Folding Forum

Hello All,

I use a laptop with the GPU Skylake GT2 (Intel) HD Graphics 520. This GPU is not supported so far.
Are there plans to support this GPU?

Thank you in advance,
Achim

Welcome to Folding@Home!

Currently no Intel GPUs are supported by F@H, although there is a zero point beta group of testers working on the topic.

https://www.techpowerup.com/gpu-specs/h ... -520.c2783

Your CPU is not going to be very powerful when/if it becomes supported. It appears to be about 1/3 as fast as an Nvidia GT 1030, about the slowest GPU that finishes all Work Units. 1/2 as fast as a GT 730, which finishes some WUs, but not all.

Unless some researcher has data they are in no hurry to get results on, I do not think it will ever fold. (I am just a folder like you, not an expert)

Integrated GPUs are adequate to display a desktop and not much else. Thy're simply too slow to use as a compute engine for a FAHCore. As was said above, they may never be supported. Gaming (discrete) GPUs are many, many times faster and speed is an important factor for FAH.

Since we are on the subject, are the newer Intel Xe graphics supported? I mean in a beta testing kinda way.

Thank you for the information,

Warm regards and stay healthy,
Achim

iero wrote:Since we are on the subject, are the newer Intel Xe graphics supported? I mean in a beta testing kinda way.

I have no real information, but my guess is that the 'wimpy' Intel laptop GPUs will never be supported, but they are using them to test with so when Intel releases discrete GPUs, F@H is ready to fold on them.

YMMV.

FAH decides on whether to support new hardware when donors report they actually have the product in their system. We don't work with pre-release hardware.

At this point only the Gen9 Intel iGPUs are semi-supported for beta test purposes - https://en.wikipedia.org/wiki/List_of_I ... units#Gen9. As I understand it, this is a combination of sufficient hardware features and mature driver support.

The Intel Xe graphics may get support at some point in the future, but I have not seen anything on that yet. It does use a new instruction set, so may need support written for it and added to the OpenMM codebase.