How about software rendering?

AFAIK PPSSPP doesn't have a software render. Ran the same game on JPCSP and it renders there on both OpenGL and software renders.

AFAIK PPSSPP doesn't have a software render. Ran the same game on JPCSP and it renders there on both OpenGL and software renders.

What is this?

You can't blame him for thinking it's doesn't exist because you need to mess with the PPSSPP.ini file for it to show up.
…
On Sat, Dec 11, 2021, 07:48 AEZ YT @.***> wrote: AFAIK PPSSPP doesn't have a software render. Ran the same game on JPCSP and it renders there on both OpenGL and software renders. What is this? [image: Screenshot_2021-12-11-15-46-30-905_org ppsspp ppsspp] https://user-images.githubusercontent.com/37603562/145668946-39ee6767-a17c-49ee-8cda-5323b8a1755a.jpg — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#15224 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/ASDWOXMQKAP4YEUN2DKUT4DUQL645ANCNFSM5J2RVGEQ .

I thought it's visible on windows platform 🤔

Yeah we hide the software rasterizer on Android by default for now, because it showed up in so many crash reports that I suspected people enabled it by accident, and it's still too slow, especially on ARM which doesn't have a pixel jit yet.

Might reconsider, but it's still not very playable in most games.

Correction: it is on Windows, but even on software renderer it's still a blackscreen.

From some bisecting someone did, it seems this may have started at 5d17ec7. If so, it's probably other timing (or even drawing timing) that's wrong...

Also, if that's the correct cause, then it's probably not only broken on Windows.

-[Unknown]

Reminder v.1.11.3 is the last working version.

Last build working is v1.11.3-1134-g92e7534f4
The issue started in v1.11.3-1138-ged3201a27

I can reproduce this for my android phone

Reminder: Do not enable DrawSyncEatCycles for Sakura-sou no Pet na Kanojo it can cause black screen.
I tested in v.1.11.3

Does that mean this works when it's not set? Doesn't seem like compat.ini by default has the game ID.

-[Unknown]

Does that mean this works when it's not set? Doesn't seem like compat.ini by default has the game ID.

-[Unknown]

NPJH50745 = true Blackscreen on Startup
NPJH50745 = false The game works

It seems Crash Tag Team Racing passed the checking memory stick screen when DrawSyncEatCycles is disabled for this game in latest builds.

Does that mean this works when it's not set? Doesn't seem like compat.ini by default has the game ID.
-[Unknown]

NPJH50745 = true Blackscreen on Startup NPJH50745 = false The game works

Okay, so if I downloaded PPSSPP onto a new phone, and ran this game - everything would work fine. However, if I manually add the game ID under DrawSyncEatCycles to compat.ini, then it starts crashing.

If that's true, it sounds like this isn't an issue, and perhaps was fixed by the cpu cycle timing changes or some other timing change. That you can make games crash or break by manually adding them to compat.ini is a feature, not a bug. There'd be no reason for that file to even exist if adding random game IDs to it had no downside.

-[Unknown]

v1.13.2-915-gc37049fa1
Still happens
log.txt

Was that with or without DrawSyncEatCycles? I'm still confused based on the previous comment.

-[Unknown]

black screen with both options even when it was set to false could not get this to work properly

NPJH50745 = true Blackscreen on Startup NPJH50745 = false The game works (didnt work for me on my PC)

@benderscruffy can you check v1.11.3-1134-g92e7534f4 and earlier if the DrawSyncEatCycles set to false works properly?

@Panderner do you know where i can get v1.11.3-1134-g92e7534f4 from ???

@benderscruffy the stable v1.11.3 works fine for this game.

where can i get it from then ??? on the site last one is 1.12.3

@benderscruffy select old versions on the download selection.

If DrawSyncEatCycles enabled for this game gets black screen on start up.

I confirm retvert 5d17ec7 can fix the issue,
I try to change another timing.
v1.16.6-333-gecab50346-windows-amd64 full debug log:
https://gist.github.com/sum2012/41fba1545753f374048259798a12f455

It is modify log
https://gist.githubusercontent.com/sum2012/63625cb8dfd16d06ae3de309fa39cb4d/raw/5487222d369f96917ff83e2a12d19875b3166601/gistfile1.txt

edit:try sceKernelUnlockMutex hleEatCycles(6000); don't work

DrawSync not taking any time causing different behavior must mean something else later happens in a different order.

I'd suggest trying to narrow down to a specific instance of it. In the log, there are a few patterns of DrawSync usage. Does the amount of cycles the syscall takes matter for each of them? Only the first one where it creates a list right after? The other ones only? You want to determine at least what area matters so you can see what else is "in flight" at that time - what other timing matters?

If the game was only doing DrawSync, with zero other threads doing anything, it is extremely likely that the timing "wouldn't matter". You could make it 100000 cycles and it at worst might affect the framerate. The cycles it consumes here matters mostly because it impacts the timing of other things happening at the same time. Think of it as racers running on a track. By deleting the cycles it takes, it's like you're warping one of the racers ahead of the others. Maybe that makes them trip over something that no longer gets in another racer's way? It can have complicated impacts like that. You have to look at the other threads to figure it out.

By the way: every syscall that consumes 0 cycles technically is a bug. I promise you, there are zero syscalls on the PSP that took zero cycles to execute. That said, some of them were pretty quick.

-[Unknown]

The last one is important. After the game run sceGeListUpdateStallAddr , sceGeDrawSync need 0 to 31 cycles to show the screen

baidu forum remind this issue .
Cannot make this game for compat ?