mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-10-22 20:43:44 +02:00
80150ed05d
As discussed in D24815, let's start the process of killing off the broken fast-math global state housed in TargetOptions and eliminate the need for function-level fast-math attributes. Here we enable two similar folds that are possible when we don't care about signed-zero: fadd nsz x, 0 --> x fsub nsz 0, x --> -x Note that although the test cases include a 'sin' function call, I'm side-stepping the FMF-on-calls question (and lack of support in the DAG) for now. It's not needed for these tests - isNegatibleForFree/GetNegatedExpression just look through a ISD::FSIN node. Also, when we create an FNEG node and propagate the Flags of the FSUB to it, this doesn't actually do anything today because Flags are silently dropped for any node that is not a binary operator. Differential Revision: https://reviews.llvm.org/D25297 llvm-svn: 284824
106 lines
2.7 KiB
LLVM
106 lines
2.7 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=avx | FileCheck %s
|
|
|
|
declare double @sin(double %f)
|
|
|
|
; When the subs are strict, they can't be removed because of signed zero.
|
|
|
|
define double @strict(double %e) nounwind {
|
|
; CHECK-LABEL: strict:
|
|
; CHECK: # BB#0:
|
|
; CHECK-NEXT: pushq %rax
|
|
; CHECK-NEXT: vxorpd %xmm1, %xmm1, %xmm1
|
|
; CHECK-NEXT: vsubsd %xmm0, %xmm1, %xmm0
|
|
; CHECK-NEXT: callq sin
|
|
; CHECK-NEXT: vxorpd %xmm1, %xmm1, %xmm1
|
|
; CHECK-NEXT: vsubsd %xmm0, %xmm1, %xmm0
|
|
; CHECK-NEXT: popq %rax
|
|
; CHECK-NEXT: retq
|
|
;
|
|
%f = fsub double 0.0, %e
|
|
%g = call double @sin(double %f) readonly
|
|
%h = fsub double 0.0, %g
|
|
ret double %h
|
|
}
|
|
|
|
; 'fast' implies no-signed-zeros, so the negates fold away.
|
|
; The 'sin' does not need any fast-math-flags for this transform.
|
|
|
|
define double @fast(double %e) nounwind {
|
|
; CHECK-LABEL: fast:
|
|
; CHECK: # BB#0:
|
|
; CHECK-NEXT: jmp sin
|
|
;
|
|
%f = fsub fast double 0.0, %e
|
|
%g = call double @sin(double %f) readonly
|
|
%h = fsub fast double 0.0, %g
|
|
ret double %h
|
|
}
|
|
|
|
; No-signed-zeros is all that we need for this transform.
|
|
|
|
define double @nsz(double %e) nounwind {
|
|
; CHECK-LABEL: nsz:
|
|
; CHECK: # BB#0:
|
|
; CHECK-NEXT: jmp sin
|
|
;
|
|
%f = fsub nsz double 0.0, %e
|
|
%g = call double @sin(double %f) readonly
|
|
%h = fsub nsz double 0.0, %g
|
|
ret double %h
|
|
}
|
|
|
|
; The 1st negate is strict, so we can't kill that sub, but the 2nd disappears.
|
|
|
|
define double @semi_strict1(double %e) nounwind {
|
|
; CHECK-LABEL: semi_strict1:
|
|
; CHECK: # BB#0:
|
|
; CHECK-NEXT: pushq %rax
|
|
; CHECK-NEXT: vxorpd %xmm1, %xmm1, %xmm1
|
|
; CHECK-NEXT: vsubsd %xmm0, %xmm1, %xmm0
|
|
; CHECK-NEXT: callq sin
|
|
; CHECK-NEXT: vxorpd {{.*}}(%rip), %xmm0, %xmm0
|
|
; CHECK-NEXT: popq %rax
|
|
; CHECK-NEXT: retq
|
|
;
|
|
%f = fsub double 0.0, %e
|
|
%g = call double @sin(double %f) readonly
|
|
%h = fsub nsz double 0.0, %g
|
|
ret double %h
|
|
}
|
|
|
|
; The 2nd negate is strict, so we can't kill it. It becomes an add of zero instead.
|
|
|
|
define double @semi_strict2(double %e) nounwind {
|
|
; CHECK-LABEL: semi_strict2:
|
|
; CHECK: # BB#0:
|
|
; CHECK-NEXT: pushq %rax
|
|
; CHECK-NEXT: callq sin
|
|
; CHECK-NEXT: vxorpd %xmm1, %xmm1, %xmm1
|
|
; CHECK-NEXT: vaddsd %xmm1, %xmm0, %xmm0
|
|
; CHECK-NEXT: popq %rax
|
|
; CHECK-NEXT: retq
|
|
;
|
|
%f = fsub nsz double 0.0, %e
|
|
%g = call double @sin(double %f) readonly
|
|
%h = fsub double 0.0, %g
|
|
ret double %h
|
|
}
|
|
|
|
; FIXME:
|
|
; Auto-upgrade function attribute to IR-level fast-math-flags.
|
|
|
|
define double @fn_attr(double %e) nounwind #0 {
|
|
; CHECK-LABEL: fn_attr:
|
|
; CHECK: # BB#0:
|
|
; CHECK-NEXT: jmp sin
|
|
;
|
|
%f = fsub double 0.0, %e
|
|
%g = call double @sin(double %f) readonly
|
|
%h = fsub double 0.0, %g
|
|
ret double %h
|
|
}
|
|
|
|
attributes #0 = { "unsafe-fp-math"="true" }
|
|
|