Opening up Folding@Home

[sk69]

I have only recently got to know about Folding@Home platform, and am super impressed on so many factors - 1) The excellent and robust distributed computing platform in terms of the engineering involved 2) The proof that it works in terms of scalability 3) The sheer number of people who are willing to learn and contribute as donors(and adding more servers for a noble cause) pumping in extreme CPU/GPU processing cycles 3) The proof that given a good cause in which a lot pf people believe, we now have a good advertised distributed computing platform in place. 4) A quickly growing community

Given the points such as above, it will be great if the Folding@Home teams begins to consider opening up/expanding the scope of projects beyond the original charter of the team. I have no idea at all what those projects are, but projects beyond the diseases (if there is compute power to spare post COVID). Are there other noble causes in the world which can benefit from such as immense distributed computing power? Perhaps there can be some form of public voting, which can help the team cherry pick what other projects can be onboarded to FAH?

Once the COVID problem is more manageable in future, it will be great to see FAH using the compute power to continue solving other interesting problems (apart from the problems/diseases it already was focusing on such as cancer etc.).

Again, maybe some or most of this is already happening and I just do not know about it as I am just starting in this space.

[Neil-B]

I am no way linked to the FAH team - just making an observation as a regular folder.

FAH has over a number of decades developed its approach and toolset for specific market which, even after COVID-19 is managed, will continue to utilise any compute resource available - although this is perhaps likely to significantly reduce as much of the current supporting compute resource returns to its original purposes and people become less focused on supporting FAH as the COVID-19 situation stabilises/abates.

There are more generalise distributed computing platforms that support a wider variation of causes - but the purpose of FAH is to advance the science within general scope of its market (which is obviously important given the massive rush to participate over the last month or so) and which is not going to go away … Now that many more people have realised the importance of FAH (and similar efforts) to redirect resource away from the science once just one challenge is managed would perhaps be a shame.

[sk69]

Thanks for the response. I am all in support of the specific market and cause being focused by FAH. The thought behind my post was mostly to think if there are ways to make use of "spare" compute power. If FAH's planned projects/missions are able to continue making use of the ramped up compute power (assuming some % of recent donor may continue to donate resources), then that is fine anyway. I completely understand and respect the specific market and critical cause which FAH drives, and my intent was just to check if there might be spare compute power in the future, and that may be better utilized, which will also keep future donors coming and excited instead of having idle donors connected to FAH with no WUs.

[Neil-B]

I am hopeful that the research/technical teams having ramped up their efforts to supply/utilise a much larger compute resource (and are still doing so - everything is still expanding) will be fully able to utilise whatever compute resource sticks with FAH after the current issues abate … They have a much larger much more powerful "toy" and they really won't want anyone to take it away as they can play much bigger "games" with it

Whilst during the current expansion for the most part people have been very patient as the teams upscale - you are right, once COVID-19 no longer has everyone's focus people may drift away from FAH if they find their "donations" of compute resource aren't being utilised - if they do though I hope that they will drift away to some other distributed project having come to realise that many small things can do huge things when they come together.

[iceman1992]

This WU shortage should be temporary, as they're right now still struggling to catch up with the ever-increasing donors (which is an awesome problem to have!), as evident from the still increasing performance => https://stats.foldingathome.org/os.
As a longer-time folder (I started in 2010), I can tell you that we didn't have this problem of not getting WUs before this sudden rush.
I certainly hope that most of the new donors that came here because of Covid-19 will stay afterwards, the FAH scientists definitely can use the compute power to work on cancer and other diseases. I'm sure we all hate cancer too as much as Covid.

[JimboPalmer]

Such a framework is available, BOINC

https://boinc.berkeley.edu/projects.php

Folding@Home however is run by biochemists, for biochemists. All it's code is organized toward folding organic molecules, mostly proteins. That is their specialty and expertise.

There are no shortage of ailments for F@H to work on once they beat COVID-19

https://foldingathome.org/diseases/

These all play into the strengths of the researchers expertise. In the wildly unlikely event that every reason to fold proteins is coped with, volunteers can swap to BOINC.

[The cynical part of me knows that as soon as COVID-19 is not a threat, 90% of folks will lose interest]

[Jonazz]

Trust me, there is an abundance of diseases that need better understanding. FAH won't run out of possible projects any time soon.

[The cynical part of me knows that as soon as COVID-19 is not a threat, 90% of folks will lose interest]

If we lose 90% of the current folders, that would still be double the amount we had before this surge!

[iceman1992]

Still hoping we can remain > 1 exaflops after Covid

[Jonazz]

Still hoping we can remain > 1 exaflops after Covid

Yeah me too. That would still be huge.

[michelv]

Coming from the other side, I am really curious and wandering why there is even a separate platform like F@H. BOINC has (nearly) everything they would need, and using BOINC would increase the user base massively. As BOINC is much larger and wider, maintenance is wider (maybe not easier) and as a user you do not need to get used to all the weird quirks of the F@H software (Oh my, this needs a jump into the 20ies. sorry).

But: maybe there is some profound reason why F@H needs to be (partly) closed source and organized the way it is. From a user perspective: please move to BOINC.
(So now I had to restart my Win10 box because MS needs to do some updates.. all my workunits in F@H are gone/lost where BOINC smoothly continues with COVID work from the Rosetta team).

[PantherX]

Welcome to the F@H Forum michelv,

Please note that F@H was running before BOINC. Moreover, there were talks about merging them into a single platform but that didn't happen since the roadmap didn't line up so F@H and BONIC continue their separate ways.

[iceman1992]

As BOINC is much larger and wider

Not sure what you mean by 'larger', but if speaking about performance and participants,
Taking data from wiki, it seems BOINC has around 137800 active participants with 41.5 petaflops of compute performance.
Whereas F@H is running at 2363 petaflops as of this week and I think over 1 million active participants.

[michelv]

Oh my, I seriously do not want to tread toes here. I am just responding to sk69 and his/her wish to open up F@H (and thus supporting JimboPalmer): there is already a very large platform for distributed computing.

Larger... well, comparing data of this week to a source on WIKI is apples and...
BOINC has a significant number of projects (eg worldcommunitygrid with over 777K users). Mind you, it is a platform. F@H is one organisation (as far as I see it) with one type of projects (What's in a name)

But I didn't know F@H has also been around for over 20 years as well.
Keep up the good work!

[Jesse_V]

Yeah, Folding@home was the second distributed computing project out there, right after SETI@home.

[HaloJones]

Third after Seti which was after GIMPS

Before doing FAH, I was doing Genome@Home and Distributed Folding both in around 2002.