Posts for Alyosha


Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
Game: Castlevania III (USA) Emulator: BizHawk: 2.8 - NESHawk Console Verification Device: TAStm32 V3 Movie: https://tasvideos.org/7743S Description of Desync: Randomly spawning enemy spawns differently on console compared to emulator ~ 1.5 minutes into the game. Research: Presumably the error is in MMC5 IRQs, as this game uses MMC5 which has not been tested or updated in some time. Possible Next Steps: Test ROMs needed Status: Open. Closed. resolution: Innacuracy in reading back state of DMC channel caused the desync. Test ROM confirmed correct timing. Additionally, unemulated open bus behaviour caused desync later in the run. This is also fixed. MMC5 IRQ timing also verified.
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
Unfortunately this run desyncs on console. About 1/4 of the time I correctly get the holy water, but then I get hit at the random spawning monsters about 1 minute later. In 4 trials of successfully getting holy water I get hit by the same enemy in this area and fall in the pit. This is presumably an issue with MMC5 IRQs, which have not been tested to the same level of precision as MMC3 IRQs. Disappointing, but thanks for making the effort to get it to work on 2.8 for testing anyway scrimpeh.
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
Oh cool you were able to do it, awesome! I'll try to console verify when I get a chance (maybe this weekend.)
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
I was able to get up (almost) dracula without too much trouble (only one enemy is visibly different between the resync and the orignial, I flying skull in level 3) with no added frames. However I got a desync on the scroll glitch right before dracula. It looks like a controller input is missed and causes a desync in the tiles generated in draculas room. I don't know how to fix precisely, if some other approach can be used, or if dracula can be fought another way. Here it is if you want to try, I think this is as far as I can go with my limited game knowledge: https://tasvideos.org/UserFiles/Info/638016425343155312
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
Look like this game has a 'press select until you get what you want' type of RNG. I was able to resync up to the first boss in 2.8 but can't get the orb to fall as fast as in this run yet. I'll keep trying though, would be cool to be able to console verify this.
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
Patashu wrote:
Probably lots of game-end-glitch/crash type stuff. Isn't there some glitchy warp stuff in Battletoads that isn't fully maxed out, or did it all get figured out by now?
I'm not sure there is anything noteworthy left in battletoads (game end glitch run already verified, so probably nothing time saving anyway.) Here is a SMB3 glitch that would be cool to replicate (around 5:30 in the video): Link to video Also I recall a boss despawn in blaster master: https://www.twitch.tv/gamesdonequick/clip/OilyTiredPorpoiseNomNom?lang=it Of course with no movie backing them up these could be hardware error, but it would be amazing if they really are re-creatable.
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
Console Verified! Link to video Really cool to see a mystery solved through the power of Jaffar bot, and bring together a known improvement on top of that. I wonder if there are any other mysteries that could be resolved in this way, I couldn't think of / remember any off the top of my head though.
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
A masterpiece, certainly the most impressive thing ever done with controller inputs. I also really enjoyed reading through the explanations and watching the videos about all the tech involved. The amount talent and tech that went into making this is amazing, really inspiring stuff. Unfortunately not entirely reproducible, which i guess is fine for a demonstration, but really too bad emulation couldn't do much to help things along here, and thus doesn't leave much afterwards either. Oh well. I'm also fascinated by the combination of real time runners and tool assistance. I think this would be an interesting direction for TASing to go in for more modern games, to still get super human results in moderate amount of time compared to doing everything frame advance for endless hours. Anyway, all around amazing, great project!
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
Console Verified! Link to video
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
I console verified this run with the improved movie, old one removed, publication page updated.
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
Another Ninja Gaiden improvement nice! Again thanks for putting in the effort to make it work in NESHawk, those DPCM effects do matter for accuracy. Hopefully you don't have to encounter a case where the conversion isn't possible, that would be a shame, but I guess we can cross that bridge if we ever come to it. Anyway I'll work on a verification eventually if no one else gets to it first. Great work!
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
More impressive bot work, nice! The bot driven races do look a bit rough though, maybe the fact that you can just bounce off the cpu's to keep pace is effecting the search (ex 25:57 in the encode.) Still cool to see another type of game tackled by your bot. Maybe try puzzle games? I wonder what it could find on ex. Adventures of Lolo games.
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
https://tasvideos.org/3589M It looks like your first fight is faster, so maybe you can improve the other 2 from there, but you certainly don't need to do the training sequence.
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
Huh, the improvement looks so obvious once you see it. Seeing stuff like this always makes me wonder what other seemingly simple tricks and optimizations are being missed out there. Nice work.
