RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-01 00:12:50 +01:00
Code Issues Projects Releases Wiki Activity
a09775b875
llvm-mirror/test/CodeGen/X86/2007-09-27-LDIntrinsics.ll

31 lines
771 B
LLVM
Raw Normal View History

Teach codegen to lower llvm.powi to an efficient (but not optimal) multiply sequence when the power is a constant integer. Before, our codegen for std::pow(.., int) always turned into a libcall, which was really inefficient. This should also make many gfortran programs happier I'd imagine. llvm-svn: 92388
2010-01-01 04:32:16 +01:00
; RUN: llc < %s | FileCheck %s
Add sqrt and powi intrinsics for long double. llvm-svn: 42423
2007-09-28 03:08:20 +02:00
target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128"
target triple = "i686-apple-darwin8"
Teach codegen to lower llvm.powi to an efficient (but not optimal) multiply sequence when the power is a constant integer. Before, our codegen for std::pow(.., int) always turned into a libcall, which was really inefficient. This should also make many gfortran programs happier I'd imagine. llvm-svn: 92388
2010-01-01 04:32:16 +01:00
define x86_fp80 @foo(x86_fp80 %x) nounwind{
Add sqrt and powi intrinsics for long double. llvm-svn: 42423
2007-09-28 03:08:20 +02:00
entry:
Teach codegen to lower llvm.powi to an efficient (but not optimal) multiply sequence when the power is a constant integer. Before, our codegen for std::pow(.., int) always turned into a libcall, which was really inefficient. This should also make many gfortran programs happier I'd imagine. llvm-svn: 92388
2010-01-01 04:32:16 +01:00
%tmp2 = call x86_fp80 @llvm.sqrt.f80( x86_fp80 %x )
ret x86_fp80 %tmp2
; CHECK: foo:
; CHECK: fldt 4(%esp)
; CHECK-NEXT: fsqrt
; CHECK-NEXT: ret
Add sqrt and powi intrinsics for long double. llvm-svn: 42423
2007-09-28 03:08:20 +02:00
}
declare x86_fp80 @llvm.sqrt.f80(x86_fp80)
Teach codegen to lower llvm.powi to an efficient (but not optimal) multiply sequence when the power is a constant integer. Before, our codegen for std::pow(.., int) always turned into a libcall, which was really inefficient. This should also make many gfortran programs happier I'd imagine. llvm-svn: 92388
2010-01-01 04:32:16 +01:00
define x86_fp80 @bar(x86_fp80 %x) nounwind {
Add sqrt and powi intrinsics for long double. llvm-svn: 42423
2007-09-28 03:08:20 +02:00
entry:
Teach codegen to lower llvm.powi to an efficient (but not optimal) multiply sequence when the power is a constant integer. Before, our codegen for std::pow(.., int) always turned into a libcall, which was really inefficient. This should also make many gfortran programs happier I'd imagine. llvm-svn: 92388
2010-01-01 04:32:16 +01:00
%tmp2 = call x86_fp80 @llvm.powi.f80( x86_fp80 %x, i32 3 )
ret x86_fp80 %tmp2
; CHECK: bar:
; CHECK: fldt 4(%esp)
; CHECK-NEXT: fld %st(0)
Heuristic: If the number of operands in the alias are more than the number of operands in the aliasee, don't print the alias. llvm-svn: 132963
2011-06-14 05:17:20 +02:00
; CHECK-NEXT: fmul %st(1)
Have the X86 back-end emit the alias instead of what's being aliased. In most cases, it's much nicer and more informative reading the alias. llvm-svn: 129497
2011-04-14 03:11:51 +02:00
; CHECK-NEXT: fmulp
Teach codegen to lower llvm.powi to an efficient (but not optimal) multiply sequence when the power is a constant integer. Before, our codegen for std::pow(.., int) always turned into a libcall, which was really inefficient. This should also make many gfortran programs happier I'd imagine. llvm-svn: 92388
2010-01-01 04:32:16 +01:00
; CHECK-NEXT: ret
Add sqrt and powi intrinsics for long double. llvm-svn: 42423
2007-09-28 03:08:20 +02:00
}
declare x86_fp80 @llvm.powi.f80(x86_fp80, i32)
Reference in New Issue Copy Permalink