Lazer gameplay stutter

[ciinTri-alt]

Type

Performance

Bug description

I have been dealing with this problem for more than 2 years now. This seems very similar to #26998, apparently it got fixed for OP, but not for my end.

These stutters often happen in the first 1/3rd of maps, doesn't always happen, but it happens about 5 out of 10 times. I noticed that the stutters will more likely appear on the first map to play after alt tabbing back into the game.

The stutter occurs at the 24 second mark in the video I attached, showing processes in Task Manager as well.
https://files.catbox.moe/lq3oe6.mp4

PC Specs:
Processor: Ryzen 7 5700X
GPU: RTX 4060
RAM: TEAMGROUP T-Force Dark Za 32GB (2x16GB) 3600MT/s
MB: ASUS ROG Strix B550-A Gaming
Storage: 1TB Samsung 980 Pro
PSU: Corsair RM650 (2023) 80+ Gold

Screenshots or videos

Stutter occurs at 24 second mark, I also showed what processes are running in the background with Task Manager.

https://files.catbox.moe/lq3oe6.mp4

Version

2025.306.0-lazer

Logs

1741518525.nauth.log
1741518524.network.log
1741518524.database.log
1741518524.performance.log
1741518524.auth.log
1741518524.runtime.log

compressed-logs.zip

[smoogipoo]

It looks like a full compacting GC happened. I haven't experienced this issue myself, but here's a few things you could try:

• Use the OpenGL renderer instead of D3D11/Deferred.
• Try a different skin, or at least try removing elements like the pp counter and the judgement counters on the side.
• When a stutter happens, pause the game and check the memory usage. You showed task manager sitting at ~550MB in song select, which is fine, but if you're seeing a large disparity during gameplay then there could be cause for concern.

Also note that having the frame statistics graphs visible will exacerbate any problems (they're not zero overhead).

[ciinTri-alt]

It seems the stutters are alleviated when all elements are removed, it took 50 minutes for the first stutter to appear with Deferred D3D11. Additionally, the stutters were noticeably shorter. Tried the argon pro (2022) skin, the stutters remained, but less severe when no elements were present. Also checked the memory usage after a stutter happens, there doesn't seem to be a disparity. I will update you probably in 1 week for testing without elements on other renderers. Although I don't think I will stick to anything other than the Deferred D3D11, it is the most responsive out of all the renderers. A piece of info that may help since I wanna help as much as I can, I don't think RAM Ranks matter in this regard (maybe even DRAM Manufacturers?), I was previously using [Dual Rank Sticks](https://www.kingston.com/datasheets/KF432C16BB1_16.pdf) (Nanya), currently on [Single Rank Sticks](https://www.teamgroupinc.com/en/product-detail/memory/T-FORCE/dark-z-alpha-ddr4/dark-z-alpha-ddr4-TDZAD432G3600HC18JDC01/) (SK Hynix). There is basically no difference in performance, the stutters are present on both of these RAM modules.

[peppy]

If you can narrow it down to a specific skin element, that would be beneficial to us.

[ciinTri-alt]

So I took my time trying to locate the issue throughout the time I made this thread, unfortunately I couldn't pinpoint the cause.

But here are the troubleshooting steps I took that didn’t help:

• Several methods of resetting windows (Keep my files, Remove Everything, Boot Drive Secure Erase and a full clean install of Windows via USB drive)
• Switching to a different GPU (RTX 3060)
• Ensuring CPU, GPU, Motherboard, PCIE Devices drivers/BIOS are up to date
• Connecting only 1 Monitor
• Minimizing Peripherals connected to my PC, and switching to different peripherals, tried with Mouse and Keyboard only as well
• Disconnecting my PCIE WIFI + Bluetooth card (ASUS PCE-AX3000)
• Disconnecting Front Panel Connectors
• Disabling High Definition Audio controller that's connected to the GPU
• Disabling AMD FTPM in BIOS
• Uninstalling as much bloatware in Windows
• Disabling Windows GameBar
• Ensuring all components aren't overheating
• Ensuring there are no bad wiring/electricity in my house

I was still able to replicate the issue with, and without configuring any Windows setting. On top of all this, it seems osu! stable doesn't have this stuttering issue. I apologize that It took me some time to give an update on this thread as I wanted to take a drastic measure to make sure I've done as much as I can on my end.

peppy: If you can narrow it down to a specific skin element, that would be beneficial to us.

