Closed CPkmn closed 11 years ago
Another confirmation, the FramebuffersCPUConvert and FramebufferstoMem on AMD cards seems to fix previous issues reported. I have an AMD 6850. Investigating working fine.
@Tittiez
Thanks for the confirmation. So that's 2 confirmed success stories so far. One on my legacy 4670 card and one on a much more modern 6870 card.
Can you post up a build of it that's already compiled? I'd like to test it, but I'm not sure how to clone/compile it on my own. I have a Radeon 5850.
@Leacero Sure, give me a few minutes to upload.
@Leacero I think this will do it: https://docs.google.com/file/d/0B1myg_QircHaVXNJbTU1bms1NHM/edit?usp=sharing
Not sure if it works on other computers yet.
Tested with the build you uploaded and it works ! I can examine the drawers and finally continue. Thanks to arnastia and Tittiez !
Thanks! So far it's working well, got past the initial roadblocks in Chapter 1. I'll keep playing and let you know if there are any other problems that come up. Again, I'm using a Radeon 5850.
@arnastia
Judging from the replies, this does seem to be the magic fix we've been waiting for.
Radeon 7950 here. Works perfectly so far (in all the formerly problematic areas). Fantastic job debugging this.
Well, good to know, but I didn't quite get if anyone got it working without CPUConvert/FramebuffersCPUConvert (meaning only FramebuffersToMem = True) on AMD/ATi.
As for Android, I'm still trying to find the sweet spot on that one (haven't made the pull request because of this). I think I have a pretty good idea about what I'm doing wrong, but it's hard to be sure when I don't have a single OpenGL ES device to test this on. Only test I did of the OpenGL ES version of the code was on my PC, and it worked, but if no objects were even detected then I have to try switching channels around some more.
@arnastia To my understanding, the Nexus 7 has OpenGL ES, being an Android device. If you can compile a build for me, I can test if it work, since I have superuser access on the 7 and thus can edit the .ini files. Come tomorrow I could have a report for you.
Edit: Sorry, turning off FramebuffersCPUConvert makes things stop working.*
@arnastia
I assume all the positive responses are with FramebuffersCPUConvert = True and FramebufferstoMem = True.
The CPU conversion is still necessary with your provided build.
I switched to using the new build with FramebuffersCPUConvert, and something is broken that's wasn't broken before. However, I can confirm it's not the CPU convert fix that broke it, as it's also broken in the last PPSSPP build (8.1.220). In the 4th chapter, the game freezes when the corpse discovery announcement is about to be made, and it gets passed this on the old build I was using (8.1.74). I guess I can try isolate exactly when this broke. I've also run into problems like this I haven't seen anyone else reporting. For instance, at the start of the trial for chapter 1 and 3, after choosing "finish preperations" the game froze using 8.1.74. I've been using JPCSP to get passed these parts.
@alyosha-pushak
It'd be very helpful if you provided a savefile made right before the freeze so that @arnastia and the rest of us can figure out what went wrong.
I've determined that it's not an issue with a particular build. The following behaviour appears to be consistent in any build passed 8.1.74: Load game -> Load save -> get to corpse discovery announcement -> game freezes Load game -> Load savestate -> get to corpse discovery announcement -> works ok Load game -> Load savestate -> Load save -> get to corpse discovery announcement -> works ok
Because the 3rd path works ok and that's how I first tested it on 8.1.74, I thought it had something to do with the builds, but it does not. If anyone wants to test it, you can get the save file and save state here:
https://docs.google.com/file/d/0B_OtR-yWg6u_R1FaSEVHaVJLTEE/edit?usp=sharing
This will spoil who the victim is in chapter 4 is, if anyone is concerned about that.
Trying to get this going, just installed what Tittiez compiled, trying to mash O back to his room.
It'd be great if anyone could post a savefile from just before that for those playtesting this OpenGL bug.
EDIT: I recovered the save files that I previously stupidly deleted. Will upload if anyone needs the post-Prologue save.
EDIT #2: So far so good, I'm on day 4, no problems yet.
EDIT #3: For what it's worth, I've got an AMD HD7750. Everything seems to be running borderline perfect now, with the exception of some of the cut-scenes slowing down, as arnastia described above. I'm beginning Chapter 2 right now, and by the way...thank you to everyone here for fixing this stuff. The game is fantastic.
Yeah, sorry, but issues with anything past free time chapter 2 and you'll have to wait for me to get there (assuming it's something I've been messing with). Needless to say, I really don't want to be spoiled on a game like this.
I'm however pretty sure you can finish the game on an Nvidia, at least with the old version of my fix, since I know people who actually have.
