mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-23 19:23:23 +01:00
068d715728
And instead just generate a libcall. My motivating example on ARM was a simple: shl i64 %A, %B for which the code bloat is quite significant. For other targets that also accept __int128/i128 such as AArch64 and X86, it is also beneficial for these cases to generate a libcall when optimising for minsize. On these 64-bit targets, the 64-bits shifts are of course unaffected because the SHIFT/SHIFT_PARTS lowering operation action is not set to custom/expand. Differential Revision: https://reviews.llvm.org/D57386 llvm-svn: 352736
135 lines
5.0 KiB
LLVM
135 lines
5.0 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc < %s -mtriple=x86_64-unknown | FileCheck %s
|
|
|
|
define i64 @f0(i64 %val, i64 %amt) minsize optsize {
|
|
; CHECK-LABEL: f0:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: movq %rsi, %rcx
|
|
; CHECK-NEXT: movq %rdi, %rax
|
|
; CHECK-NEXT: # kill: def $cl killed $cl killed $rcx
|
|
; CHECK-NEXT: shlq %cl, %rax
|
|
; CHECK-NEXT: retq
|
|
%res = shl i64 %val, %amt
|
|
ret i64 %res
|
|
}
|
|
|
|
define i32 @f1(i64 %x, i64 %y) minsize optsize {
|
|
; CHECK-LABEL: f1:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: movq %rsi, %rcx
|
|
; CHECK-NEXT: movq %rdi, %rax
|
|
; CHECK-NEXT: # kill: def $cl killed $cl killed $rcx
|
|
; CHECK-NEXT: shlq %cl, %rax
|
|
; CHECK-NEXT: # kill: def $eax killed $eax killed $rax
|
|
; CHECK-NEXT: retq
|
|
%a = shl i64 %x, %y
|
|
%b = trunc i64 %a to i32
|
|
ret i32 %b
|
|
}
|
|
|
|
define i32 @f2(i64 %x, i64 %y) minsize optsize {
|
|
; CHECK-LABEL: f2:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: movq %rsi, %rcx
|
|
; CHECK-NEXT: movq %rdi, %rax
|
|
; CHECK-NEXT: # kill: def $cl killed $cl killed $rcx
|
|
; CHECK-NEXT: sarq %cl, %rax
|
|
; CHECK-NEXT: # kill: def $eax killed $eax killed $rax
|
|
; CHECK-NEXT: retq
|
|
%a = ashr i64 %x, %y
|
|
%b = trunc i64 %a to i32
|
|
ret i32 %b
|
|
}
|
|
|
|
define i32 @f3(i64 %x, i64 %y) minsize optsize {
|
|
; CHECK-LABEL: f3:
|
|
; CHECK: # %bb.0:
|
|
; CHECK-NEXT: movq %rsi, %rcx
|
|
; CHECK-NEXT: movq %rdi, %rax
|
|
; CHECK-NEXT: # kill: def $cl killed $cl killed $rcx
|
|
; CHECK-NEXT: shrq %cl, %rax
|
|
; CHECK-NEXT: # kill: def $eax killed $eax killed $rax
|
|
; CHECK-NEXT: retq
|
|
%a = lshr i64 %x, %y
|
|
%b = trunc i64 %a to i32
|
|
ret i32 %b
|
|
}
|
|
|
|
define dso_local { i64, i64 } @shl128(i64 %x.coerce0, i64 %x.coerce1, i8 signext %y) minsize optsize {
|
|
; CHECK-LABEL: shl128:
|
|
; CHECK: # %bb.0: # %entry
|
|
; CHECK-NEXT: pushq %rax
|
|
; CHECK-NEXT: .cfi_def_cfa_offset 16
|
|
; CHECK-NEXT: movzbl %dl, %edx
|
|
; CHECK-NEXT: callq __ashlti3
|
|
; CHECK-NEXT: popq %rcx
|
|
; CHECK-NEXT: .cfi_def_cfa_offset 8
|
|
; CHECK-NEXT: retq
|
|
entry:
|
|
%x.sroa.2.0.insert.ext = zext i64 %x.coerce1 to i128
|
|
%x.sroa.2.0.insert.shift = shl nuw i128 %x.sroa.2.0.insert.ext, 64
|
|
%x.sroa.0.0.insert.ext = zext i64 %x.coerce0 to i128
|
|
%x.sroa.0.0.insert.insert = or i128 %x.sroa.2.0.insert.shift, %x.sroa.0.0.insert.ext
