Hi,
we finally have a new PhD student starting in June and another one by the end of the year.
So our usage of GPUGRID will raise again after June.

thanks.
gdf

SO more work for us to crunch ? good good :)

Wwlcome to both in advance and, like Maxinna said, it's a great "crunchy"good news for us !!!

Eric

this is very good news :-)
Congratulations !!!

Will you have work for Nvidia Tesla cards?

Virtual911

Good news!
Thanks for the heads-up Gianni :)

---

[CSF] Thomas H.V. Dupont
Founder of the team CRUNCHERS SANS FRONTIERES 2.0
www.crunchersansfrontieres

Nice - so my GTX750 will be able to crunch continously.

But what about my AMD R9 280X? (Finally it will catch a cold....)

Guys, is there a reason why don't you support other project with you resources, while you don't have any work unit to crunch? Like Einstein, Rosetta or World Community Grid.

Thanks for the answer.

Honestly, I'm becoming ambivalent to BOINC projects and have crunched for Rosetta for many years also many others.

Lack of any feeling of a purpose usually because scientists of many projects think that a project and its users are "set and forget". You tend to lose the inspiration that you might actually be contributing to something worthwhile.

A lot of projects fade away and I wonder whether we are doing any real science or just the "grunt" work to prove a concept for someone whom doesn't want to pay the cost of doing "grunt" work on their own clusters or someone elses.

GPUGrid is my last project standing.

EDIT TO ADD

I also believe that "Scientists" in general do NOT have any confidence in BOINC or the concept of distributed computing.

Betting Slip,
I can buy in to some of that. You do begin to wonder whether your efforts are worthwhile. But you have to remember that if these projects were major efforts by big pharmaceutical companies, they would probably be run on their own computers. But in fact, there seems to be a bias by a lot of people that these projects are not proprietary, which means that they will be published and perhaps that is as far as they get.

So these projects are often student exercises. There is nothing wrong with that; some of the most creative ideas (right or wrong) come during your early years, and you may not have the chance to try them out in later employment. And you may not know whether the results will be useful for many years even after they are published. So look at the science, and if it looks worthwhile, I would give it a go.

But you have to remember that if these projects were major efforts by big pharmaceutical companies, they would probably be run on their own computers.

They do this kind of research.
See http://top500.org/statistics/sublist/.
Select "All" lists and try different "Application area"s like Biology, Biochemistry, Life science, Medicine, Pharmaceutics.
Not all supercomputers are public.
Or try google pharmaceutics supercomputer or supercomputing pharma.
The main reason for doing this kind of research via BOINC is that the results are published, so it won't be the intellectual property of one big pharma company.

In theory, 100 GeForce GTX 1080 cards would have 887,300 TFlops SP and would scrape into that top 100 super computer list.

Noticed this one:
92nd Barcelona Supercomputing Center Spain MareNostrum
- iDataPlex DX360M4, Xeon E5-2670 8C 2.600GHz, Infiniband FDR
IBM 48,896 925.1 1,017.0 1,015.6 KW

6112 CPU's @ $1552/CPU = $9.5M + rest of the 2S systems and 1MW electric!

100 GTX1080's would use ~18KW. The rest of the systems might bring that up to 20KW - 25KW. Total purchase Cost ~$100,000.

< 1/100th of the purchase cost and < 1/40th the electricity, smaller footprint...

While I appreciate GPU's can't do everything CPU's can, only one 'super-computer' makes any sense.

Many research groups may be able to afford 25 computers with 4 such GPU's, but not all and what better way to do research than to reach out to the masses and facilitate all those non-researchers/enthusiasts/crunchers interested in supporting sensible, affordable scientific research, and allow them to actively participate via DC not just in scientific research but the greater scientific community.

---

FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

What is an amazing discovery is that out of the three replys to my post NON have addressed anything I said.

Guys, is there a reason why don't you support other project with you resources, while you don't have any work unit to crunch? Like Einstein, Rosetta or World Community Grid.

Einstein@home: Not biomedical, but have GPU (CUDA) app
Rosetta@home: Biomedical, but CPU only
WCG: it has Biomedical projects, but right now CPU only

By the way I've changed my mind long ago, and I've set Einstein@home as my backup (GPU) project. SETI@Home also have a GPU (CUDA) app.

What is an amazing discovery is that out of the three replys to my post NON have addressed anything I said.

Maybe our replies didn't directly respond to your comments, but perhaps that's because we get where you are coming from.
Generally there is a gap between the DC researchers and the crunchers. Here that gap has been bridged somewhat, however this is about the most complex scientific research being performed by DC and it's not a big, well funded group. All the research papers that resulted from our work are available to us and the general public. Perhaps the scientists could better elaborate on the results; explain what they mean, how they are/may be useful, but they have put some effort in.
With respect to DC it's viable as I've pointed out. Boinc isn't perfect but it does let us contribute to several/many projects, if we want to - good when there is a drought.

---

FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

SETI@Home also have a GPU (CUDA) app.

If attaching to SETI for NV work note that you will receive both cuda and opencl tasks. Make sure that you're running 362.00 (or previous) drivers or the new 368.22:

https://www.gpugrid.net/forum_thread.php?id=4291&nowrap=true#43513

What is an amazing discovery is that out of the three replys to my post NON have addressed anything I said.

I did get the fact that you were unhappy, but other than that you weren't very clear. If you have a question, why don't you ask it? Otherwise, we are free (and required) to interpret what you said for ourselves.

But I always point out, on this and other forums, that unless the research is applied, it doesn't do any good. Supporting very basic reseach without any immediate application is fine; it should be done. But you can't then expect to see any great results right away either.

Well, groups also don't have the funding or manpower of a Pharma to do the whole pipeline of drug discovery so they limit themselves to more basic research which is then applied in some cases by Pharma.
Pharmas also is in most cases simulating confidential structures so that would go a bit against the principles of some BOINC projects.
Also some stuff like my last paper on our software and my next paper on OPM are a proof-of-concept in a way to convince people of the power of the software and such simulations. Might sound quite indirect, however if you consider that it can convince companies to use it in applied research (already happening at least in one case) which could improve their productivity it's not such an indirect gain in my eyes. I actually (personally) feel this has more of an impact than our group studying a specific system related to a disease because most of the times people in pharmaceuticals have more experience doing that.

Einstein@home: Not biomedical, but have GPU (CUDA) app
Rosetta@home: Biomedical, but CPU only
WCG: it has Biomedical projects, but right now CPU only

Poem@Home (opencl)

Einstein@home: Not biomedical, but have GPU (CUDA) app
Rosetta@home: Biomedical, but CPU only
WCG: it has Biomedical projects, but right now CPU only
Poem@Home (opencl)

POEM has both CPU and GPU workunits. The GPU application used to crash a lot, but now seems stable. Their opencl workunits are well behaved and finish fast.
The project science is similar to Rosetta, predict spatial structure of protein. Beside that I don't now what they simulate. Have always plenty of work. Admins are quiet.

OT: What does [VENETO] or other in squarebrackets in usernames mean?

---