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/pr15267.ll

147 lines
4.9 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
Fix PR15267 - When extloading from a vector with non-byte-addressable element, e.g. <4 x i1>, the current logic breaks. Extend the current logic to fix the case where the element type is not byte-addressable by loading all bytes, bit-extracting/packing each element. llvm-svn: 175642
2013-02-21 02:04:21 +08:00
; RUN: llc < %s -mtriple=x86_64-pc-linux -mcpu=corei7-avx | FileCheck %s
define <4 x i3> @test1(<4 x i3>* %in) nounwind {
[X86] Refreshed sign extension tests. llvm-svn: 245358
2015-08-19 05:21:35 +08:00
; CHECK-LABEL: test1:
; CHECK: # BB#0:
; CHECK-NEXT: movzwl (%rdi), %eax
; CHECK-NEXT: movl %eax, %ecx
; CHECK-NEXT: shrl $3, %ecx
; CHECK-NEXT: andl $7, %ecx
; CHECK-NEXT: movl %eax, %edx
; CHECK-NEXT: andl $7, %edx
; CHECK-NEXT: vmovd %edx, %xmm0
; CHECK-NEXT: vpinsrd $1, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movl %eax, %ecx
; CHECK-NEXT: shrl $6, %ecx
; CHECK-NEXT: andl $7, %ecx
; CHECK-NEXT: vpinsrd $2, %ecx, %xmm0, %xmm0
; CHECK-NEXT: shrl $9, %eax
; CHECK-NEXT: andl $7, %eax
; CHECK-NEXT: vpinsrd $3, %eax, %xmm0, %xmm0
; CHECK-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
%ret = load <4 x i3>, <4 x i3>* %in, align 1
Fix PR15267 - When extloading from a vector with non-byte-addressable element, e.g. <4 x i1>, the current logic breaks. Extend the current logic to fix the case where the element type is not byte-addressable by loading all bytes, bit-extracting/packing each element. llvm-svn: 175642
2013-02-21 02:04:21 +08:00
ret <4 x i3> %ret
}
define <4 x i1> @test2(<4 x i1>* %in) nounwind {
[X86] Refreshed sign extension tests. llvm-svn: 245358
2015-08-19 05:21:35 +08:00
; CHECK-LABEL: test2:
; CHECK: # BB#0:
; CHECK-NEXT: movzbl (%rdi), %eax
; CHECK-NEXT: movl %eax, %ecx
; CHECK-NEXT: shrl %ecx
; CHECK-NEXT: andl $1, %ecx
; CHECK-NEXT: movl %eax, %edx
; CHECK-NEXT: andl $1, %edx
; CHECK-NEXT: vmovd %edx, %xmm0
; CHECK-NEXT: vpinsrd $1, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movl %eax, %ecx
; CHECK-NEXT: shrl $2, %ecx
; CHECK-NEXT: andl $1, %ecx
; CHECK-NEXT: vpinsrd $2, %ecx, %xmm0, %xmm0
; CHECK-NEXT: shrl $3, %eax
; CHECK-NEXT: andl $1, %eax
; CHECK-NEXT: vpinsrd $3, %eax, %xmm0, %xmm0
; CHECK-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
%ret = load <4 x i1>, <4 x i1>* %in, align 1
Fix PR15267 - When extloading from a vector with non-byte-addressable element, e.g. <4 x i1>, the current logic breaks. Extend the current logic to fix the case where the element type is not byte-addressable by loading all bytes, bit-extracting/packing each element. llvm-svn: 175642
2013-02-21 02:04:21 +08:00
ret <4 x i1> %ret
}
define <4 x i64> @test3(<4 x i1>* %in) nounwind {
[X86] Refreshed sign extension tests. llvm-svn: 245358
2015-08-19 05:21:35 +08:00
; CHECK-LABEL: test3:
; CHECK: # BB#0:
; CHECK-NEXT: movzbl (%rdi), %eax
; CHECK-NEXT: movq %rax, %rcx
; CHECK-NEXT: shlq $62, %rcx
; CHECK-NEXT: sarq $63, %rcx
; CHECK-NEXT: movq %rax, %rdx
; CHECK-NEXT: shlq $63, %rdx
; CHECK-NEXT: sarq $63, %rdx
; CHECK-NEXT: vmovd %edx, %xmm0
; CHECK-NEXT: vpinsrd $1, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movq %rax, %rcx
; CHECK-NEXT: shlq $61, %rcx
; CHECK-NEXT: sarq $63, %rcx
; CHECK-NEXT: vpinsrd $2, %ecx, %xmm0, %xmm0
; CHECK-NEXT: shlq $60, %rax
; CHECK-NEXT: sarq $63, %rax
; CHECK-NEXT: vpinsrd $3, %eax, %xmm0, %xmm0
[DAGCombiner] Fold CONCAT_VECTORS of EXTRACT_SUBVECTOR (or undef) to VECTOR_SHUFFLE. Check to see if this is a CONCAT_VECTORS of a bunch of EXTRACT_SUBVECTOR operations. If so, and if the EXTRACT_SUBVECTOR vector inputs come from at most two distinct vectors the same size as the result, attempt to turn this into a legal shuffle. Differential Revision: http://reviews.llvm.org/D12125 llvm-svn: 245490
2015-08-20 04:09:50 +08:00
; CHECK-NEXT: vpmovsxdq %xmm0, %xmm1
; CHECK-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[2,3,0,1]
[X86] Refreshed sign extension tests. llvm-svn: 245358
