
; X64-NEXT: retq
[x86] Regenerate precise checks for a couple of test cases and remove a test case that was just grepping the debug stats output rather than actually checking the generated code for anything useful. llvm-svn: 218951
2014-10-03 09:50:08 +08:00
%tmp12 = shl i32 %a, 12
%tmp21 = insertelement <2 x i32> undef, i32 %tmp12, i32 1
%tmp22 = insertelement <2 x i32> %tmp21, i32 0, i32 0
%tmp23 = bitcast <2 x i32> %tmp22 to x86_mmx
store x86_mmx %tmp23, x86_mmx* %P
ret void
Implement vector shift up / down and insert zero with ps{rl}lq / ps{rl}ldq. llvm-svn: 51667
2008-05-29 16:22:04 +08:00
}
define <4 x float> @t2(<4 x float>* %P) nounwind {
[X86][SSE] Regenerated the vec_insert tests. llvm-svn: 265179
2016-04-02 03:42:23 +08:00
; X32-LABEL: t2:
; X32: # BB#0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movaps (%eax), %xmm1
; X32-NEXT: xorps %xmm0, %xmm0
; X32-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,0],xmm0[2,0]
; X32-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,0]
; X32-NEXT: retl
;
; X64-LABEL: t2:
; X64: # BB#0:
; X64-NEXT: movaps (%rdi), %xmm1
; X64-NEXT: xorps %xmm0, %xmm0
; X64-NEXT: shufps {{.*#+}} xmm1 = xmm1[0,0],xmm0[2,0]
; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,0]
; X64-NEXT: retq
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%tmp1 = load <4 x float>, <4 x float>* %P
[x86] Regenerate precise checks for a couple of test cases and remove a test case that was just grepping the debug stats output rather than actually checking the generated code for anything useful. llvm-svn: 218951
2014-10-03 09:50:08 +08:00
%tmp2 = shufflevector <4 x float> %tmp1, <4 x float> zeroinitializer, <4 x i32> < i32 4, i32 4, i32 4, i32 0 >
ret <4 x float> %tmp2
Implement vector shift up / down and insert zero with ps{rl}lq / ps{rl}ldq. llvm-svn: 51667
2008-05-29 16:22:04 +08:00
}
define <4 x float> @t3(<4 x float>* %P) nounwind {
[X86][SSE] Regenerated the vec_insert tests. llvm-svn: 265179
2016-04-02 03:42:23 +08:00
; X32-LABEL: t3:
; X32: # BB#0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
[X86][SSE] Enable commutation between MOVHLPS and UNPCKHPD Assuming SSE2 is available then we can safely commute between these, removing some unnecessary register moves and improving memory folding opportunities. VEX encoded versions don't benefit so I haven't added support to them. llvm-svn: 277930
2016-08-07 02:40:28 +08:00
; X32-NEXT: xorps %xmm0, %xmm0
; X32-NEXT: movlps {{.*#+}} xmm0 = mem[0,1],xmm0[2,3]
[X86][SSE] Regenerated the vec_insert tests. llvm-svn: 265179
2016-04-02 03:42:23 +08:00
; X32-NEXT: retl
;
; X64-LABEL: t3:
; X64: # BB#0:
[X86][SSE] Enable commutation between MOVHLPS and UNPCKHPD Assuming SSE2 is available then we can safely commute between these, removing some unnecessary register moves and improving memory folding opportunities. VEX encoded versions don't benefit so I haven't added support to them. llvm-svn: 277930
2016-08-07 02:40:28 +08:00
; X64-NEXT: xorps %xmm0, %xmm0
; X64-NEXT: movlps {{.*#+}} xmm0 = mem[0,1],xmm0[2,3]
[X86][SSE] Regenerated the vec_insert tests. llvm-svn: 265179
2016-04-02 03:42:23 +08:00
; X64-NEXT: retq
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%tmp1 = load <4 x float>, <4 x float>* %P
[x86] Regenerate precise checks for a couple of test cases and remove a test case that was just grepping the debug stats output rather than actually checking the generated code for anything useful. llvm-svn: 218951
2014-10-03 09:50:08 +08:00
%tmp2 = shufflevector <4 x float> %tmp1, <4 x float> zeroinitializer, <4 x i32> < i32 2, i32 3, i32 4, i32 4 >
ret <4 x float> %tmp2