Thought people might like to know, I experienced a crash at the the ballot. Luckily. I had a save state and only lost 2 minutes worth of work. The 2nd time I tried it didn't crash. I'd advise to make a savestate shortly before you know the ballot will be cast.
I finished Chapter 2 yesterday and had no crashes so far. Is this on chapter 4 ?
So I downloaded the one tittiez uploaded and I can't get past the first room .then I enabled buffered rendering but it lags very much .using windows 7. what gives ?
@bneundh, disable all other video options or simply get a better GPU.
Ok, so I pushed another try to my repository. Would appreciate it if anyone could test it on AMDs (with and without FramebuffersCPUConvert = True) and Android (FramebuffersCPUConvert changes nothing on this one).
If nothing's wrong with the AMD version I'll make a pull request later today. If the Android version still doesn't work, well, I'll probably only fix it when I can get my hands on a rooted Android device. I'll also look into doing things in SSE/NEON later.
Repository: https://github.com/arnastia/ppsspp And a x86 build for those with AMD cards: https://dl.dropboxusercontent.com/u/3475202/PPSSPPWindows.zip I'd post an Android build, but don't have the tools to compile it right now.
EDIT: PS, you still need FramebuffersToMem = True on both of them, and the slow video bug isn't fixed in these builds. Just use turbo/start to get past it.
@arnastia I can't comment on the entire game, but I'm in my room right now just at the start of the 3rd Chapter with FramebuffersCPUConvert = False and I can still investigate things fine.
AMD 6850
I haven't tried it with my older build, but at least the newer build doesn't seem to need it for AMD GPUs. Good job!
now when I enabled buffered rendering it just freeze . I must have done something wrong ..
Actually yeah, I found a few problems I haven't noticed before. I'll post them, hold on.
I'm aiming at my door from a far away distance, yes? If I press X, it takes me into the BATH.
And another one:
If I press X right now on Kirigiri's room, it'll take me into MY room.
I should be too far away from the rooms I'm aiming at for that to even show up.
This is with FramebuffersCPUConvert = False would you like to to test True as well?
Just to be clear, that's with FramebuffersCPUConvert = False, right? With FramebuffersCPUConvert enabled everything should still work.
If so then Nvidia and AMD probably interpret the type parameter in glReadPixels differently, and there isn't much I can do other than add an option to assume either AMD or Nvidia card (or is there some way to detect which card the user has?). Otherwise just keep FramebuffersCPUConvert enabled.
@arnastia in the worst case, we can get the vendor and do something differently based on it. Should be no reason to ask the user if they are using AMD/ATI, since their video card can answer that question for us.
I wonder if anything is being returned in glGetError()?
-[Unknown]
@arnastia Yes, that is with FramebuffersCPUConvert set to False. Setting it to True seems to fix the issue.
Is it a known problem that texture filtering isn't working properly? It seems to get randomly deactivated, and sometimes you can reactivate by switching it off and back on.
I guess it's a side-effect of the fix.
ok, i know this is the wrong place to ask, and if you can tell me where to ask stuff like that pls do, but anyway: I would like to know how to build this projekt. I have basicly no real knowledge of C++ or programming, but i do know what building is for and so on... i've installed GitHub, Visual Studio and everything i need to have installed (i think, at least)... now i only have the problem that the native library does not load... and i am not totally sure that i understood:
"First of all, after having checked out the source, don't forget to run:
git submodule update --init
in order to get the "native" library."
because it does not work as far as i understand, i have to open a shell in the repositorie and then enter "git submodule update --init" i did that and it looked like it worked but it didn't solve the problem, i still couldn't load "native" in Visual Studio (everything else works i think)
edit: Well, the thing is, i do not get an response from the command and nothing changes, but thanks, i hope i'll figure it out on my own anyway :P
Better to ask stuff like this on the forum. Remember to run the command inside the directory you cloned ppsspp into. Post in Development (ideally with whatever error message or response from that command you're getting, etc.) and I'll look a bit later.
-[Unknown]
I'm running PPSSPP on my Android 4.0.4. Is there any solutions to this? If there already is, can anyone link the comment? Thanks, (also, please @mention me)
You must root it if I'm not wrong. Use a computer instead, it's going to lag a lot anyway.
@HermitCrap I doesn't lag much on my phone (not after I go change a lot of configurations to find out how to fix this issue (of course didn't work)). But it's always more enjoyable playing on consoles because using a PC to play games or watch movies always makes me feel unproductive and irritating :)
If you have an Android PPSSPP Build with the fix integrated to the code, you need to root and then use an app like "Root Browser" to access the hidden app files within Android/Data/Data/PPSSPP. From there all you have to do is switch the pertaining values from "False" to "True" within ppsspp.ini with an Android text editor.
