My laptop has limited RAM (only 16 GB) a really crappy nVidia GPU with 4GB of VRAM and because the Intel APU is faster on most benchmarks I never choose to run on the discrete GPU, so the idea is to make some use of the otherwise wasted HW…
Any reason as to why a discrete Nvidia GPU of capacity as large as 4GB underperforms the APU? Could you confirm if you are using the proprietary drivers?
Hi toxic coder, indeed I do not want to use proprietary drivers from nvidia, so yes, currently the open drivers for the intel APU is much faster and supports higher OpenGL levels & features than the nuvoueueou driver… look at the glxinfo output:
nVidia
Vendor: nouveau (0x10de)
Device: NV117 (0x139b)
Version: 20.2.1
Accelerated: yes
Video memory: 4095MB
Unified memory: no
Max core profile version: 4.3
Intel
Vendor: Intel (0x8086)
Device: Mesa Intel(R) HD Graphics 530 (SKL GT2) (0x191b)
Version: 20.2.1
Accelerated: yes
Video memory: 3072MB
Unified memory: yes
Max core profile version: 4.6
Thanks, but I am aware that I could install proprietary drivers from nVidia, however the point here is, I do not even want to play hi-end games on my laptop, but could use the 4GB additional RAM for my apps or accelerating my swap which is on my disk at the moment…
Good news, ArchWiki to the rescue: https://wiki.archlinux.org/index.php/Swap_on_video_RAM
You’ll have to check whether you have to use MTD or vramfs can run on POCL or, with recent improvements to OpenCL on nouveau, mesa’s Clover.