SCE Joins Stanford's Folding@home Program (B3D ID=32377)

[Kryton]

It would be more like below for something closer to the human brain.
- 100 billion neurons
- each neuron with 10,000 connections
- working at 10,000 hertz

The total floating point power required would be:
10**9 x 10**4 x 10**4 x 2 = 2 x 10**17 floating point operations

or 200 Petaflops

Maybe in 5 or 10 years with a DC simulation using PS4 or PS5


Right now it is 988 Tfp.

The human brain only has a 100Hz clock Also, there are only about 1000 synapses per neuron. Not that it means anything

 

[ban25]

intelligent design

Was that a Freudian slip?

 

[joker454]

They need to start offering achievement points for different milestones, like at 10 work units, 100 work units and 1000 work units completed.

 

[pascal]

The human brain only has a 100Hz clock Also, there are only about 1000 synapses per neuron. Not that it means anything

I decided to verify the numbers using wikipedia and ... http://en.wikipedia.org/wiki/Neuron

The human brain has a huge number of synapses. Each of the 10**12 neurons — one billion (long scale) has on average 7,000 synaptic connections to other neurons. It has been estimated that the brain of a three-year-old child has about 10**16 synapses (10,000 billion). This number declines with age, stabilizing by adulthood. Estimates vary for an adult, ranging from 10**15 to 5 x 10**15 synapses (1,000 to 5,000 thousand billion).[3]

OBS: the billion above has the meaning of million x million.

So being optimist it is about 10** 15 synapses. Now the frequency IIRC is in the KHz range. We need to have fast reflex.

 

[Arwin]

Was that a Freudian slip?

It was a play on the position of man in the creation of sentient computers ...

Highest number I saw (total) was 988, then when I looked about an hour later it had dropped waaaay down! I think we've seen our first peak.

 

[wco81]

Has anyone set up a B3D group?

 

[pascal]

The Beyond3D Team number is 32377

 

[Kryton]

I decided to verify the numbers using wikipedia and ... http://en.wikipedia.org/wiki/Neuron


OBS: the billion above has the meaning of million x million.

So being optimist it is about 10** 15 synapses. Now the frequency IIRC is in the KHz range. We need to have fast reflex.

Your brain doesn't work at Khz, it's about 100Hz but massively parallel (forget the clock operating at the same 100Hz everywhere, think about the physical clock distribution along chemical pathways and you'll realise it's impossible!). This article explains it all best: 500Tb capacity, 10-peta neural operation (which he argues is akin to 2 FLOP: one mul, one add).

Anyway, why did the PS3 FLOP count just get halved?

 

[Onlooker1]

Your brain doesn't work at Khz, it's about 100Hz but massively parallel

Doesn't anyone see the irony of comparing the power of the DC array to a human brain when the job it is trying to do is to simulate the motion of a few dozen simple proteins, over a time-scale of a millisecond, that might be whizzing around in one single human cell?

Anyway, why did the PS3 FLOP count just get halved?

My guess is the finger in the wind stats reporting system just knocked off all the initial signups who started a few days ago, produced a WU for kicks after 6 hours, then shut down their PS3?

 

[ban25]

Anyway, why did the PS3 FLOP count just get halved?

A PS3 is only considered active if it submits a work unit within two days of receipt. It's much higher for the other platforms (10+ days).

 

[patsu]

Slipping from: 732 TFLOPs (29876 active PS3s out of 35176)
To: 492 TFLOPs (30931 active PS3s and 36848)

Assuming they have not changed the FLOP calculation methods, almost half the people may have disrupted their folding (playing games on Saturday night ?). I think the TFLOP number is based on submitted WU only.

Hopefully more will continue during weekdays (since we have more active PS3 folders now)


EDIT: Oops... it just went up a little:
526 TFLOPs (30589 active PS3s out of 36535)

 

[pascal]

Your brain doesn't work at Khz, it's about 100Hz but massively parallel (forget the clock operating at the same 100Hz everywhere, think about the physical clock distribution along chemical pathways and you'll realise it's impossible!). This article explains it all best: 500Tb capacity, 10-peta neural operation (which he argues is akin to 2 FLOP: one mul, one add).

Anyway, why did the PS3 FLOP count just get halved?

I trust more the wikipedia data.

The concept is of distributed assychronous system with many parts with diferent speed of iteraction. There is no clock, but one can simulate this system with a minimum frequency which should be enough to sample/simulate all iteractions.
At 100Hz my guess no one would have samples enough to simulate somes reflex and activities.


Back to folding with PS3, looks like many people are playing games.

 

[homy]

To those who compared the TFLOPs between systems, don't do it it's not accurate.

