Re: [hatari-devel] Hatari speed accuracy

[Laurent Sallafranque <laurent.sallafranque@xxxxxxx>]

> The biggest challenge will be to get the same level with falcon 
emulation :)

I agree ;)
That's a huge challenge.

Laurent




Le 15/04/2016 22:37, Nicolas Pomarède a écrit :
> Le 15/04/2016 20:52, Roger Burrows a écrit :
> > I've been doing some performance tweaks in EmuTOS recently, and it is 
> > much more
> > convenient to test them under Hatari than on my real ST. Occasionally 
> > I've had
> > results that surprised me, so I have compared the timings running 
> > under Hatari
> > and on a real ST.  Timings were based on clock ticks.  In all the 
> > cases I
> > checked, Hatari timings were within 1% of ST timings, and the 
> > variability in ST
> > timings was around 0.5% anyway.
> >
> > So congratulations to everyone, a very fine job!
> >
> > Roger
>
> Hi
>
> it depends on what method you used, and against which type of ST you 
> compared (pal, ntsc, there can be some variations). But in fact, to 
> run all the overscan demos and similar tricks, I would say that 
> current accuracy should even be nearly 100% compared to a real ST.
>
> The biggest challenge will be to get the same level with falcon 
> emulation :)