Implement vector shift up / down and insert zero with ps{rl}lq / ps{rl}ldq. llvm-svn: 51667
2008-05-29 16:22:04 +08:00
}
define <4 x float> @t4(<4 x float>* %P) nounwind {
[X86][SSE] Regenerated the vec_insert tests. llvm-svn: 265179
2016-04-02 03:42:23 +08:00
; X32-LABEL: t4:
; X32: # BB#0:
; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
; X32-NEXT: movaps (%eax), %xmm0
; X32-NEXT: xorps %xmm1, %xmm1
; X32-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,0],xmm1[1,0]
; X32-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[2,3]
; X32-NEXT: retl
;
; X64-LABEL: t4:
; X64: # BB#0:
; X64-NEXT: movaps (%rdi), %xmm0
; X64-NEXT: xorps %xmm1, %xmm1
; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[3,0],xmm1[1,0]
; X64-NEXT: shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[2,3]
; X64-NEXT: retq
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%tmp1 = load <4 x float>, <4 x float>* %P
[x86] Regenerate precise checks for a couple of test cases and remove a test case that was just grepping the debug stats output rather than actually checking the generated code for anything useful. llvm-svn: 218951
2014-10-03 09:50:08 +08:00
%tmp2 = shufflevector <4 x float> zeroinitializer, <4 x float> %tmp1, <4 x i32> < i32 7, i32 0, i32 0, i32 0 >
ret <4 x float> %tmp2
Implement vector shift up / down and insert zero with ps{rl}lq / ps{rl}ldq. llvm-svn: 51667
2008-05-29 16:22:04 +08:00
}
X86: Make shuffle -> shift conversion more aggressive about undefs. Shuffles that only move an element into position 0 of the vector are common in the output of the loop vectorizer and often generate suboptimal code when SSSE3 is not available. Lower them to vector shifts if possible. We still prefer palignr over psrldq because it has higher throughput on sandybridge. llvm-svn: 182102
2013-05-17 22:48:34 +08:00
define <16 x i8> @t5(<16 x i8> %x) nounwind {
[X86][SSE] Regenerated the vec_insert tests. llvm-svn: 265179
2016-04-02 03:42:23 +08:00
; X32-LABEL: t5:
; X32: # BB#0:
; X32-NEXT: psrlw $8, %xmm0
; X32-NEXT: retl
;
; X64-LABEL: t5:
; X64: # BB#0:
; X64-NEXT: psrlw $8, %xmm0
; X64-NEXT: retq
[x86] Regenerate precise checks for a couple of test cases and remove a test case that was just grepping the debug stats output rather than actually checking the generated code for anything useful. llvm-svn: 218951
2014-10-03 09:50:08 +08:00
%s = shufflevector <16 x i8> %x, <16 x i8> zeroinitializer, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 17>
ret <16 x i8> %s
X86: Make shuffle -> shift conversion more aggressive about undefs. Shuffles that only move an element into position 0 of the vector are common in the output of the loop vectorizer and often generate suboptimal code when SSSE3 is not available. Lower them to vector shifts if possible. We still prefer palignr over psrldq because it has higher throughput on sandybridge. llvm-svn: 182102
2013-05-17 22:48:34 +08:00
}
define <16 x i8> @t6(<16 x i8> %x) nounwind {
[X86][SSE] Regenerated the vec_insert tests. llvm-svn: 265179
2016-04-02 03:42:23 +08:00
; X32-LABEL: t6:
; X32: # BB#0:
; X32-NEXT: psrlw $8, %xmm0
; X32-NEXT: retl
;
; X64-LABEL: t6:
; X64: # BB#0:
; X64-NEXT: psrlw $8, %xmm0
; X64-NEXT: retq
[x86] Regenerate precise checks for a couple of test cases and remove a test case that was just grepping the debug stats output rather than actually checking the generated code for anything useful. llvm-svn: 218951
2014-10-03 09:50:08 +08:00
%s = shufflevector <16 x i8> %x, <16 x i8> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
ret <16 x i8> %s
