'Folding@Home Now More Powerful Than World's Top 7 Supercomputers, Combined' "The fight against coronavirus goes nuclear" - Tom's

[StefanR5R]

ServeTheHome recent joined the Tom's Hardware team with an impressively-fast Gigabyte G481-S80 server armed with a whopping eight Nvidia Tesla P100-SXM2's running for the cause.

Is that why their pulling ahead now?

Anybody check the stats on that boy ? what ppd is it getting ?

I don't see a particularly strong user in their team stats. (Edit, what borandi said.)
Except EXT64, but this user was running for Tom's with high output for months/ years already.

Tom's advantage is purely that they have more users, thus certainly more folding slots which keep requesting work.

When work scarcity subsides, e.g. when LTT users withdraw, we'll be at an advantage again. LTT are holding a Folding event from March 27 to April 10. So, work availability will go up after April 10 23:59 GMT, which is 4½ days before the end of our race against Tom's.

PS, the output of Tom's, PCMR, and ourselves dropped at the same time when LTT's output rose sharply. Ours dropped considerably more than Tom's.

 

[Assimilator1]

Every time we double our server resources, the number of Donors trying to help goes up by a factor of 4, outstripping whatever we do.

Wow! Lol, that's an amazing influx of donors!

Btw, anyone know how Rosetta have been handling new donors? And how many they're getting?

 

[Markfw]

Wow! Lol, that's an amazing influx of donors!

Btw, anyone know how Rosetta have been handling new donors? And how many they're getting?

No idea, but I have about 500 cpu (including smts as cores) and doing 300k ppd working for our team.

 

[Endgame124]

Wow! Lol, that's an amazing influx of donors!

Btw, anyone know how Rosetta have been handling new donors? And how many they're getting?

I saw a post on the Rosetta forums that they have had their active users more then double and they were running out of work units. They are working to adjust to the new demand.

Sounds familiar , eh?

 

[Assimilator1]

Ok, so heading that way, but not their yet.
I looked at R@H wesbite :-

Computing status
Work

Tasks ready to send

20392

 

[StefanR5R]

To be fair to the supercomputer guys: They are typically measuring peak FLOPS with the linpack benchmark, which is based on double precision arithmetic. The total contributed FLOPS count at Folding@home is probably referring to single precision arithmetic (FP32), which is well supported by consumer-level GPUs, unlike double precision.

The current supercomputing top dog, Summit, has got 9,216 POWER9 22-core CPUs and 27,648 Tesla V100 GPUs. Theoretical combined peak FP32 throughput of the latter is 434 PetaFLOPS.

Still, they can't hold a candle to F@h. :-P ;-)

 

[mopardude87]

No idea, but I have about 500 cpu (including smts as cores) and doing 300k ppd working for our team.

Ty for your massive contribution.That is insane but i know you got your reasons and i hope that contribution goes towards that as well!

I do love the fact ever nerd/geek is uniting on a front and contributing so massively to this, it's our time to shine and it puts a big smile on my face to see numbers push exascale. I had to google that cause yeah i never heard of such a thing till now!

 

[Assimilator1]

Me neither!
And in case I haven't said it to you before Mark, thankyou on all fronts to your massive contribution!