Thread 105695960 - /g/ [Archived: 813 hours ago]

Anonymous
6/25/2025, 3:53:30 AM No.105695960
sddefault
sddefault
md5: 2f0732230f2a3eb1b80f0380a351b44d🔍
Why is tearing still so precarious on wayland? Some desktops refuse to support the protocol, and even those that have support hide it behind a config file. Only Vulkan windows will then go ahead and use it. OpenGL is SOL without patching Mesa.
Replies: >>105696020 >>105696395 >>105697620 >>105697704
Anonymous
6/25/2025, 4:06:16 AM No.105696020
8bfa6ea0b7d9090a525d1cf835ad6d2c-140600317
8bfa6ea0b7d9090a525d1cf835ad6d2c-140600317
md5: 86f93bc1b0ef4b4ab196bc3e70ead866🔍
>>105695960 (OP)
https://wayland-book.com/protocol-design/design-patterns.html

>The most important of the Wayland protocol design patterns is atomicity. A stated goal of Wayland is "every frame is perfect". To this end, most interfaces allow you to update them transactionally, using several requests to build up a new representation of its state, then committing them all at once.
Replies: >>105697620 >>105699486
Anonymous
6/25/2025, 4:09:44 AM No.105696033
Hyprland gave me issues, but I got tearing to work across the board on Sway.
Anonymous
6/25/2025, 5:25:29 AM No.105696395
>>105695960 (OP)
>still doesnt work on intel integrated graphics
PLEAAAASE INTLELAVIV I WANT TO PLAY TF2 WITHOUT FORCED VSYNC :<<
Replies: >>105697255
Anonymous
6/25/2025, 5:57:45 AM No.105696590
I don't have that issue on my 5090/240hz monitor
Anonymous
6/25/2025, 8:03:47 AM No.105697255
>>105696395
Doesn't work on AMD either now.
https://gitlab.freedesktop.org/drm/amd/-/issues/4263
Anonymous
6/25/2025, 9:18:44 AM No.105697620
>>105695960 (OP)
see>>105696020
basically tearing is antithetical to their stated goal of "every frame is perfect"

see also https://github.com/swaywm/sway/issues/4763#issuecomment-578462619

tldr all semi modern gpus have at least 2 hardware planes, a cursor plane and a primary plane, rendering occurs on the primary plane, the purpose of the cursor plane is so that if rendering stalls, your cursor is unaffected and doesn't lock up. the old drm interface exposes this and on X11 this works perfectly, however in this case with sway, they use the new atomic interface to sync the cursor plane to the primary plane because if they don't they can get tearing on the cursor plane... and the effect of this is that you effectively get a software cursor, so if your compositor's framerate dips so does your cursor....

truly awe inspiring levels of retardation.
Replies: >>105697659 >>105698794 >>105698819
Anonymous
6/25/2025, 9:26:42 AM No.105697659
1747882808101535
1747882808101535
md5: d58cfd90dfa65489362eed98e2881619🔍
>>105697620
I'll never understand waytroons and their asinine vsync obsession. I turn off my compositor for drawing because my pen always feels sluggish as fuck with it on. X11 lets me do it, so why can't they? Or do they think nobody has a use-case for low input latency workflows? Actual blithering retards.
Replies: >>105697775
Anonymous
6/25/2025, 9:37:08 AM No.105697704
>>105695960 (OP)
Because you're living in 2015
Anonymous
6/25/2025, 9:52:10 AM No.105697775
reactiontime
reactiontime
md5: 380091deace6a1c8cd76b803bfa19562🔍
>>105697659
They never used CRTs and they don't do anything but use a web browser+web browser LARPing as an IDE. There is really nothing more to it than that. They think multiple ms response times are normal because they've never done anything creative in their lives and they've never played vidya at a level where that 1-2ms reaction time matters. There are plenty of use cases where I want "tearing" because I'm doing stuff like drawing to a CRT or something I'm using wants to be able to draw before waiting for vsync. There is absolutely no reason why you shouldn't be able to turn the compositor off. I mean, that was a basic requirement for most people playing competitive FPS games like Q3A and 2D fightan games. I don't give a fuck about tearing when I'm doing that. I'd rather have the ability to see what other players are doing in that -/+6ms where it matters most.