As for this, I took longer sample sizes of sessions to make sure. There actually is no change, issue persists even without any skin element present.

If you have any suggestions, please do. I am still willing to cooperate. But for now, I am sticking to osu! stable.

[pigeonskew]

Try playing on Borderless instead of Fullscreen/Windowed. It solved latency issues for someone. Might deal with the stutters you're experiencing.

[ciinTri-alt]

Try playing on Borderless instead of Fullscreen/Windowed. It solved latency issues for someone. Might deal with the stutters you're experiencing.

None of the 3 screen modes helped unfortunately, I actually replicated the stutter in Borderless in the video link I posted above.

https://files.catbox.moe/lq3oe6.mp4

[ciinTri-alt]

Fix TLDR: Signing in, then restarting the game fixes the stuttering issue.

So I've been playing the game signed in for the past 3 days, and I am yet to encounter the stuttering issue. Basically, I had to sign in, then restart the game to let osu!Lazer update 18 beatmaps. Is it possible for those 18 beatmaps to be causing the stutters? I have Standard Beatmap Packs 200-1627 installed, and the stutters were recurring way too frequently in all of my previous testings when I wasn't playing signed in. I also signed into the osu!Stable client, but I can't be certain whether that had a factor, as there was no update in osu!Stable after logging in and restarting. I haven't confirmed if the stutters will occur again after logging out, but I'll probably give an update in a couple of days if I'm not busy anymore.

By the time logged in and I restarted Lazer, only 18 beatmaps got updated in the notification area. Would there happen to be some under the hood updates being pushed out exclusively when logged in?

I've also done a recording session for about an hour and a half, showing no signs of the stutter issue mentioned in this thread. Tested with both Direct3D11 and Direct3D11 (Experimental) Fullscreen/Borderless modes, as well as showing the Windows Gaming Settings at the end. There should be timestamps ready on the video, otherwise check the video description. https://youtu.be/Aij3goJvjjw

Logs:
compressed-logs.zip

[ciinTri-alt]

I was able to replicate the stutters 3 times in just under 15 minutes of gameplay by playing when signed out. So it seems like being signed out is really a part of the reason of the stutters.

Here's another log file. Logs created at 2025-05-28 should be where to look at.
compressed-logs.zip

[peppy]

my guess is different GC pattern (less cycling) due to lower load, leading to larger collections.

we are looking to change the GC engine which will alleviate this.

[peppy]

@ciinTri-alt you could test again on tachyon release stream to see if things are any better.

[ciinTri-alt]

Things definitely improved, the stutters don't occur too often and are not too disruptive to the gameplay on Tachyon. However, I did notice a behavior depending on the skin, and the renderer being used. I recorded this test that you can access here, it also includes timestamps in a text file, and the OBS log just incase you wanted to see the render/encoding lag numbers. https://mega.nz/folder/fxBj3KLb#JJal3AA2EBwWxfv5sUFuSg

Direct3D 11 Experimental was seeing stutters on both Tri-gon and Triangles skin, but Tri-gon skin was seeing significantly less stutters. I've played plenty of short sample sizes with Direct3D 11 Experimental on the Tri-gon skin without recording for the past few days, the results are basically the same, the stutters are rare. There was one instance of stutter in the recording, but I'm not sure whether it was visible in the recording.