|
|
%conv = sext i8 %y to i32
|
|
%sh_prom = zext i32 %conv to i128
|
|
%shl = shl i128 %x.sroa.0.0.insert.insert, %sh_prom
|
|
%retval.sroa.0.0.extract.trunc = trunc i128 %shl to i64
|
|
%retval.sroa.2.0.extract.shift = lshr i128 %shl, 64
|
|
%retval.sroa.2.0.extract.trunc = trunc i128 %retval.sroa.2.0.extract.shift to i64
|
|
%.fca.0.insert = insertvalue { i64, i64 } undef, i64 %retval.sroa.0.0.extract.trunc, 0
|
|
%.fca.1.insert = insertvalue { i64, i64 } %.fca.0.insert, i64 %retval.sroa.2.0.extract.trunc, 1
|
|
ret { i64, i64 } %.fca.1.insert
|
|
}
|
|
|
|
define dso_local { i64, i64 } @ashr128(i64 %x.coerce0, i64 %x.coerce1, i8 signext %y) minsize optsize {
|
|
; CHECK-LABEL: ashr128:
|
|
; CHECK: # %bb.0: # %entry
|
|
; CHECK-NEXT: pushq %rax
|
|
; CHECK-NEXT: .cfi_def_cfa_offset 16
|
|
; CHECK-NEXT: callq __ashrti3
|
|
; CHECK-NEXT: popq %rcx
|
|
; CHECK-NEXT: .cfi_def_cfa_offset 8
|
|
; CHECK-NEXT: retq
|
|
entry:
|
|
%x.sroa.2.0.insert.ext = zext i64 %x.coerce1 to i128
|
|
%x.sroa.2.0.insert.shift = shl nuw i128 %x.sroa.2.0.insert.ext, 64
|
|
%x.sroa.0.0.insert.ext = zext i64 %x.coerce0 to i128
|
|
%x.sroa.0.0.insert.insert = or i128 %x.sroa.2.0.insert.shift, %x.sroa.0.0.insert.ext
|
|
%conv = sext i8 %y to i32
|
|
%sh_prom = zext i32 %conv to i128
|
|
%shr = ashr i128 %x.sroa.0.0.insert.insert, %sh_prom
|
|
%retval.sroa.0.0.extract.trunc = trunc i128 %shr to i64
|
|
%retval.sroa.2.0.extract.shift = lshr i128 %shr, 64
|
|
%retval.sroa.2.0.extract.trunc = trunc i128 %retval.sroa.2.0.extract.shift to i64
|
|
%.fca.0.insert = insertvalue { i64, i64 } undef, i64 %retval.sroa.0.0.extract.trunc, 0
|
|
%.fca.1.insert = insertvalue { i64, i64 } %.fca.0.insert, i64 %retval.sroa.2.0.extract.trunc, 1
|
|
ret { i64, i64 } %.fca.1.insert
|
|
}
|
|
|
|
define dso_local { i64, i64 } @lshr128(i64 %x.coerce0, i64 %x.coerce1, i8 signext %y) minsize optsize {
|
|
; CHECK-LABEL: lshr128:
|
|
; CHECK: # %bb.0: # %entry
|
|
; CHECK-NEXT: pushq %rax
|
|
; CHECK-NEXT: .cfi_def_cfa_offset 16
|
|
; CHECK-NEXT: movzbl %dl, %edx
|
|
; CHECK-NEXT: callq __lshrti3
|
|
; CHECK-NEXT: popq %rcx
|
|
; CHECK-NEXT: .cfi_def_cfa_offset 8
|
|
; CHECK-NEXT: retq
|
|
entry:
|
|
%x.sroa.2.0.insert.ext = zext i64 %x.coerce1 to i128
|
|
%x.sroa.2.0.insert.shift = shl nuw i128 %x.sroa.2.0.insert.ext, 64
|
|
%x.sroa.0.0.insert.ext = zext i64 %x.coerce0 to i128
|
|
%x.sroa.0.0.insert.insert = or i128 %x.sroa.2.0.insert.shift, %x.sroa.0.0.insert.ext
|
|
%conv = sext i8 %y to i32
|
|
%sh_prom = zext i32 %conv to i128
|
|
%shr = lshr i128 %x.sroa.0.0.insert.insert, %sh_prom
|
|
%retval.sroa.0.0.extract.trunc = trunc i128 %shr to i64
|
|
%retval.sroa.2.0.extract.shift = lshr i128 %shr, 64
|
|
%retval.sroa.2.0.extract.trunc = trunc i128 %retval.sroa.2.0.extract.shift to i64
|
|
%.fca.0.insert = insertvalue { i64, i64 } undef, i64 %retval.sroa.0.0.extract.trunc, 0
|
|
%.fca.1.insert = insertvalue { i64, i64 } %.fca.0.insert, i64 %retval.sroa.2.0.extract.trunc, 1
|
|
ret { i64, i64 } %.fca.1.insert
|
|
}
|
|
|