CPS2 SSF2T Speed Test - Real Arcade Hardware vs MiSTer vs FBNeo vs 30th Anniversary

[Harineto]

Follow up from CPS1 testing I present SSF2T testing. Enjoy!

 

[djsheep]

Thanks for doing the science @Harineto — when I used MiSTer I always felt that the CPS cores were janky. I remember mentioning it to @rtw at the time and he felt something was off too.

Seeing proof like this is great.

 

[thomasbender]

Hi everyone,

This is very interesting!

@Harineto do you have SFZ2 to run this test?

I own that board + a Mister so if anything, I can replicate this setup and contribute.

 

[Harineto]

Hi everyone,

This is very interesting!

@Harineto do you have SFZ2 to run this test?

I own that board + a Mister so if anything, I can replicate this setup and contribute.

I own two CPS2's one is running darksoft multi so I could test SFZ2 if people want to see how it is.

 

[msamiullah]

I love mister but definitely there is some lag (usb) and sound emulation is not the same for sure, that being said, may be its 90% of original hardware may be 95% , still not bad for what it is.

 

[djsheep]

MAME > MiSTer.

 

[msamiullah]

MAME is

 

[thomasbender]

Alright I want to go forward with SFZ2 on my end (I don’t own a multi) and run a similar test.
For science!

• Arcade CPS2
• Mister
• 30th anniversary
• Fightcade FB Neo
• GroovyMAME

The latter two I’m more interested in running in 15 KHz. I’ve ordered an old Radeon for those two.

Any other ports/emulators worthwhile? RetroArch’s FB Neo?

 

[TechnicalMonkey]

Follow up from CPS1 testing I present SSF2T testing. Enjoy!

I know that some JT builds were better than others in the speed category. I'd like to see a shootout of the builds. I heard that one was released recently, so it would be nice to compare that one with much older builds of the core, and just get the best one to compare to CPSx hardware.

 

[Harineto]

I know that some JT builds were better than others in the speed category. I'd like to see a shootout of the builds. I heard that one was released recently, so it would be nice to compare that one with much older builds of the core, and just get the best one to compare to CPSx hardware.

I tried a few of the recent builds going back a few months they all have speed issues. It's so hard to pinpoint a build that has better speed. I'll do some research for sure.

 

[TechnicalMonkey]

I tried a few of the recent builds going back a few months they all have speed issues. It's so hard to pinpoint a build that has better speed. I'll do some research for sure.

look up fluxcore on yt, and you will get an idea of what was used in his comparisons.

also a repository of older rbf files.
https://github.com/sofia-rose/old_jtbin/tree/master/mister

 

[thomasbender]

https://x.com/topapate/status/2004551158283862213?s=46&t=jj2-3Y_L_1gCriUgZ7-WDQ

 

[TechnicalMonkey]

https://x.com/topapate/status/2004551158283862213?s=46&t=jj2-3Y_L_1gCriUgZ7-WDQ

When is the next scheduled Binary?

 

[terminator2k2]

if you dont want to wait for JT's official release you can try his nightly build here:

https://github.com/jotego/jtcores/actions/runs/20533678482/artifacts/4973940675

 

[ShootTheCore]

 

[Harineto]

Once the build is out officially I will rerun my tests.

 

[djsheep]

Once the build is out officially I will rerun my tests.

I reckon they’ll still show inconsistencies.

Given how this Jotego clown defends his sloppy cores, there’s no way Mr. e-beggar is actually doing his due diligence and actually testing his work thoroughly.

Keep the videos coming man. Nice to see the whole “cycle accurate” myth be debunked

 

[Harineto]

Updated test:

View: https://www.youtube.com/watch?v=fR3D0fbRo2w

 

[djsheep]

 

[nem]

Huge improvement. MiSTer now performs better than the other two emulation setups.

As someone who only dabbles in original hardware, I wonder how long would it take for two "identical" CPS2 boards to fall out of sync? Hours or days?