Overclocking the EE (discussion) (testers wanted)

[Blyss Sarania]

Edit: I added a poll about this, so please vote.

A lot of people seem to have the idea that overclocking the EE will allow games to run at full speed on hardware where it otherwise wouldn't - this is not the case. In fact it is just the opposite - if you can't keep the game at full speed without overclock, it will be even slower with it. What it can do is increase INTERNAL FPS and make gameplay smoother overall, under the right circumstances.

---

Update: Latest tester is here!

---

Also, thanks to Saiki for pointing me in the right direction when I started working on this. He saved me a good bit of work!

---

First off, I put this here instead of developer discussion so it would get more eyes.

Okay so I noticed(and posted in) this issue on Github about overclocking the EE: https://github.com/PCSX2/pcsx2/issues/153

I have only very briefly mentioned it around here(due to not wanting to be bombarded with requests for it), but half a year back I worked on doing just that. I modified the EE slider to go both ways - to under or overclock the EE on the same slider. I also, of course, modified the scalarLow/scalarMid/scalarHigh variables accordingly. I had refraction look over my code, and he said it should work for the most part, but that my highest setting(most overclock) might cause issues because of not enough delay on the blocks.

Anyway, I had Fezzer test it out on his powerful rig. We discovered a few things.

• It increases system requirements IMMENSELY. +50% will cause slowdown even on uber hardware. +33% is doable even on my rig though, for some games
  
• It doesn't help many games. Basically for it to help them they need to either be frame uncapped or be running below their frame cap, and they need to be EE bound. This is not many, but a few.
  
• It doesn't seem to affect compatibility much at all. He tested many games and only found one that broke with it.
  

Anyway since I saw the issue on Github, I thought I'd start a discussion here. What are you guys thoughts on this? Is it worth adding to the main branch? If so, I can submit a pull request for my changes(but the final decision will be with the main coders of course.) But remember, it doesn't really help all that much. So is it a useful or useless feature?

If anyone would like a build with it to test then let me know. I still have my source, but its based on code from about 6 months ago so I need to port it over to the latest GIT and then build.

Please do test if you ask for a build - and remember this is technically an unofficial build. Though I am a mod, I am not a PCSX2 coder. I can tinker. Therefore, do not ask for support on the build if you get it.

Links will only be given out in PM and are not to be shared. Remember, this is currently just a test. I can imagine the feeding frenzy that might occur if the build got out and people thought it would make their games run super fluid on old hardware when in reality that is not the case. That's not how overclocking the EE works. If you can't keep a game at full speed normally then the OC will make it worse. What it can do in some cases is increase the INTERNAL FPS(i.e. not the one in the title bar) and make things more fluid.

So, anyway. If you want to test it for me then post here and I'll PM you a link to the build. Please test as many games as possible and note:

• The name of the game
  
• Your hardware
  
• Any non default settings you used
  
• Whether it works as normal or breaks
  
• Whether you used 33% or 50% OC or both
  
• Whether there is any subjective improvement in internal FPS
  

Lastly, here is a video showing it in action. I used an SNES emulator for PS2 running inside PCSX2. The reason for that is it has a frame counter that I can see changes in. You will see at first(around 1:02) in the area it is 42 FPS, but after ramping up the EE, it goes to 60.

[Nobbs66]

I personally don't feel it's needed. I think it will only cause new users to be confused about the 2-way slider. Sure it may make a few games a bit smoother, but I don't feel it's enough to make adding it worthwhile.

[Mkilbride]

I want this.

Project64 has this. Or 1964, I can't remember. You can OC the internal CPU to up to 6x. In games that had slow downs because of the actual consoles hardware, ran full speed / smoother. It does increase the load on the system, I know that.

I still want it.

I imagine SoTC would really benefit from it.

[Blyss Sarania]

SoTC does benefit some, yes.

So does Kingdom Hearts with the 60FPS patch.

@Mkilbride - I'll send you a build to test once I get it sorted.

[H.A Treesong]

This could be interesting to test for while though.
As I'll get a new, powerful rig tomorrow I'd be happy to give it a try and report if it's got any benefits on my games (got about 50 games going from 2D arcade ports to full 3D madness like ZOE2).

[Blyss Sarania]

I personally don't feel it's needed. I think it will only cause new users to be confused about the 2-way slider. Sure it may make a few games a bit smoother, but I don't feel it's enough to make adding it worthwhile.

Yes, I am concerned about that too. That's why I made this thread - to determine if the benefits outweigh that risk.

[jlwmanagement]

You mentioned that only a few games gain from this (assuming I understand properly). Do you have a list of those games? If I own any of them I would be happy to test this for you.

[vsub]

I don't really see what's the big deal is.
I tried KH2FM patch and the game did look kinda different(smoother)but I wouldn't make the cpu do more work just for this...it's good enough for me even without the patch

Still,I would like to see how ToA will work with this.

[Nobbs66]

I think I might like a build just to test.

[H.A Treesong]

Still,I would like to see how ToA will work with this.

Oh yes, me too!