exxos's Atari, Amiga & retro forum

Dan1975 wrote: 23 Jun 2020 08:41 :lol: You've done a lot better than me then :lol:

:lol:

It would be interesting if you could time how long a intro runs in something like frontier elite. Like time how long it takes to get back to the start in 8MHz and 32MHz modes. This way we would finally have a some clue how much faster 3D games run in 32MHz mode.

No problem. Will try that later if get a chance.

I filmed the start of Damocles in both 8mhz and 32mhz and there is a clear noticeable difference when you watch them back to back.

Dan1975 wrote: 23 Jun 2020 15:22 No problem. Will try that later if get a chance.

I filmed the start of Damocles in both 8mhz and 32mhz and there is a clear noticeable difference when you watch them back to back.

:thumbup:
Yeah, I saw a good gain in castle master. The general idea is about 30% speed boost, but always seems like more in the actual game :shrug:

exxos wrote: 23 Jun 2020 14:52

Dan1975 wrote: 23 Jun 2020 08:41 :lol: You've done a lot better than me then :lol:

:lol:

It would be interesting if you could time how long a intro runs in something like frontier elite. Like time how long it takes to get back to the start in 8MHz and 32MHz modes. This way we would finally have a some clue how much faster 3D games run in 32MHz mode.

FYI, a single loop of all GB tests takes about half the time at 32Mhz, 3 minutes instead of 6.

rubber_jonnie wrote: 23 Jun 2020 17:00 FYI, a single loop of all GB tests takes about half the time at 32Mhz, 3 minutes instead of 6.

Yep, but games are different because RAM is still 100%.. and games use RAM more than anything, so the speed is always biased towards RAM speed.

exxos wrote: 23 Jun 2020 17:05

rubber_jonnie wrote: 23 Jun 2020 17:00 FYI, a single loop of all GB tests takes about half the time at 32Mhz, 3 minutes instead of 6.

Yep, but games are different because RAM is still 100%.. and games use RAM more than anything, so the speed is always biased towards RAM speed.

This seems to be the case with Frontier. At both speeds the intro takes exactly the same time from the beginning until it starts to loop again - approx 3 mins 36 seconds. However it is noticeably smoother and appears faster at 32mhz

Dan1975 wrote: 23 Jun 2020 17:50 This seems to be the case with Frontier. At both speeds the intro takes exactly the same time from the beginning until it starts to loop again - approx 3 mins 36 seconds. However it is noticeably smoother and appears faster at 32mhz

Ah dammit, yeah I bet the frame rate goes up but nothing else. So not sure how it could be tested now :roll:

Try Driller, the faster the CPU, the faster the playback.

Will keep trying different games and reporting back.

Can't do anymore this evening as the wife, who was supposed to be working until 6 and then going for a run so shouldn't have been home until about 7, got in just after 5 as they let her finish early and she said its too hot to go running :sigh:

I wonder if EMUTOS could count the frames drawn in games ? @thorsten.otto ? I half of the feeling it had a frame rate display in it but I could have imagined it.. But if there was some way to actually count the number of frames in something like 60 seconds, while running frontier or something, then there would be one possible method of working out how fast accelerators actually make games.