Core 22 version 0.0.18 observation, possible iGPU conflict

If you think it might be a driver problem, see viewforum.php?f=79

Moderators: Site Moderators, FAHC Science Team

Post Reply
BobWilliams757
Posts: 493
Joined: Fri Apr 03, 2020 2:22 pm
Hardware configuration: ASRock X370M PRO4
Ryzen 2400G APU
16 GB DDR4-3200
MSI GTX 1660 Super Gaming X

Core 22 version 0.0.18 observation, possible iGPU conflict

Post by BobWilliams757 »

I just wanted to post this in the event any other AMD iGPU users observe this conflict. This is a one time event at this point, but I'll try to duplicate with other GPU and CPU projects and post findings.

At times I will fold on both GPU and CPU on my Ryzen 2400G. Due to architecture and shared resources the end result is less than the sum of the CPU and GPU folding individually. So a CPU unit that returns 50k on average folding with a CPU unit that returns 70K on average might fold simultaneously and return only 90-100k points. This varies somewhat by project, and at times the sum of points (progress) might be higher or lower, but always more than just folding one WU individually.

I happened to notice the new core download and install the other day, and folded a few GPU work units with no changes in performance that I could notice. This is the norm so far with this system, slow but stable and cores, system updates, drivers, etc.... nothing seems to impact it much. So it's GPU folding, and towards the evening I unpaused the CPU slot so it would pick up a CPU work unit. Often the PPD will rise up near the sum of the two run individually, then settle lower as the system finds how it's going to share the resources and such. But in this case as it settled, it reached the point of slowing below what even the GPU alone would produce. It was still slowly moving down, so I checked all the system basics to make sure there was no machine problem. Nothing found, I finally just paused the GPU to let the CPU project progress quicker. There was plenty of time until the GPU timeout, so it seemed the quickest way to get both projects completed.

In this case the GPU work unit was 18201 (47483, 0, 20) and the CPU work unit was 18210 (11893, 0, 31). Neither of these projects has ever created any unusual behavior when run individually on this machine. I couldn't find any other instances of both of them running at once in my logs though.


I'm going to observe going forward and see if this slowdown duplicates with other WU combinations. I'm not certain it's a Core update related issue, but it's never happened at all over my history of folding on this machine.



Just a heads up to the iGPU users. These things are slow enough as is, we don't want to risk slowing them even more!
Fold them if you get them!
Post Reply