Why isn't my GPU using all dedicated memory before using shared memory?

Question

According to Task Manager, I'm only using <6 GB of my GTX 1070's 8GB of dedicated GPU memory before starting to use shared memory. IIUC, using shared memory imposes a heavy performance penalty, as system RAM is far slower than dedicated GPU RAM.

Why is this happening, and how can I make the GPU use all the available dedicated GPU memory?

Solution

Shared memory does indeed have a heavy performance hit compared to the actual GPU memory, but you are using only 0.2GB with 5.8GB in use on your GPU. That is approximately 3.5% of the total memory use.

It is obvious that the vast majority of the work is being done on your graphics card and it is entirely possible that the only reason that shared memory is in use is either as a pass-through buffer, or a shared area for the CPU to do work (such as physics, shader caching or other tasks) before passing on to the GPU. Using shared memory is not an immediate failure. It just means that there are some jobs that need the CPU to do some work as well as the GPU.

If shared memory was up at 6GB as well as or instead of the dedicated GPU then you might have cause to be concerned, but at such a tiny amount it is not doing any significant amount of work in that area.

Comments

[gronostaj]

Wild guess: your video outputs are physically connected to integrated GPU and the dedicated GPU is exposing its rendered frames through RAM. (I don't know if that makes any sense)

[Hannu]

GENERALLY: "Shared memory" = Available to both GPU and CPU. Must be used to transfer data to/from the GPU; i.e. CPU writes data into shared mem to make it available for the GPU.

[Community]

As it’s currently written, your answer is unclear. Please edit to add additional details that will help others understand how this addresses the question asked. You can find more information on how to write good answers in the help center.