X86: Make shuffle -> shift conversion more aggressive about undefs. Shuffles that only move an element into position 0 of the vector are common in the output of the loop vectorizer and often generate suboptimal code when SSSE3 is not available. Lower them to vector shifts if possible. We still prefer palignr over psrldq because it has higher throughput on sandybridge. llvm-svn: 182102
2013-05-17 22:48:34 +08:00
}
define <16 x i8> @t7(<16 x i8> %x) nounwind {
[X86][SSE] Regenerated the vec_insert tests. llvm-svn: 265179
2016-04-02 03:42:23 +08:00
; X32-LABEL: t7:
; X32: # BB#0:
; X32-NEXT: pslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2]
; X32-NEXT: retl
;
; X64-LABEL: t7:
; X64: # BB#0:
; X64-NEXT: pslldq {{.*#+}} xmm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,xmm0[0,1,2]
; X64-NEXT: retq
[x86] Regenerate precise checks for a couple of test cases and remove a test case that was just grepping the debug stats output rather than actually checking the generated code for anything useful. llvm-svn: 218951
2014-10-03 09:50:08 +08:00
%s = shufflevector <16 x i8> %x, <16 x i8> undef, <16 x i32> <i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 1, i32 2>
ret <16 x i8> %s
X86: Make shuffle -> shift conversion more aggressive about undefs. Shuffles that only move an element into position 0 of the vector are common in the output of the loop vectorizer and often generate suboptimal code when SSSE3 is not available. Lower them to vector shifts if possible. We still prefer palignr over psrldq because it has higher throughput on sandybridge. llvm-svn: 182102
2013-05-17 22:48:34 +08:00
}
[X86][SSE] psrl(w/d/q) and psll(w/d/q) bit shifts for SSE2 Patch to match cases where shuffle masks can be reduced to bit shifts. Similar to byte shift shuffle matching from D5699. Differential Revision: http://reviews.llvm.org/D6649 llvm-svn: 228047
2015-02-04 05:58:29 +08:00
define <16 x i8> @t8(<16 x i8> %x) nounwind {
[X86][SSE] Regenerated the vec_insert tests. llvm-svn: 265179
2016-04-02 03:42:23 +08:00
; X32-LABEL: t8:
; X32: # BB#0:
; X32-NEXT: psrldq {{.*#+}} xmm0 = xmm0[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],zero
; X32-NEXT: retl
;
; X64-LABEL: t8:
; X64: # BB#0:
; X64-NEXT: psrldq {{.*#+}} xmm0 = xmm0[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],zero
; X64-NEXT: retq
[X86][SSE] psrl(w/d/q) and psll(w/d/q) bit shifts for SSE2 Patch to match cases where shuffle masks can be reduced to bit shifts. Similar to byte shift shuffle matching from D5699. Differential Revision: http://reviews.llvm.org/D6649 llvm-svn: 228047
2015-02-04 05:58:29 +08:00
%s = shufflevector <16 x i8> %x, <16 x i8> zeroinitializer, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 8, i32 9, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 17>
ret <16 x i8> %s
}
define <16 x i8> @t9(<16 x i8> %x) nounwind {
[X86][SSE] Regenerated the vec_insert tests. llvm-svn: 265179
2016-04-02 03:42:23 +08:00
; X32-LABEL: t9:
; X32: # BB#0:
; X32-NEXT: psrldq {{.*#+}} xmm0 = xmm0[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],zero
; X32-NEXT: retl
;
; X64-LABEL: t9:
; X64: # BB#0:
; X64-NEXT: psrldq {{.*#+}} xmm0 = xmm0[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15],zero
; X64-NEXT: retq
[X86][SSE] psrl(w/d/q) and psll(w/d/q) bit shifts for SSE2 Patch to match cases where shuffle masks can be reduced to bit shifts. Similar to byte shift shuffle matching from D5699. Differential Revision: http://reviews.llvm.org/D6649 llvm-svn: 228047
2015-02-04 05:58:29 +08:00
%s = shufflevector <16 x i8> %x, <16 x i8> undef, <16 x i32> <i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 7, i32 8, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 14, i32 undef, i32 undef>
ret <16 x i8> %s
}