Do some MVS/AES boards overclock better than others?

This is something I have been wondering for a little while and was wondering if anyone had experimented with different motherboard models/revisions to see if some systems could handle the overclocked speed better than others.

I see the problem more with the way the games are programmed, some handle the extra speed better and some garble up.

They all run off a 68k so they should be the same unless the SMD ones are better than the DIP ones.

I remember reading a post or two here were someone said it was the ram not being able to keep up with the CPU or something like that which caused the artifacting when overclocked. It made me wonder if maybe a certain set of Neo boards had better or faster components that worked better at the higher speed.

Vectorman0 said:

I remember reading a post or two here were someone said it was the ram not being able to keep up with the CPU or something like that which caused the artifacting when overclocked. It made me wonder if maybe a certain set of Neo boards had better or faster components that worked better at the higher speed.

Click to expand...

The reported GFX glitches come from VRAM not being written properly and the rest of the game works fine (that's the summary of it, right XX?). It goes through 68k->LSPC->VRAM and the LSPC is rather shit in terms of speed. It is already a bit too slow even with a stock 12mhz 68k writing to it. You can probably class overclockability based on what parts make up that part of the system which is the LSPC/68KB of RAM it has attached to it. Maybe faster VRAMs can help mask the LSPC slowness but I don't know the exact fault. Depending on how gets data into VRAM when fed by the 68k it might not do anything at all to improve overclockability.

Everything seems to come down to the actual programming. I guess when SNK made Metal Slug 2 they learned a hard lesson with the programming. All the slugs after MS2 have no difference and no glitches even running at 16mhz.

I don't think all the newer releases were optimized like the slugs were as some releases still had glitches in them.

If one were desperate enough they could replace the P ROM(s) with de-optimized code that accesses VRAM slower, that will eliminate the garbling by giving LSPC enough time to write it to VRAM. Most games don't seem to have much headroom with VRAM access since they access it very quickly unlike the slugs which I have seen use terribly slow code for this. On the other hand not many would want to hack up an otherwise legit game to degrade performance just to get it working properly on a select few systems.