Re: [hatari-devel] Re: 040/060 emulation issue with BadMood

Hi Laurent,

Yes it does sound familiar. Hmm.

In fact I recently encountered some weird problems on my 68040 machine as well, which seemed to indicate the host port sync was reporting the wrong result in some circumstances (very tight polling/reading). And this only happens when the code executes from TT-ram.

This is really weird, because all of the tight looping code is fully cached, and runs at the same speed from the cache regardless of TT-ram. Still a chance that some once-only setup code is running into timing issues but there is very little of that ... its very puzzling.

(On my 040 @ 40/20, the TT-ram can be put in copyback-cache mode which prevents immediates writes to the bus - but not by default, and it only affects data not code. And using TT-ram for data causes no issues anyway, which just adds to the puzzle).

For now I'm assuming these are just bugs in my code on faster machines but it is strange to see you are getting similar sorts of freezes in the same MHz range on CT60...

Once I get some time to check out BM on Hatari in 060 mode I'll try for some new tests on HW at faster speeds...

On 2 March 2015 at 16:11, <laurent.sallafranque@xxxxxxx> wrote:

Hi Doug,

Just for your info, while I finished Racer 2, Templeton did many tests for CT60 compatibility (I don't own a CT60 and my target is the stock 4 mo Falcon, but if the game could run also for CT60, why not doing it ;)

So, the results were :
- at fast frquencies, (75 Mhz and more), the game freezed after some time (1 minute or so)
- at lower frequencies (like 66 Mhz), the game freezed, but after 20 minutes to 1 hour of gaming
- at 50 Mhz, he didn't detect any freeze after 1 hour of gaming.

I wrote something about downgrading the CT60 frequency into the README.TXT file of Racer 2.

Maybe this is not related, but it sounds strangely similar to the problem some users encounter with BadMood and fast CT60, no ?

For info, in Racer 2, I only use the DSP for Mod playing + 4 channels sounds mixing.

I suspect a problem with the CT60 drivers, or video drivers, or something like that (that would become unstable at high frequencies when strongly used (like in a game).

Regards

Laurent

----- Mail original -----
De: "Douglas Little" <doug694@xxxxxxxxxxxxxx>
À: "Eero Tamminen" <oak@xxxxxxxxxxxxxx>
Cc: hatari-devel@xxxxxxxxxxxxxxxxxxx
Envoyé: Lundi 2 Mars 2015 12:00:10
Objet: [hatari-devel] Re: 040/060 emulation issue with BadMood

Real users of 060 machines have complained of the game freezing when the CPU speed goes above 75MHz, so its very likely the problem is still with the game, not Hatari. I mean some timing differences must exist in Hatari anyway, but this should be absorbed by the game code since real machines vary also....

I'll test with recent Hatari and see what I can find.

D

On 1 March 2015 at 18:27, Douglas Little < doug694@xxxxxxxxxxxxxx > wrote:

Hi,

Both work with higher CPU clock (32Mhz) and TT-RAM, but freeze at level
start when I try Hatari's 040 or 060 emulation. There's no Hatari output
when that happens (e.g. about bus errors).

It looks like a timing mismatch. What were the exact names of the build archives?

And which build was used to get the disasm? One of the earlier builds should sync on all transactions so timing issues shouldn't happen unless I missed one. The later ones relax the use of sync in some of the inner loops.

Thanks,

Doug.