It all starts making a lot more sense when you understand things like wayland aren't designed for people using workstations for actual work and play. It's designed for kiosks and other similar applications like POS terminals and displaying menus and advertisements on those annoying LCDs at fast food places. You know the kind that replaced the actual menus and do nothing but play food porn while you're standing in line waiting for your slop.
Replies: >>105697905
Anonymous
6/25/2025, 10:12:03 AM No.105697905
>>105697775
>They think multiple ms response times are normal
more people need to point this shit out. 1 frame.
Replies: >>105697996 >>105698051 >>105700184
Anonymous
6/25/2025, 10:27:40 AM No.105697996
>>105697905
Most modern "competitive" games are designed around shit response times caused by bluetooth controllers and modern display devices. It's so bad that when crossplay was introduced to a fightan game I played we discovered it had a 8+ms built in lag to all inputs. The game was available on both PC and PS4 at the time. A mod was released for the PC version to take about 2-3ms off the input lag. It really improved the game and made it viable to do a lot of combos without guessing. It was much better in local vs. play and didn't hurt anyone's experience in online play at all.

What does the company do? They declare this is cheating and install a rootkit on everyone's PC in an attempt to stop it. They never bothered to fix it either even though it would have been easy to fix on the console version as well and they'd just pushed out a patch to install said rootkit and fix other bugs.

Later on people dug into the source code some more and discovered that the game was programmed in such a way that the PC player always experienced lag from the horrible rollback netcode. A fix was developed for this to split it between both sides. What does the company do? They declare this is cheating and push out another patch to punish the PC players even more. After that I gave up on playing it all together. At least in online play.

They were dead set on catering to the console side of things and giving them every advantage. It was so horrible. My friend had it on console and lived down the street from me. He used a wifi connection so the lag was horrible. On his end though it looked smooth and was nearly as good as local play. On my end he was always teleporting all over the place. I'd score a hit and be 4 hits further into a combo then it'd roll back to the point where I hit him and of course he's now blocking. He claimed I was just mad and this wasn't really happening until I recorded some matches to show him just how horrible it was.
Replies: >>105698051 >>105698091
Anonymous
6/25/2025, 10:36:49 AM No.105698051
>>105697905
>>105697996
Another example I can give you was this FPS game I used to play somewhat competitive. A bunch of us dug into the Xbox 360 version because we were attempting to figure out a way to let them join and play on the PC servers. Since the company refused to push out updates/patches any longer and we wanted to modify some rules for the team vs. mode.

We discovered that the best way to play the game on console was to use an old CRT TV. Since there was programmed input lag if the console was connected via HDMI or outputting video above 480i. If it detected it was running in non-HD resolution it would by-pass the programmed input lag and it picked up about 20-30fps. Anyone playing on CRTs had a huge advantage. But we also discovered that there was some kind of aim bot programmed into the console version. Bullets would curve towards targets if you got the cross hairs within about an inch of the target. This made going things like shooting rocks out of the air or flying enemies really really easy. It was so easy that if you hooked up a mouse and used an adapter to make the console think it was an analog stick then jacked up the sensitivity beyond the level it allowed you to set through the GUI using a hacked save file you were near unstoppable. You could just spin around firing randomly and score head shots whenever you wanted without trying.

We ended up developing a way for the console players to join our servers through hacked save files. They could only play the default maps but they could play with us under our custom rule set (we could control spawns of items+NPCs). This was good fun for about a week until they figured out the aim bot thing. We ended up having to ban them off our server because it was impossible to have an honest game with them. They could use the mouse thing plus take advantage of a bug we couldn't patch to steam roll any team no matter how skilled.
Replies: >>105698098
Anonymous
6/25/2025, 10:43:29 AM No.105698091
>>105697996
>They were dead set on catering to the console side of things and giving them every advantage. It was so horrible. My friend had it on console and lived down the street from me. He used a wifi connection so the lag was horrible. On his end though it looked smooth and was nearly as good as local play. On my end he was always teleporting all over the place. I'd score a hit and be 4 hits further into a combo then it'd roll back to the point where I hit him and of course he's now blocking. He claimed I was just mad and this wasn't really happening until I recorded some matches to show him just how horrible it was.

its really not surprising, the primary reason crossplatform play was avoided for so long was because pc players will wipe the floor against console players, especially when the performance gap was so huge back in the day(imagine being on a state of the art athlon 64 with a geforce 6800 ultra getting 100fps on an hd monitor vs fucking ps2 players on a crt tv),
by nerfing pc , they protect their business, being reletlessly Tbagged by players you can't win against isn't very fun, pc is a small market compared to the 50+ million console players. so it is the obvious and easy thing to do.
Replies: >>105698140
Anonymous
6/25/2025, 10:44:40 AM No.105698098
>>105698051
Name it. You know you want to, but someone won't let you do it.
Replies: >>105698149
Anonymous
6/25/2025, 10:46:17 AM No.105698108
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26906

