#RAM issue

1 messages ยท Page 1 of 1 (latest)

granite storm
south wagon
#

Hello!
Thank you for your report!
Does the issue repeat every time when you summon your ship?

We are aware that our game is quite resource-have and constantly working on optimization of both ram and gpu processing :(

granite storm
#

no no no !! first time i've got a crash !
in the log i've got a warning cause my ram is high, but i dont monitoring my system ram before the crash.
And btw, for a UE5 game, beside the writing disk issue, you game is "light" ๐Ÿ˜‰
i'll monitore my system ram to see if it's happen again โค๏ธ

south wagon
#

Thanks! We are collecting all the reports to track data, so your is helpful too :)

granite storm
#

it's seems it's a community website "gaming tool" how eat my ram more than the game... well, i'll close all thing and see if it's ok now, lol

#

sorry if it's a false report

#

after world load, let's see in a couple of hours o_o

granite storm
#

so.. not bad ๐Ÿ˜‰

granite storm
#

the only "leak" i can force, it's to fast travel 6 time in a row ๐Ÿ˜‰

granite storm
granite storm
#

After more testing, this does not look like a constant RAM leak.

Memory usage is mostly stable during normal gameplay, but repeated fast travel / world transitions / heavy loading can push it to a higher plateau, and the game does not seem to release it quickly enough afterward.

So it looks more like memory accumulation or slow cleanup after transitions than a permanent leak.

so, for now, ticket close ๐Ÿ˜‰

granite storm
#

after all my crash, and analysing them:
This does not look like a pure random crash to me.

It looks more like a transition-related runtime issue: repeated reload / fast travel / re-entry seems to recreate some rendering resources (DLSS / cloth / async-loaded content), while cleanup is delayed or incomplete.

Memory then climbs to a higher plateau instead of dropping back down, and after enough repetitions the runtime seems to end in a GPU-side D3D12 crash.

So this feels closer to a resource lifecycle / cleanup issue than to a simple hardware instability.
โค๏ธ

#

so, maybe the GPU-side crash with "VRAM issue" on lower GPU can be cause by that thing... and to go deeper, this is on you side :p

granite storm
#

Fast travel pattern captured: memory usage increases in a step-like way during transitions and then stabilizes at a higher plateau instead of returning to the previous baseline.
(yep its a mess xD)

granite storm
#

Confirmed with runtime monitoring: after fast travel, the game keeps a higher RAM/VRAM plateau even during idle time in base.

So at least on my side, resources loaded during transitions do not appear to be released properly, or not aggressively enough, once the transition is over.

granite storm
#

much better !
(gpu_usage seems off, i log the 3D side not the compute ><)

vivid fulcrum
#

Thanks for the detailed update! We'll look into it.

granite storm
#

i'm good for tonight โค๏ธ , script ready ahah

granite storm
#

ok, normal session is clean, let's see if this change in the next session ๐Ÿ‘Œ

granite storm
#

i've got a little freeze 500ms/1sec (guess) and here we can see the game free up RAM, pagefile goes up.
btw, upgrade the page file to 32GB from 8GB remove OOM ๐Ÿ˜ฎ โค๏ธ

#

DB write / transaction latency + PSO waits + DLSS recreate can create a severe runtime/render stall.

That stall may not be a GPU crash by itself, but it can push the D3D12 pipeline into an unstable state, which may later surface as a device removed / driver-like crash.

1 GB/s write spike on disk when the freeze happen ๐Ÿ˜ฎ

granite storm
#

fine today beside the freeze โค๏ธ

granite storm
#

Little session after update