llvm-6502/test/CodeGen/X86/negative-sin.ll

; RUN: llc < %s -enable-unsafe-fp-math -march=x86-64 | FileCheck %s
; CHECK-NOT:     {{addsd|subsd|xor}}

declare double @sin(double %f)

define double @foo(double %e)
{
  %f = fsub double 0.0, %e
  %g = call double @sin(double %f) readonly
  %h = fsub double 0.0, %g
  ret double %h
}
test/CodeGen/X86/negative-sin.ll: FileCheck-ize. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@122619 91177308-0d34-0410-b5e6-96231b3b80d8 2010-12-29 03:58:47 +00:00			`; RUN: llc < %s -enable-unsafe-fp-math -march=x86-64 \| FileCheck %s`
			`; CHECK-NOT: {{addsd\|subsd\|xor}}`
New test case. DAGCombiner should be able to fold -sin(-x) in -enable-unsafe-fp-math mode. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37841 91177308-0d34-0410-b5e6-96231b3b80d8 2007-07-02 15:43:20 +00:00
			`declare double @sin(double %f)`

			`define double @foo(double %e)`
			`{`
Split the Add, Sub, and Mul instruction opcodes into separate integer and floating-point opcodes, introducing FAdd, FSub, and FMul. For now, the AsmParser, BitcodeReader, and IRBuilder all preserve backwards compatability, and the Core LLVM APIs preserve backwards compatibility for IR producers. Most front-ends won't need to change immediately. This implements the first step of the plan outlined here: http://nondot.org/sabre/LLVMNotes/IntegerOverflow.txt git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@72897 91177308-0d34-0410-b5e6-96231b3b80d8 2009-06-04 22:49:04 +00:00			`%f = fsub double 0.0, %e`
Add readonly to some sin and cos calls; transformations being checked aren't valid without it. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@82786 91177308-0d34-0410-b5e6-96231b3b80d8 2009-09-25 18:15:29 +00:00			`%g = call double @sin(double %f) readonly`
Split the Add, Sub, and Mul instruction opcodes into separate integer and floating-point opcodes, introducing FAdd, FSub, and FMul. For now, the AsmParser, BitcodeReader, and IRBuilder all preserve backwards compatability, and the Core LLVM APIs preserve backwards compatibility for IR producers. Most front-ends won't need to change immediately. This implements the first step of the plan outlined here: http://nondot.org/sabre/LLVMNotes/IntegerOverflow.txt git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@72897 91177308-0d34-0410-b5e6-96231b3b80d8 2009-06-04 22:49:04 +00:00			`%h = fsub double 0.0, %g`
New test case. DAGCombiner should be able to fold -sin(-x) in -enable-unsafe-fp-math mode. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@37841 91177308-0d34-0410-b5e6-96231b3b80d8 2007-07-02 15:43:20 +00:00			`ret double %h`
			`}`