Re: [hatari-devel] Emulation on separate CPU cores

[Eero Tamminen <oak@xxxxxxxxxxxxxx>]

Hi,

I watched the guy's video of the emulator, and I think most interesting 
thing about it is it being 100% deterministic, which allows:
* Doing movie recording that records just the user inputs, instead of 
video output
* When replaying that movie, user continuing playing from *any* place in 
that "movie"


	- Eero

On 4.12.2022 15.44, Eero Tamminen wrote:
> Hi,
>
> On 4.12.2022 15.20, Thorsten Otto wrote:
> > On Sonntag, 4. Dezember 2022 11:55:34 CET Miro Kropáček wrote:
> >
> > > That guy is really good, so I wouldn't dismiss his achievement simply as
> > > "nah, surely his emulator isn't that accurate and that's why it's so 
> > > fast".
> >
> > I agree that he surely did a great job. If something in my previous 
> > post could
> > be interpreted differently, i apologize.
> >
> > But his goals are different than what Hatari does. Its more like what 
> > Aranym
> > does, running supported programs at maximum achievable speed.
> >
> > Having that said, it also makes me wonder why he used Musashi. It is 
> > not a bad
> > emulator, and in projects like his it is easy to integrate, but it is 
> > surely
> > not the fastest emulator around.
>
> What other m68k emulation cores there are?
>
> Qemu - still too inaccurate
>
> OldUAE - obsolete / too inaccurate
>
> WinUAE - complex, emulates lot of different 680x0 variants with 
> different emulation accuracy options
>
> Any others?
>
>
> Maybe threading was easy to integrate to Musashi?
>
> If it is e.g. less complex and has less global state...
>
>
>      - Eero