MAMEWorld Forums - News - MAME 0.281 Icons / Snaps / Benchmarks

Game Speed Seconds
------------------------------------------------
blitz 542.72 89
crusnusa 513.33 89
cubeqst 485.10 89
cyvern 2,426.76 89
dkong 5,292.75 89
drivedge 622.90 89
gauntleg 840.46 89
gnbarich 3,801.61 89
gradius4 788.26 89
gtfore06 1,222.37 89
harddriv 1,137.90 89
kidniki 868.15 89
kof98 2,257.63 89
mario 2,208.34 89
mk4 473.96 89
pacman 23,672.03 89
pinkswts 3,022.75 89
pong 897.71 89
propcycl 501.32 89
radikalb 494.29 89
roadblst 2,047.67 89
robotron 9,462.19 89
rvschool 619.64 89
scud 148.81 89
sf2 4,096.25 89
sfa2 3,197.52 89
sfiii 3,161.08 89
slrasslt 905.07 89
spacfury 380.15 89
starblad 339.44 89
starsldr 212.67 89
sushibar 148.73 89
tekken 988.18 89
tekken3 561.25 89
vf2 286.65 89
vfkids 274.10 89

> The ARM back-end is already executing more IR instructions per host CPU clock cycle
> than the x86-64 back-end, and I’ve mostly just been going for easy gains so far.
> (That’s part of the reason some of the MIPS, PowerPC and SuperH games do so well on
> ARM – the back-end is actually pretty impressive, even to me as the person who did
> most of the performance tuning for it.)
>
> I know I can get quite a bit more performance out of ARM, but it requires some
> significant architectural changes to the back-end (won’t require changes to the
> emulated CPU cores themselves to benefit). The stuff I want to do can’t be done
> incrementally, and it’s going to require non-trivial time investment.
>
> I optimised the Hyperstone E1 recompiler a lot already. Between that and optimising
> the x86-64 back-end, it now generates less than half the number of native
> instructions, while emulating more CPU features with more accuracy and matching the
> interpreter on cycle counts. There are definitely gains to be had by optimising the
> code generation for the other recompiling CPUs, but it isn’t a rabbit hole I want to
> go down right now.
>
> On the bright side, the state the code was in at the beginning of the year meant it
> was relatively easy to get some significant performance gains already while also
> fixing a bunch of long-standing issues along the way. The trouble is I’ll inevitably
> run out of low-hanging fruit and have to start making bigger architectural changes to
> keep making progress.

Thank you for your amazing work.

I have rough understand of how a DRC works and can only imagine the time and skill it takes to make one work so well.

I can understand you taking slow but steady approach, rather than jumping in headfirst and risking burning out.

I'm happy to see performance gains on any front, and if ARM is the way forward MAME performance, I might have to buy my first Mac since the LC 475, though Qualcomm are making a big song and dance about the new Snapdragon X2 Elite Extreme so maybe there will be a second option for a fast ARM based systems soon.

I really hope you keep going with the great work you have started, and I look forward to future developments.