If you do have a compiled PPSSPP build with the CPUConvert fix, hook me up, will ya? My Comp has been pretty weird lately. RAM problems, maybe the connectors. Want to avoid using it much if I can help it.
--- Original Message ---
From: "potasmic" notifications@github.com Sent: July 7, 2013 10:12 PM To: "hrydgard/ppsspp" ppsspp@noreply.github.com Cc: "NewDCD" diegoj10@hotmail.com Subject: Re: [ppsspp] Danganronpa bug (#1686)
@HermitCrap I doesn't lag much on my phone (not after I go change a lot of configurations to find out how to fix this issue (of course didn't work)). But it's always more enjoyable playing on consoles because using a PC to play games or watch movies always makes me feel unproductive and irritating :)
Reply to this email directly or view it on GitHub: https://github.com/hrydgard/ppsspp/issues/1686#issuecomment-20583731
Alright, I've made it to Chapter 5 and cannot continue. I'd like to go into detail on where the crash is, but I don't want to give away spoilers. Basically, at a certain point of dialog, the screen fades to white and PPSSPP simply stops working.
Is there an error log that I can copy/paste here?
Hey I'm experiencing a problem extremely similar to Kwazzi's. The only difference is that mine happens in chapter 6. (Not to be rude but are you sure it is chapter 5 kwazzi?)
So like kwazzi I am wondering if there is something I can post like a save file or error log.
@MegaPear Is it after Kirigiri says a line in the bath changing room? I might be mistaken on the chapter, it's happening during an exchange with Aoi.
Yeah, that's when it happens.
With me as well, after the flashback to Kirigiri's dialog in the bath, the screen fades to white and then the emulator crashes.
If you're crashing in chapter 6 during a conversation with Aoi in the data processing room (specifically at the Kirigiri flashback) try turning off Hardware transform in the options menu.
Turning off hardware transform let me continue past that section.
@peleides
Good call, thanks for the tip. Works perfectly.
@peleides
Excellent advice. Will add that to my thread in the forums.
Just tried playing PPSSPP on my AMD Build, the performance hit is really big with FrameCPUConvert on. Since PPSSPP starts to stutter a lot when loading or when you open the ElectroID. Other than that, it works fine. System Specs: I5 - 4670K with HD 7870 PCS+ Must Edition. I suspect the stuttering have something to do with my slow hard drive but yeah.
@HermitCrab it doubt its your cpu or gpu, because it runs full speed for me and I have a Phenom ii x4 955 and a 6850.
Are you sure you are using a dev build?
humm that's mean only SW T&L works only in chapter 6 , just wonder any log during crash ? we may be able to debug it.
Yeah that is weird, I don't know any other cases of HW T&L causing crashes...
I am using a version of PPSSPP posted earlier here. (Version by @Tittiez http://puu.sh/3B7qB.png) I am at chapter 5, everything works perfectly fine thanks to that version. (Normal PPSSPP made it impossible to select objects with the crosshair) HOWEVER, the game's screen fades to black, and then nothing happens anymore, making it impossible for me to continue playing this game.
When this happens (may contain spoilers, but I don't think so): Monokuma act XVI, monokuma shows a video of another game. After the video ends, he says "This is the era of ninja! Samurai are a thing of the past!" "By the way, why didn't I get an offer?" "I'll buy it anyway, though!" "I'll buy -two- copies! grin" (Screen fades to black) "Diiiing doooong... Diiiing doooong..." (I press X, Screen fades from black to your room) (after a second it fades from your room back to black) (It stays like this forever, the buttons are completely unresponsive and nothing ever happens)
I use a radeon hd 6970. (ati)
I attached a .gif of the issue:
I hope someone knows how to fix this, I really don't want to stop playing here.
Oh and, good work you guys, thanks for all the effort.
EDIT: @solarmystic right below me This fixed the issue, thanks! The game looks much better too!
Github won't let me write a new reply though so I have to edit this one
I tried a search in the issues here to see if anyone reported this yet, but it seems no one did.
While playing Danganronpa (both the demo and the full release) in PPSSPP I noticed objects are not interactable. This basically makes the game unplayable since the requires object interaction on many occasions. I notice the same issue on JPCSP, however the software rendering mode in JPCSP works with the objects (and has since at least revision 2450, the first software rendering public release; https://code.google.com/p/jpcsp/source/detail?r=2450). Unfortunately I can't check earlier releases' software rendering modes since I don't know how to compile JPCSP.
I'll attach a picture of what PPSSPP and JPCSP (in software rendering mode) get when hovering over a should-be-interactable object.
(JPCSP in software rendering mode)
(PPSSPP)