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
I didn't expect to see an improvement to this game, nice work. Console Verification: Link to video
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
I'll be using this thread to keep track of work I'm doing related to GBA emulation and testing with the goal of console verification of GBA TASes. Code: https://github.com/alyosha-tas/GBAHawk This summer I had spent a lot of time working on GBA emulation. I figure this is the next most likely console to see new console verification progress, so I decided to make an emulator core myself to understand the console and hopefully use to make console verified TASes. Originally I wanted to experiment with a modern time keeping / task scheduling style used in other GBA emulators, but after looking over the specs on GBA Tek, there looked to be too many moving parts for me to easily keep track of. So, I went with my usual approach of emulating the whole system one clock cycle at a time. I'm glad I did, as there a lot of nuances, and I spent pretty much all my time focused just on accurate timing of the basic building blocks (cpu, timers, dma, prefetcher) and I haven't even gotten to the ppu or sound much yet. So far though I have really enjoyed the process, this console does a lot of cool things, and I'm really impressed with how everything is put together, and it has been fun learning about a slightly more modern architecture than what I am used to. Anyway, I have just enough working to present some early results. So far I can pass the Aging Cart tests, get 2020 on mGBA suite ( https://github.com/mgba-emu/suite ) timing tests, and 1256 on the DMA tests, but other emulators can already do that. Additionally, I can get 936 on the mGBA suit count up tests, pass the irq_delay and isr tests ( https://github.com/destoer/gba_tests ) and pass the AGBEEG cart tests ( https://github.com/GhostRain0/AGBEEG-Aging-Cartridge ) all of which I believe no other emulator can do so far: PPU emulation is my next goal, I only have enough timing details down to pass the basic tests, a lot of work left to do. I haven't properly started rendering yet either, so I have a long road ahead of me. That's about it for now, my next immediate test goal is the misc. edge case tests on the mGBA suite, which aside from PPU timing I also need to sort out a few details on open bus emulation for. I'm hoping to be able to make a decently detailed post every month or 2 with progress and ideas.
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
Everything in the previous movie besides the final boss fight was done by Aglar and Randil, so if anything only they need to be added as authors, although personally I think its fine for eien86 to remain sole author.
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
Wow what a great improvement, the fight looks so smooth now! Thanks for taking on this challenge, this is some really impressive work. It's really cool to see general purpose botting advancing like this, it's something I wouldn't have considered possible even 5 years ago. Console Verification: Link to video
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
eien86 wrote:
Not really sure. I hear suggestions!
I think it would be really interesting to see what it can find on the last boss of [2777] NES Ironsword: Wizards & Warriors II by Aglar, Randil, Alyosha, rchokler & Samsara in 07:14.32. EDIT: Updated Console Verification Video: (removed, re-verified with improved movie) I also looked into the pacifist run, but it desyncs in the latest BizHawk version.
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
TAS is console verified! (Took ~10 tries) -EDIT: video updated below I forgot to turn on 60 fps recording so the quality isn't very good, I'll replace it with a better one later this week. Really cool to finally see a verified Castlevania run, what game do you think you will try your bot on next eien86?
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
Amazing work, great combination of innovative botting combined with deep technical knowledge of the game. Great write up too. There are already several console verifications of games with RNG as you describe here, its not that bad. Chance of success is something like 1/4 or 1/8. I'll try it assuming no one else does it first and I can figure out what a PRG0 cart looks like.
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
hndfhng wrote:
Is 1 audio channel missing a normal thing for this game?
For some reason I had one of the audio lines disconnected to my capture device. Strange, I'll re-encode. EDIT: Done
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
I converted the movie to work in bizhawk 2.8 and in GBA mode in gambatte. I tried to console verify and it fails. Tried both floor and ceiling in the dump script. I also made sure SRAM was set to 0xFF. The movie also syncs in dev build. It desyncs in the first level, donkey kong gets hit by an enemy shortly after collecting the 'K'. I'll try and see if I can narrow down what's going wrong at all, would really like to see this one verified. Turns out I was initially using a wrong version of the script. Using the correct version it still fails though, desyncs at loopy lights. Console verified! Link to video
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
https://tasvideos.org/UserFiles/Info/637948905739299716 saved a couple more frames in 4-1 with better jumping at the start, but i messed up the ending and don't know how to fix it (fade out takes longer), i think the level can still be a bit faster too. Also since you can clip in 2-3 now, can the 1.1 version be used?
Alyosha
He/Him
Editor, Experienced Forum User, Published Author, Expert player (3536)
Joined: 11/30/2014
Posts: 2733
Location: US
similarly around frame 23080, take a small jump off the spring then a big jump up the ladder.