How should the FLOPS per client be interpreted?
We stress that one should not divide "current TFLOPS" by "active CPU" to estimate the performance of that CPU running without interruption. Note that if donors suspend the FAH client (e.g. to play a game, watch a movie, etc) they enlarge the time between getting the WU and delivering the result. This in turn reduces the FLOPS value, as more time was needed to deliver the result.

 

[Arwin]

I've so far left my PS3 on 24h a day, and will continue to do so for at least a week ... When I'm not playing games, it's folding. Now that the work-week has started again, it should get a chance to do at least two WU a day.

After a full 24/7 week, I will see whether or not I will continue raising the electricity bill like that. No promises!

 

[Geo]

The numbers for B3D are interesting so far:

http://folding.extremeoverclocking.com/team_summary.php?s=&t=32377

Clearly their statistic servers are having issues. No points on Saturday, a whopping big points (10x higher than our typical, and 4.5x bigger than the first PS3 day) on Sunday. Nothing today yet.

If one were to assume that Sunday's 41k points were actually Saturday + Sunday, that would still be very impressive. Like a 7x increase over Team Beyond3D's recent daily averages.

Of course, there are two countervailing issues to consider. 1) Right now there is a strong "Lookie-Lou" factor. How many of these people will settle in for the long haul? 2) It hasn't rolled out to everyone yet, likely, and of course PS3 users will keep increasing over time. The total number of PS3 users is nearly a blip right now compared to what it will be in two years time.

 

[inefficient]

I've so far left my PS3 on 24h a day, and will continue to do so for at least a week ... When I'm not playing games, it's folding. Now that the work-week has started again, it should get a chance to do at least two WU a day.

After a full 24/7 week, I will see whether or not I will continue raising the electricity bill like that. No promises!

At my utility rate, I've calculated at 200W it should cost me about $30 USD a month to keep my PS3 folding 24hrs a day.

BTW, they've put up some good information here that I didn't catch before:
http://folding.stanford.edu/FAQ-PS3.html

Q.What type of calculations the PS3 client is capable of running?

A.The PS3 right now runs what are called implicit solvation calculations, including some simple ones (sigmodal dependent dielectric) and some more sophisticated ones (AGBNP, a type of Generalized Born method from Prof. Ron Levy's group at Rutgers). In this respect, the PS3 client is much like our GPU client. However, the PS3 client is more flexible, in that it can also run explicit solvent calculations as well, although not at the same speed increase relative to PC's. We are working to increase the speed of explicit solvent on the PS3 and would then run these calculations on the PS3 as well. In a nutshell, the PS3 takes the middle ground between GPU's (extreme speed, but at limited types of WU's) and CPU's (less speed, but more flexibility in types of WU's).


Q. It seems that the PS3 is more than 10X powerful as an average PC. Why doesn't it get 10X PPD as well?

A. We balance the points based on both speed and the flexibility of the client. The GPU client is still the fastest, but it is the least flexible and can only run a very, very limited set of WU's. Thus, its points are not linearly proportional to the speed increase. The PS3 takes the middle ground between GPU's (extreme speed, but at limited types of WU's) and CPU's (less speed, but more flexibility in types of WU's). We have picked the PS3 as the natural benchmark machine for PS3 calculations and set its points per day to 900 to reflect this middle ground between speed (faster than CPU, but slower than GPU) and flexibility (more flexible than GPU, less than CPU).

 

[DieH@rd]

So sunday was the best. Looks like "Sunday Foldathlon" was a sucsess.

 

[Rainbow Man]

At my utility rate, I've calculated at 200W it should cost me about $30 USD a month to keep my PS3 folding 24hrs a day.

Wow. 30 bucks for a 200W load that sounds like a lot to me.. Is electricity very expensive in japan?

BTW, they've put up some good information here that I didn't catch before:

Awesome find. I've no idea what sigmodal dependent dielecrtic and all that other stuff means but it's cool to know sort of the extents of he capabilities of the PS3 client.

Maybe one day they make it as general as the full CPU version.
Peace.

 

[Shifty Geezer]

Wow. 30 bucks for a 200W load that sounds like a lot to me.

It does sound a lot. A quick search places Uk electricity at less than 4p per kWH. 200W is 144 kWH per 30 day month, or <£6.

Perhaps EU consumer electronics prices are balanced out by lower service charges

 

[inefficient]

I'm looking at last months bill right now. I paid 7,914 JPY ($67 USD) for just 342 kWh.

So If 200W is 144 kWH per 30 day month, I should pay 3332 yen (~28USD) - Basically another 1/3 to by power bill just to Fold@Home!


Edit: Actually now that I think about it, some of that 7,914 might be the basic service charge and then the per kWh rate is on top of that. Because it does seem too high per kWh

Edit2: Sigh... it's very complicated: http://www.tepco.co.jp/en/service/custom/guide/guide04-e.html But it looks like I am going to get screwed. Maybe I should get the Night-Only package and Fold only at night.