Distributive computer specs

[jamesdsimone]

I did a quick search on folding at home and the specs they list look hopelessly outdated. So the questions are how many cores do these projects scale to? Can they use 2 processors? How much ram do you need? Do you need a GPU?

 

[mmonnin03]

Folding @ Home is mainly a GPU driven project. The credit done by GPUs is far, far higher than any CPU.

I believe FAH will scale to all of today's desktop CPUs. If not then 2 tasks can be run at once. Per watt though, go with as much Nvidia GPU you can afford.

 

[jamesdsimone]

Folding @ Home is mainly a GPU driven project. The credit done by GPUs is far, far higher than any CPU.

I believe FAH will scale to all of today's desktop CPUs. If not then 2 tasks can be run at once. Per watt though, go with as much Nvidia GPU you can afford.

Are there projects that are better suited to CPU's? I already have the CPU's but no plans on buying a GPU just for this.

 

[Markfw]

Are there projects that are better suited to CPU's? I already have the CPU's but no plans on buying a GPU just for this.

none of the F@H projects works well on CPUs. You are better off with Rosetta@home.

 

[TennesseeTony]

If you are not too worried about points, but want to contribute instead purely for the science being done, Folding@home does need your CPU cycles, as not all their research can be done via GPU.

There are many other projects to contribute to as well, using BOINC. https://boinc.berkeley.edu/

If you need help setting up either F@H or BOINC, just ask away.

 

[Markfw]

If you are not too worried about points, but want to contribute instead purely for the science being done, Folding@home does need your CPU cycles, as not all their research can be done via GPU.

There are many other projects to contribute to as well, using BOINC. https://boinc.berkeley.edu/

If you need help setting up either F@H or BOINC, just ask away.

Yes, when I say Rosetta@home, I assume people are adding it via BOINC. I don't even know how to do it without BOINC.

 

[jamesdsimone]

If you are not too worried about points, but want to contribute instead purely for the science being done, Folding@home does need your CPU cycles, as not all their research can be done via GPU.

There are many other projects to contribute to as well, using BOINC. https://boinc.berkeley.edu/

If you need help setting up either F@H or BOINC, just ask away.

I just got an HP z840. It looks to be as new. The plastic film has not even been pulled off the sides of the case. It came stock with 2 E5-2620 V3 but I might swap them for 2 E5-2690 V4's. So would be interested in something that could best utilize 28 cpu cores.

 

[Skillz]

I just got an HP z840. It looks to be as new. The plastic film has not even been pulled off the sides of the case. It came stock with 2 E5-2620 V3 but I might swap them for 2 E5-2690 V4's. So would be interested in something that could best utilize 28 cpu cores.

You could run it on World Community Grid which has Mapping Cancer Markers CPU work. Link to join WCG and our Team

 

[StefanR5R]

Older versions of Folding@Home's CPU-only application ( = not GPU accelerated application) did scale easily to, for example, 128 threads of a dual-socket 32c/64t EPYC computer of mine. That is, the performance of this application was not hurt by cross-socket communications much, or at all. It used, and I believe it still uses, an inter-thread (inter-process) mechanism for this (MPI) which is also used in HPC clusters for communication not only across sockets but also over Infiniband, Ethernet, or other node-to-node/ cabinet-to-cabinet interconnects.

Last time I checked though, Folding@Home cannot scale to that many threads anymore. Past a certain thread count, which I don't remember anymore, it would either plainly fail or run extremely slowly. You can test this yourself quite easily. Just configure the thread count of the "CPU slot" differently and watch F@H's log.

Whether or not cross-socket communications still perform as well as in the past is not known to me.

So in short, F@H on high-core-count CPUs (or on CPUs in general) is a drag nowadays. But as @TennesseeTony pointed out, the F@H consortium runs different experiments on CPUs than they do on GPUs, therefore CPU contributions are not obsolete. They do however give poor points per Watt hour compared to most (discrete) GPUs.

The alternative which several pointed out here already, i.e. various BOINC based projects, tends to scale to huge core counts easily because most of these projects use n copies of single-threaded tasks on such hosts, with n = as many of the host's hardware threads as you want.

 

[mmonnin03]

All that talk about communication between cores in FAH brings back memories of running two VMs on a Q66 as FAH was faster running two clients, two threads each vs a single 4 thread even with QRB. But the Q66 was a 4 core hack with two 2-core E66s on one substrate and communicate between them was bad (maybe back to the northbridge?) so FAH really slowed down if one WU was ran across all cores.

 

[craiggloyd]

FAH can use help with CPU-only projects! According to their project stat page (https://apps.foldingathome.org/serverstats) they always seem to have about 1.2M WUs for 0xa8 (CPU). I have about 21 6-cores boxes for that cause.. half the cores for FAH, half for BOINC using ProcessLasso.