mirror/dolphin
1
0
Fork 0
mirror of https://github.com/dolphin-emu/dolphin.git synced 2024-12-26 09:27:16 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Jit_FloatingPoint: Prefer BLENDVPD over VBLENDVPD

Browse Source 
Prefer BLENDVPD over VBLENDVPD if the latter doesn't save any
instructions.

VBLENDVPD allows separate source and destination registers, which can
eliminate a MOVAPD/MOVSD. However, on Intel since Skylake, VBLENDVPD
takes additional uops to execute compared to BLENDVPD (according to
https://uops.info). On AMD and older Intel microarchitectures there is no
difference.
This commit is contained in:
Sintendo 2024-12-07 10:53:53 +01:00
parent 0a84d93a8e
commit 065165f749
1 changed files with 15 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
Source/Core/Core/PowerPC/Jit64/Jit_FloatingPoint.cpp
View File

@ -644,6 +644,20 @@ void Jit64::fselx(UGeckoInstruction inst)
if (cpu_info.bAVX)
{
// Prefer BLENDVPD over VBLENDVPD if the latter doesn't save any
// instructions.
//
// VBLENDVPD allows separate source and destination registers, which can
// eliminate a MOVAPD/MOVSD. However, on Intel since Skylake, VBLENDVPD
// takes additional uops to execute compared to BLENDVPD (according to
// https://uops.info). On AMD and older Intel microarchitectures there is no
// difference.
if (d == c)
{
BLENDVPD(Rd, Rb);
return;
}
X64Reg src1 = XMM1;
if (Rc.IsSimpleReg())
{
@ -654,7 +668,7 @@ void Jit64::fselx(UGeckoInstruction inst)
MOVAPD(XMM1, Rc);
}
if (d == c || packed)
if (packed)
{
VBLENDVPD(Rd, src1, Rb, XMM0);
return;