I'm not aware of any that are Linux only, but most everything has a Linux app these days. GPUGrid and FAH will push your cards to their limit. I'm going to have to look at the Windoze machine at some point, one of the GPUs is hitting 80-81 on GPUGrid which is causing it to throttle. I'm probably losing 10-15% efficiency.My Dell (GTX 760) machine was crashing at an ever increasing frequency. Eliminate 90% of possible hardware causes. Only MB, CPU and OS left. So I switched back to Mint to eliminate the OS possibility. Now able to run GAIA.
Are there any NVIDIA GPU science projects for Linux only?
Hope eventually to replace the Dell quad core with one of the Xeon 6/12's I'm working on.
Giving my machines a break from FAH. I had not noticed until now that other BOINC GPU projects are running a little cooler than FAH.
Folding@Home I'm Finally Back!
- Thread starter Nick Name
- Start date
Decided to run GPUGRID last night. Same here, 80-81C. Have to relearn Linux and how to load coolbits again. Argh.
But, Dell ran without a hitch on Mint. Hoping that means previous glitch was OS. I'll give it more time. Hopefully I can use the Win key on one of my other projects.
But, Dell ran without a hitch on Mint. Hoping that means previous glitch was OS. I'll give it more time. Hopefully I can use the Win key on one of my other projects.
I don't know why the basic Nvidia fan control on Linux is so bad, especially since it's the compute card of choice and compute is much more demanding than gaming loads. The fans don't run nearly fast enough to handle the heat without the coolbits tweaking you mention. It's really aggravating.
Upgrading to the latest client plus tweaking the router a bit seem to have mostly fixed the problems I was having. Output has been pretty steady, around 5 mil/day for the last week or so. There are bound to be occasional hiccups, server load is still high and new work has to be queued etc. For now I'm happy with how things are running. I'd like to upgrade the AMD card new a new Nvidia card but as we know the newer GPUs are impossible to find at a reasonable price, if at all.
That almost makes me want to sell my VII! GPU pricing is crazy right now.
I gave up on the Einstein project and I'm back to FAH again (I guess I'm more interested in health project...)!
We've advanced more than one hundred positions in less than a month (now at 808th). Which was our best place ever, @NickName?
It seems Doneske is still with us!
We've advanced more than one hundred positions in less than a month (now at 808th). Which was our best place ever, @NickName?
It seems Doneske is still with us!
Yes. If you scroll to the bottom of our Extreme OC stats page it lists the top members. We only have 9 active folders right now. Thanks to those who kept folding during the summer.
The FAH ecosystem isn't nearly as robust as BOINC, it's almost like the stats were added as an afterthought. There's really no team control other than creation, there's no team communication like BOINC has. If it weren't for the gracious EOC stats I wouldn't even be able to see who's active on the team now. It's just as well there aren't any challenges, we aren't competitive here...yet.
I don't know what our highest rank was, I guess I should have kept track. There doesn't seem to be granular history like that available. I know we were higher than we are now.I gave up on the Einstein project and I'm back to FAH again (I guess I'm more interested in health project...)!
We've advanced more than one hundred positions in less than a month (now at 808th). Which was our best place ever, @NickName?
It seems Doneske is still with us!
I see my daily average has been declining, I think the client is hanging now and then and not getting new work.
Have you kept the switch in the 'switch' "I support research finding" in the default position ("Any disease")?
I think it's the best way to ensure your client will always have work to do.
I think it's the best way to ensure your client will always have work to do.
I've accessed FAH fórum today and saw this message (seems to be good news):
Welcome to the Folding Forum! Before you start, there are a few things we need to tell you:
We thank you for your support of the project!
Please remember, before starting a new thread/topic, use the search function (top right).
The community members that support this forum are volunteers like you, and like the servers, we are being overwhelmed by demand, so please keep that in mind.
Welcome to the Folding Forum! Before you start, there are a few things we need to tell you:
- The Folding@home team know about the work unit shortage.
- It's happening because of an approximately 20x increase in demand.
- The Folding@home team are working on it and hope to have a solution very soon.
- Keep your machines running, they will eventually fold on their own.
We thank you for your support of the project!
Please remember, before starting a new thread/topic, use the search function (top right).
The community members that support this forum are volunteers like you, and like the servers, we are being overwhelmed by demand, so please keep that in mind.
I had a couple of machines that were limited on disk space so I signed them up for FAH after being able to find a FAHControl for Python3. The FAHControl from Stanford was built for Python2 and would have dependency issues during install. Don't know why they recommend using the --nodeps flag as then the program installs but is unusable due to missing modules. The FAHControl for Python3 installed without any issues and works quite well. These systems are desktop type systems that I upgraded to the new Ryzen 3900 processor. Once the RX6800 XT becomes available I will probably get those and shutdown 2 more servers. There have been issues getting work lately and I will notice that I have about 50% of the clients waiting on download at times. I have also noticed that even with all clients running, the PPD calculation has dropped by about 25%. They must have changed some scoring for some of the new projects they added.
UPDATE: 6 of 9 clients now waiting for work. These are all CPU clients
UPDATE: 6 of 9 clients now waiting for work. These are all CPU clients
Last edited:
My preference is set to Alzheimer's, but I am getting mostly COVID work. The project preference is only slightly respected, if at all.
This message has been up for months, nothing new there.
A declining average could also mean lower credit work. As I've said I have BOINC set to run on that GPU if there's no FAH work running, specfically PrimeGrid ppsieve since those run quickly. I can see a 24 hour history in BOINCTasks, so I can tell if the FAH client has been idle for a substantial amount of time. I know it was hung earlier today because I saw it waiting to up/download, but I don't have any completed tasks in my BT history so it wasn't stopped long enough for any PrimeGrid work to complete. I can't comment on the FAH CPU work as I haven't tried that.