Direct3D 11 seemed to remain stutter free on my personal Tri-gon skin? at least on the 30 minute sample size, I will give this a bigger sample size and give an update as soon as I can. On the other hand, Triangles skin + hitlighting and beatmap colours disabled is seeing the same stutters as it would in Direct3D11 Experimental. I'm not certain if enabling hitlighting and beatmap colours caused some stutters.

Giving an update to this thread totally slipped in my mind, I apologize for that! I often couldn't find enough time and energy to give a reasonable sample size for proper testings.

Logs:
compressed-logs.zip

[peppy]

Please do not use experimental renderers.

[pigeonskew]

Just wanna point out that this also happens on my laptop even on Tachyon (2025.702.0-tachyon)

Specs:
AMD Ryzen 3 5300U with Radeon Graphics (Vega 6 i think)
8GB RAM
Zorin OS 17.3

Just using OpenGL.

Here is a video showing the stutters (visible on the recording):
https://streamable.com/t5x36r

Kinda ruins playing taiko for me. On standard, it's not noticeable.

[BaroqueViola]

Hi, I might be experiencing the same class of stutter, but my ETW trace shows a non-compacting Gen 1 collection instead of a full compacting GC. This may be related to #26998 as mentioned earlier in the discussion.

Description

I get visual stutters, which happens at random and not at a reproducible part of a beatmap. It occurs during both live gameplay and replay playback. It occurs about once a minute, but it seems less frequent on my old laptop.

Audio does not audibly stutter. Stutters still occurs even when I don't make mouse or keyboard inputs, and disconnecting peripherals did not fix the issue.

In the attached screenshots, you can see that there are spikes in Input, Audio, Update, and Draw at the same time. They only appear in different positions because they run at different frame rates (480/960/1000 fps). At the stutter, there are green and yellow pixels. A red pixel also rarely displays.

System information

• osu! version: 2026.804.2-lazer
• Windows 11 Home 25H2, build 26200.9168
• CPU: Intel Core Ultra X7 358H
• GPU: Intel Arc B390
• Graphics driver: Intel 32.0.101.8991
• Display: 3120 x 2080 at 120 Hz
• Mouse polling rate: 1000 Hz

Old laptop:

• osu! version: 2026.804.2-lazer
• Windows 11 Pro 25H2, build 26200.9278
• CPU: Intel Core i7 1165G7
• GPU: Intel Iris Xe Graphics G7 (96 EU)
• Graphics driver: Intel 32.0.101.7088
• Display: 2560 x 1600 at 60 Hz

I tried the following, which didn't solve the issue:

• Switching to the default skin
• Closing Discord and other overlay apps
• Closing Xiaomi background/helper apps
• Changing the FPS limit to 2x refresh rate or VSync
• Testing OpenGL, Direct3D 11, and Direct3D 11 Experimental (Vuklan is unavailable)
• Changing the display's refresh rate to 60 Hz
• Testing SDL2 with OSU_SDL3=0

Diagnosis

I captured a stutter using PerfView with Thread Time enabled. Nothing much happens at first. When I noticed the stutter, I quickly paused the replay, alt tabbed, then stopped PerfView. The trace says the stutter happens at ~58.95 s, and gameplay activity dropped at ~60.1 s (when I paused and alt tabbed).

Findings from the trace:

• Serveral high-frequency osu! threads experienced simulatenous ready-to-run delays of ~14.3-15.0 ms.
• Serveral other priority-15 osu! threads executed continuously for ~10.9-15.7 ms.
• SuspendForGCPrep began at 58.965 s.
• A non-concurrent Gen 1 collection began at 58.967 s.
• PerfView reports 2.582 ms for the GC pause.
• The system had plenty of idle CPU capacity during that time. No other app was using significant CPU time.

I am not sure if the collection causes the scheduling disruption, or it just occurs as part of the same event. I included these values to help locate the stutter in the trace.

Attachments

They can be found on MEGA

• Compressed osu! logs
• Frame time graph screenshots from both new and old laptops
• PerfView Thread Time ETW trace