2015-08-19 05:21:35 +08:00
; CHECK-NEXT: vpmovsxdq %xmm0, %xmm0
[DAGCombiner] Fold CONCAT_VECTORS of EXTRACT_SUBVECTOR (or undef) to VECTOR_SHUFFLE. Check to see if this is a CONCAT_VECTORS of a bunch of EXTRACT_SUBVECTOR operations. If so, and if the EXTRACT_SUBVECTOR vector inputs come from at most two distinct vectors the same size as the result, attempt to turn this into a legal shuffle. Differential Revision: http://reviews.llvm.org/D12125 llvm-svn: 245490
2015-08-20 04:09:50 +08:00
; CHECK-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
[X86] Refreshed sign extension tests. llvm-svn: 245358
2015-08-19 05:21:35 +08:00
; CHECK-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
%wide.load35 = load <4 x i1>, <4 x i1>* %in, align 1
Fix PR15267 - When extloading from a vector with non-byte-addressable element, e.g. <4 x i1>, the current logic breaks. Extend the current logic to fix the case where the element type is not byte-addressable by loading all bytes, bit-extracting/packing each element. llvm-svn: 175642
2013-02-21 02:04:21 +08:00
%sext = sext <4 x i1> %wide.load35 to <4 x i64>
ret <4 x i64> %sext
}
Fixes a bug in vector load legalization that confused bits and bytes. Differential Revision: http://reviews.llvm.org/D7400 llvm-svn: 228168
2015-02-05 02:54:01 +08:00
define <16 x i4> @test4(<16 x i4>* %in) nounwind {
[X86] Refreshed sign extension tests. llvm-svn: 245358
2015-08-19 05:21:35 +08:00
; CHECK-LABEL: test4:
; CHECK: # BB#0:
; CHECK-NEXT: movq (%rdi), %rax
; CHECK-NEXT: movl %eax, %ecx
; CHECK-NEXT: shrl $4, %ecx
; CHECK-NEXT: andl $15, %ecx
; CHECK-NEXT: movl %eax, %edx
; CHECK-NEXT: andl $15, %edx
; CHECK-NEXT: vmovd %edx, %xmm0
; CHECK-NEXT: vpinsrb $1, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movl %eax, %ecx
; CHECK-NEXT: shrl $8, %ecx
; CHECK-NEXT: andl $15, %ecx
; CHECK-NEXT: vpinsrb $2, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movl %eax, %ecx
; CHECK-NEXT: shrl $12, %ecx
; CHECK-NEXT: andl $15, %ecx
; CHECK-NEXT: vpinsrb $3, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movl %eax, %ecx
; CHECK-NEXT: shrl $16, %ecx
; CHECK-NEXT: andl $15, %ecx
; CHECK-NEXT: vpinsrb $4, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movl %eax, %ecx
; CHECK-NEXT: shrl $20, %ecx
; CHECK-NEXT: andl $15, %ecx
; CHECK-NEXT: vpinsrb $5, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movl %eax, %ecx
; CHECK-NEXT: shrl $24, %ecx
; CHECK-NEXT: andl $15, %ecx
; CHECK-NEXT: vpinsrb $6, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movl %eax, %ecx
; CHECK-NEXT: shrl $28, %ecx
; CHECK-NEXT: vpinsrb $7, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movq %rax, %rcx
; CHECK-NEXT: shrq $32, %rcx
; CHECK-NEXT: andl $15, %ecx
; CHECK-NEXT: vpinsrb $8, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movq %rax, %rcx
; CHECK-NEXT: shrq $36, %rcx
; CHECK-NEXT: andl $15, %ecx
; CHECK-NEXT: vpinsrb $9, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movq %rax, %rcx
; CHECK-NEXT: shrq $40, %rcx
; CHECK-NEXT: andl $15, %ecx
; CHECK-NEXT: vpinsrb $10, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movq %rax, %rcx
; CHECK-NEXT: shrq $44, %rcx
; CHECK-NEXT: andl $15, %ecx
; CHECK-NEXT: vpinsrb $11, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movq %rax, %rcx
; CHECK-NEXT: shrq $48, %rcx
; CHECK-NEXT: andl $15, %ecx
; CHECK-NEXT: vpinsrb $12, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movq %rax, %rcx
; CHECK-NEXT: shrq $52, %rcx
; CHECK-NEXT: andl $15, %ecx
; CHECK-NEXT: vpinsrb $13, %ecx, %xmm0, %xmm0
; CHECK-NEXT: movq %rax, %rcx
; CHECK-NEXT: shrq $56, %rcx
; CHECK-NEXT: andl $15, %ecx
; CHECK-NEXT: vpinsrb $14, %ecx, %xmm0, %xmm0
; CHECK-NEXT: shrq $60, %rax
; CHECK-NEXT: vpinsrb $15, %eax, %xmm0, %xmm0
; CHECK-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
%ret = load <16 x i4>, <16 x i4>* %in, align 1
Fixes a bug in vector load legalization that confused bits and bytes. Differential Revision: http://reviews.llvm.org/D7400 llvm-svn: 228168
2015-02-05 02:54:01 +08:00
ret <16 x i4> %ret
}