Post subject: FCEUX/Mesen/puNES Overclocking Compatibility List
AxlRocks
He/Him
Joined: 8/26/2016
Posts: 5
Hello all, as per feos' suggestion, I've gotten around to posting the list of games I've tested for overclocking compatibility; 124 of them at the time of typing this post. as of May 30, 2017
    273 NES and FDS titles tested (some multiple times), 271 Good, 2 w/ Issues, 0 Unplayable 10 Re-tests in Mesen, 10 Good
The List: https://docs.google.com/spreadsheets/d/1exirXMWNp20OWdgRjosIrRh38-cpXLouga_P6A-jHp0/edit?usp=sharing Now, about the testing itself. All games were tested from start to finish in FCEUX. That includes demo loops, cutscenes, and credits sequences. In the case of arcade style games which lack a proper ending, they were either played until the levels looped back to level 1 or for anywhere between 15-120 minutes. No-Intro NES ROM set 2014-08-25 and No-Intro FDS ROM set 2016-12-25 used for all tests except where absolutely necessary. All games were unchanged with only two types of exceptions (both of which are always noted in the spreadsheet):
    1.) Some games had incorrect headers. This applied to:
      a.)Chiller (USA) (Unl) b.)Ms. Pac Man (USA) (Unl) (Tengen) c.)Saiyuuki World (Japan) d.)Galaxian (Japan) requires an overdump to play in FCEUX because it's too small for regular ines header specifications.
    2.) If translation patches were needed, like in RPGs or text adventures.
I typically used 240 Post-Render Scanlines, however games which required more or required VBlank Scanlines are noted. Many games were tested on various FCEUX revisions prior to FCEUX 2.2.3 releasing. Unfortunately, I did not take note of the versions used but I believe very few, if any, games will be affected negatively by changes between r3120 and 2.2.3. Games which are affected:
    1.) Kirby's Adventure (USA) - Demo loop and credits sequence has sped up audio after FCEUX r3223/r3224 and in Mesen(Not a problem in 0.9.1) and puNES.
I'm always looking forward to suggestions or anyone else insane enough to test more games.
Site Admin, Skilled player (1236)
Joined: 4/17/2010
Posts: 11268
Location: RU
Wow, that's some insane amount of work right there! Thank you! A few questions: What's the difference between Perfect and Good? What about Castlevania 2 and 3? Regarding versions, I don't expect any regressions between first overclock commit and 2.2.3 either, while all the upgrades due to FHorse's ideas are expected. So yeah, whatever was already working well back then, should work the same now.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
AxlRocks
He/Him
Joined: 8/26/2016
Posts: 5
Thank you! The only difference between Perfect and Good is that Perfect can be added if the tester knows the game very well or has played it several times without issues. The former is kind of subjective so I've debated on whether to remove that and only stick with multiple playthroughs or even remove the Perfect rating altogether. Castlevania 2 and 3 have been on there for a very long time and I just marked them as "To Be Determined." I haven't gotten to them due to the fact I'm not very familiar with CV2. But I might finish CV3 first since it's a more straight-forward platformer. That's good to hear regarding versions too.
Site Admin, Skilled player (1236)
Joined: 4/17/2010
Posts: 11268
Location: RU
I think Perfect can be dropped then, since it doesn't prove there's something special in a given game's overclocked emulation that makes it better than Good.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Player (170)
Joined: 7/1/2016
Posts: 263
In "Life Force" (mostly on the second level) if there is to much objects on screen your ship can become invisible for the couple of seconds (not a couple of frames) instead of flickering or something. This is pretty annoying.
AxlRocks
He/Him
Joined: 8/26/2016
Posts: 5
feos wrote:
I think Perfect can be dropped then, since it doesn't prove there's something special in a given game's overclocked emulation that makes it better than Good.
I agree. I went ahead and removed it.
Koh1fds wrote:
In "Life Force" (mostly on the second level) if there is to much objects on screen your ship can become invisible for the couple of seconds (not a couple of frames) instead of flickering or something. This is pretty annoying.
That's weird. For some reason, I can't seem to get that to happen. I'll replay it a couple times and see if I can.
Player (170)
Joined: 7/1/2016
Posts: 263
I see now. It happens even without overclocking in japanese version only. My bad.
AxlRocks
He/Him
Joined: 8/26/2016
Posts: 5
Koh1fds wrote:
I see now. It happens even without overclocking in japanese version only. My bad.
Oh, hmm, that's interesting! And it's no problem :) I got another full test from trying that out anyway, so it's all good.
AxlRocks
He/Him
Joined: 8/26/2016
Posts: 5
feos wrote:
Regarding versions, I don't expect any regressions between first overclock commit and 2.2.3 either, while all the upgrades due to FHorse's ideas are expected. So yeah, whatever was already working well back then, should work the same now.
Yesterday I found an exception to this finally. Kirby's Adventure (any NTSC version). The PAL versions don't have this problem. The demo loop and credits sequence have sped up audio, however, the rest of the game is fine. It seems to have broken in r3223 or r3224. In r3222, the audio is normal like in the rest of the game. The same issue occurs in Mesen and puNES too. Any ideas about what is causing this? If it's actually more accurate behavior that other games might require, could an option be added to specifically fix Kirby's Adventure, and possibly other games with this same issue? (Assuming there are any others.) EDIT: 8-21-17 Mesen 0.9.1 can now overclock Kirby's Adventure without these issues by using a Before NMI overclock of 480. I haven't done a full playthrough yet, but so far appears to be just fine.