I have a HP Pavilion laptop with an integrated Intel Iris GPU, as well as a dedicated Nvidia Geforce MX 450. The Nvidia has a dedicated 2 GB VRAM, and the the two cards share an additional 16 GB of memory. Unreal Editor only detects the 2 GB, and claims being overbudget while using just 3 GB of video memory. This makes the editor borderline unusable, with at most 10 FPS. The computer has an Intel Core i5 with 32 GB RAM, so these are not an issue. Is there any way to make UE5 use all available resources?
Add a comment
|
1 Answer
\$\begingroup\$
\$\endgroup\$
3
"Shared" video RAM is usually most relevant for integrated GPUs, which don't have their own dedicated memory. Ideally, your dedicated GPU should not need to access shared memory, because shared memory is very slow and terrible for performance. In particular, the MX 450 doesn't support shared memory (refer to spec table here).
I haven't worked with Unreal, but it sounds like the engine is saying that you are using all 2GB of dedicated VRAM + 1GB of additional memory that is stored in system RAM and probably has to be swapped back and forth between system RAM and the GPU. You might be able to fix this issue by adjusting your texture streaming settings; a web search for 'unreal memory overbudget' seems to return many relevant results. However, for the best long-term results, you should probably start saving up for a new laptop. The MX450 is an entry-level GPU from two years ago and 2GB of VRAM is pretty skimpy by today's standards.
-
\$\begingroup\$ This is admittedly a weak answer, but seemed a little too long to leave as a comment. \$\endgroup\$– KevinAug 2, 2022 at 1:08
-
1\$\begingroup\$ Added a reference to strengthen your initial claim. The way I read it, UE is using 3gb of ram placing it overbudget because only 2gb was available from the Nvidia card. \$\endgroup\$– PikalekAug 2, 2022 at 2:57
-