llvm-6502/test/CodeGen/X86/avx512-cvt.ll
Andrew Trick 6a7770b7ae Enable MI Sched for x86.
This changes the SelectionDAG scheduling preference to source
order. Soon, the SelectionDAG scheduler can be bypassed saving
a nice chunk of compile time.

Performance differences that result from this change are often a
consequence of register coalescing. The register coalescer is far from
perfect. Bugs can be filed for deficiencies.

On x86 SandyBridge/Haswell, the source order schedule is often
preserved, particularly for small blocks.

Register pressure is generally improved over the SD scheduler's ILP
mode. However, we are still able to handle large blocks that require
latency hiding, unlike the SD scheduler's BURR mode. MI scheduler also
attempts to discover the critical path in single-block loops and
adjust heuristics accordingly.

The MI scheduler relies on the new machine model. This is currently
unimplemented for AVX, so we may not be generating the best code yet.

Unit tests are updated so they don't depend on SD scheduling heuristics.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@192750 91177308-0d34-0410-b5e6-96231b3b80d8
2013-10-15 23:33:07 +00:00

218 lines
4.5 KiB
LLVM

; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl | FileCheck %s
; CHECK-LABEL: sitof32
; CHECK: vcvtdq2ps %zmm
; CHECK: ret
define <16 x float> @sitof32(<16 x i32> %a) nounwind {
%b = sitofp <16 x i32> %a to <16 x float>
ret <16 x float> %b
}
; CHECK-LABEL: fptosi00
; CHECK: vcvttps2dq %zmm
; CHECK: ret
define <16 x i32> @fptosi00(<16 x float> %a) nounwind {
%b = fptosi <16 x float> %a to <16 x i32>
ret <16 x i32> %b
}
; CHECK-LABEL: fptoui00
; CHECK: vcvttps2udq
; CHECK: ret
define <16 x i32> @fptoui00(<16 x float> %a) nounwind {
%b = fptoui <16 x float> %a to <16 x i32>
ret <16 x i32> %b
}
; CHECK-LABEL: fptoui01
; CHECK: vcvttpd2udq
; CHECK: ret
define <8 x i32> @fptoui01(<8 x double> %a) nounwind {
%b = fptoui <8 x double> %a to <8 x i32>
ret <8 x i32> %b
}
; CHECK-LABEL: sitof64
; CHECK: vcvtdq2pd %ymm
; CHECK: ret
define <8 x double> @sitof64(<8 x i32> %a) {
%b = sitofp <8 x i32> %a to <8 x double>
ret <8 x double> %b
}
; CHECK-LABEL: fptosi01
; CHECK: vcvttpd2dq %zmm
; CHECK: ret
define <8 x i32> @fptosi01(<8 x double> %a) {
%b = fptosi <8 x double> %a to <8 x i32>
ret <8 x i32> %b
}
; CHECK-LABEL: fptrunc00
; CHECK: vcvtpd2ps %zmm
; CHECK-NEXT: vcvtpd2ps %zmm
; CHECK-NEXT: vinsertf64x4 $1
; CHECK: ret
define <16 x float> @fptrunc00(<16 x double> %b) nounwind {
%a = fptrunc <16 x double> %b to <16 x float>
ret <16 x float> %a
}
; CHECK-LABEL: fpext00
; CHECK: vcvtps2pd %ymm0, %zmm0
; CHECK: ret
define <8 x double> @fpext00(<8 x float> %b) nounwind {
%a = fpext <8 x float> %b to <8 x double>
ret <8 x double> %a
}
; CHECK-LABEL: funcA
; CHECK: vcvtsi2sdqz (%
; CHECK: ret
define double @funcA(i64* nocapture %e) {
entry:
%tmp1 = load i64* %e, align 8
%conv = sitofp i64 %tmp1 to double
ret double %conv
}
; CHECK-LABEL: funcB
; CHECK: vcvtsi2sdlz (%
; CHECK: ret
define double @funcB(i32* %e) {
entry:
%tmp1 = load i32* %e, align 4
%conv = sitofp i32 %tmp1 to double
ret double %conv
}
; CHECK-LABEL: funcC
; CHECK: vcvtsi2sslz (%
; CHECK: ret
define float @funcC(i32* %e) {
entry:
%tmp1 = load i32* %e, align 4
%conv = sitofp i32 %tmp1 to float
ret float %conv
}
; CHECK-LABEL: i64tof32
; CHECK: vcvtsi2ssqz (%
; CHECK: ret
define float @i64tof32(i64* %e) {
entry:
%tmp1 = load i64* %e, align 8
%conv = sitofp i64 %tmp1 to float
ret float %conv
}
; CHECK-LABEL: fpext
; CHECK: vcvtss2sdz
; CHECK: ret
define void @fpext() {
entry:
%f = alloca float, align 4
%d = alloca double, align 8
%tmp = load float* %f, align 4
%conv = fpext float %tmp to double
store double %conv, double* %d, align 8
ret void
}
; CHECK-LABEL: fpround_scalar
; CHECK: vmovsdz
; CHECK: vcvtsd2ssz
; CHECK: vmovssz
; CHECK: ret
define void @fpround_scalar() nounwind uwtable {
entry:
%f = alloca float, align 4
%d = alloca double, align 8
%tmp = load double* %d, align 8
%conv = fptrunc double %tmp to float
store float %conv, float* %f, align 4
ret void
}
; CHECK-LABEL: long_to_double
; CHECK: vmovqz
; CHECK: ret
define double @long_to_double(i64 %x) {
%res = bitcast i64 %x to double
ret double %res
}
; CHECK-LABEL: double_to_long
; CHECK: vmovqz
; CHECK: ret
define i64 @double_to_long(double %x) {
%res = bitcast double %x to i64
ret i64 %res
}
; CHECK-LABEL: int_to_float
; CHECK: vmovdz
; CHECK: ret
define float @int_to_float(i32 %x) {
%res = bitcast i32 %x to float
ret float %res
}
; CHECK-LABEL: float_to_int
; CHECK: vmovdz
; CHECK: ret
define i32 @float_to_int(float %x) {
%res = bitcast float %x to i32
ret i32 %res
}
; CHECK-LABEL: uitof64
; CHECK: vcvtudq2pd
; CHECK: vextracti64x4
; CHECK: vcvtudq2pd
; CHECK: ret
define <16 x double> @uitof64(<16 x i32> %a) nounwind {
%b = uitofp <16 x i32> %a to <16 x double>
ret <16 x double> %b
}
; CHECK-LABEL: uitof32
; CHECK: vcvtudq2ps
; CHECK: ret
define <16 x float> @uitof32(<16 x i32> %a) nounwind {
%b = uitofp <16 x i32> %a to <16 x float>
ret <16 x float> %b
}
; CHECK-LABEL: @fptosi02
; CHECK vcvttss2siz
; CHECK: ret
define i32 @fptosi02(float %a) nounwind {
%b = fptosi float %a to i32
ret i32 %b
}
; CHECK-LABEL: @fptoui02
; CHECK vcvttss2usiz
; CHECK: ret
define i32 @fptoui02(float %a) nounwind {
%b = fptoui float %a to i32
ret i32 %b
}
; CHECK-LABEL: @uitofp02
; CHECK vcvtusi2ss
; CHECK: ret
define float @uitofp02(i32 %a) nounwind {
%b = uitofp i32 %a to float
ret float %b
}
; CHECK-LABEL: @uitofp03
; CHECK vcvtusi2sd
; CHECK: ret
define double @uitofp03(i32 %a) nounwind {
%b = uitofp i32 %a to double
ret double %b
}