Replies: 1 comment
-
No, it didn't improve speed. To perform top speed, GPU needs to have large data for calculation. When a task is too light, GPU may finish the task before speed-up. After GPU has enough task to become full speed, more data would not be required. I limited VRAM usage upto 25%. Then, it uses 4 GB on a GPU with 16 GB VRAM. The 25% limit is came from NVIDIA OpenCL driver's restriction. It allocates 25% of VRAM size. Though other (AMD or Intel) drivers support more allocation, using more VRAM doesn't improve speed. If a fast GPU has very small VRAM, this 25% limit may be bad. But, normally fast GPU seems to have enough large VRAM nowadays. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
Been using your app for many years. Thank you for continuing to make it better!
I monitored my GPU memory usage during PAR archive creation. The maximum runtime usage is approximately ~4GB but my GPU card has 16GB of onboard memory.
Would using more of the onboard GPU memory help to speed up the calculations even more?
Just an idea!
Beta Was this translation helpful? Give feedback.
All reactions