Jump to content

Is it possible to dedicate more gpu memory?

srrs234
Go to solution Solved by da na,

The dedicated is how much VRAM is on your card. Shared is how much of your GPU VRAM can spill over into your ram once your VRAM is full. GPU memory is the sum of the two.

My graphics card is a 2060 6gb, but in task manager is says it has 21.9gb available and says only 6 are dedicated. Is it possible to dedicate more of the memory?

gpu.png

Link to comment
Share on other sites

Link to post
Share on other sites

The dedicated is how much VRAM is on your card. Shared is how much of your GPU VRAM can spill over into your ram once your VRAM is full. GPU memory is the sum of the two.

Fan of workstation/server hardware and old stuff that just works.

Top LTT Forum reaction giver/recipient 2022

Link to comment
Share on other sites

Link to post
Share on other sites

4 minutes ago, srrs234 said:

My graphics card is a 2060 6gb, but in task manager is says it has 21.9gb available and says only 6 are dedicated. Is it possible to dedicate more of the memory?

gpu.png

Huh 馃槷

Your GPU has 6GB VRAM, that can only be used as VRAM so it's "dedicated", and there's no way to add more unless soldering it yourself...聽

But it *can* use your "CPU" RAM as well, that's why you see more "shared" RAM available. But RAM is way slower than VRAM so when it uses it it stutters and give low FPS

System : AMD R9 5900X聽/ X570 AORUS PRO/ 2x16GB Corsair Vengeance 3600CL18聽/聽ASUS TUF Gaming AMD Radeon RX 7900 XTX OC Edition GPU/ Phanteks P600S case /聽聽Eisbaer 280mm AIO聽(with 2xArctic P14 fans)/聽0.5TB Sabrent Rocket +聽2TB WD SN850 NVme Gen4 + 4TB Toshiba X300 HDD聽drives/ Corsair RM850x PSU/ 34" 120Hz聽3440x1440p Alienware AW3420DW monitor / Logitech G915TKL keyboard (wireless) / Logitech G PRO X Superlight聽mouse

Link to comment
Share on other sites

Link to post
Share on other sites

3 hours ago, da na said:

The dedicated is how much VRAM is on your card. Shared is how much of your GPU VRAM can spill over into your ram once your VRAM is full. GPU memory is the sum of the two.

But not all games or apps that tap a GPU can use shared GPU memory, right?

I think AI stuff like Stable Diffusion or local chat models require dedicated VRAM, because I had some 12GB models crash on my 3080 10GB despite having tons of shared memory (given my 32 GB RAM).

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now