>Still not merged after a year
Anonymous
6/25/2025, 10:46:37 AM No.105698112
1750841176
1750841176
md5: 1fd972491dd8f7b3bbff1059c74a75c5🔍
wayland on suicide watch
Replies: >>105698143 >>105698170
Anonymous
6/25/2025, 10:51:59 AM No.105698140
>>105698091
With that game in particular. They were doing it because they had a deal with Sony to use their consoles at events. In the prior version of the game all the events used Xbox 360s for years. Since the 360 version of the last game was the only one that didn't have random input lag. You could do 2-frame combos if you were skilled enough on 360. On the PS3 version they would randomly drop half of the time because it was programmed like shit. This was embarrassing for them because it meant they had to use an American console at all the Japanese events for years. They tried to force use of the PS3 one time but the players said they'd boycott any events going forward if they attempted it again.

It would have been different if they put the PC version on an even playing field with the PS4. But they were actively attempting to fuck over PC players that had dropped $120+ on the game. The rootkit was the final slap in the face and a lot of people refuse to buy anything that company produces even today.

I can't remember the exact details now but the netcode was fucking awful. I'll give you an example; At the time I was playing the game on a PC with a GTX980 and using a 144hz monitor. I had a 1Gbps fiber connection with 5-10ms ping times to players within my region. I was on properly configured network (ports open etc.). I used a joystick with a low latency PCB.

Basically, the better your connection+display was the more you got punished. I couldn't play with 90+% of players in online play. Since my connection was better than theirs and a lot of them on console were using wireless as it was the default for them. Even if I limited player pool to PC only I usually got shafted by the netcode.

I had to artificially cause lag on my end to even play the game with another player. That's how bad it was.

Any other game I played. Including other crossplay games were fine. Even with 100+ms ping times to the other player.
Anonymous
6/25/2025, 10:52:12 AM No.105698143
whores
whores
md5: c366e283c11ff4078258da35ceab5830🔍
>>105698112
>AUR
Anonymous
6/25/2025, 10:53:00 AM No.105698149
>>105698098
Left 4 Dead 1.
Anonymous
6/25/2025, 10:56:52 AM No.105698170
>>105698112
I will celebrate when it's on main repo.
Replies: >>105698176
Anonymous
6/25/2025, 10:57:06 AM No.105698172
so here's the tldr

redhat did not want to deal with xorg no more and decided to make something from scratch tailored for their product needs ie tight security

ok no problem but it became a problem when their employees started sabotaging xorg for everyone else and pushing their niche solution (wayland) into everyone's throats

now it's obvious anyone and everyone who spent time re/writing software for wayland did it for free to append to corporate product and was a waste of time and effort because xorg can still exist
Anonymous
6/25/2025, 10:58:19 AM No.105698176
>>105698170
>I will celebrate when chinaman makes it for me.
Anonymous
6/25/2025, 12:48:12 PM No.105698794
>>105697620
I was always curious about this.
On Wayland, the cursor does not drift away from the underlying action. Either the cursor lags more than on Xorg, or the compositor miraculously reacts instantly to mouse movement. I suspect the former is true, but I was never able to measure it.
Anonymous
6/25/2025, 12:53:55 PM No.105698819
>>105697620
you forgot to tell everyone that cursor plane updates going to 0 also tanked your performance of you were using VRR. Software cursor is less problematic overall.
Replies: >>105698839
Anonymous
6/25/2025, 12:56:39 PM No.105698839
>>105698819
I noticed this. Is this preventable, or are hardware cursors inherently broken on Wayland? I don't want to regress to fucking software cursors.
Replies: >>105700048
Anonymous
6/25/2025, 2:38:14 PM No.105699486
>>105696020
claims to have a model that works more like modern hardware
and they do this lmao
Anonymous
6/25/2025, 3:52:40 PM No.105700048
>>105698839
>are hardware cursors inherently broken on Wayland?
correct.
Anonymous
6/25/2025, 4:02:15 PM No.105700131
i'd much rather have tearing than higher latency. why are devs so insistent that i don't get to choose?
Anonymous
6/25/2025, 4:08:56 PM No.105700184
>>105697905
every time i point it out, i get 10 redditor-types saying that adding double-digit milliseconds of latency (on top of an already laggy hardware and software path) is imperceptible. show them research that found otherwise, and they just ignore it.