Folding Forum

toTOW wrote:This account is impressive by the number of active client, but each one taken individually is not that great. Each of their node folds at ~5500 PPD, which is basically a 2 GHz quad core ...

Anyway, it's great that such a cluster uses useful computations for their burn tests

I politely disagree with your maths...

According to official stats, they're running 688-692 clients (depending on how you count it) and ~39,000,000 points per day, therefore I'd put each node at closer to 55,000 ppd. I'd estimate that to be a stock clocked 2p quad core if they're working with Xeon...

Alternatively, each 3 hourly update on kakaostats the 688-692 clients are completing roughly 800 WU for roughly 6,000 points each, which works out at about 55,000ppd.

....But agree with your conclusion that it's great that such a cluster uses useful computations for their burn tests

Jorge1950 wrote:JV I'm interested in everything. I actually have fun analyzing statistics.

See the impact of the death of Steve Jocks on the production of Mac OS X/Intel (about 70 Tflops native). Or the great blackout of FAH, as 160 TFLOPS native, etc.

Though linking the change in contribution of OS X/Intel to Steve Jobs death would be ignoring a number of other reasons. His death also coincided with a shortage of WU's for OS X, there is and was just one server with work for OS X/Intel for folders running the 6.29 client and using the A3 SMP core. There are other servers with A3 work, but they require a newer version of that core which has not been released for OS X. It is available for Windows and Linux.

At the time the server often showed less than 100 WU's available. Later more work was added to that server, but persons folding with OS X reported they would often have to wait long periods to get another WU. Adding -advmethods to their folding setup helped. But not all of the folders check in with this and other forums to get the latest configuration help.

Some have heard about the V7 beta, and moved to that to get a more steady supply of work. But the OS X version is not ready for general issue, and at this point has a number of initial setup issues to be taken care of before I and others could recommend it for the average user. In the meantime some have just stopped folding, hopefully a good number will return.

Joe_H: "Though linking the change in contribution of OS X/Intel to Steve Jobs death would be ignoring a number of other reasons. His death also coincided with a shortage of WU's for OS X, there is and was just one server with work for OS X/Intel for folders running the 6.29 client and using the A3 SMP core. ..."

I like to think that the death of Steve Jobs; It was that brought about the fall of the collaboration; partners of "Mac os x/Intel". On Friday 11/4/2011 they collaborated with 141 TFLOPS, today 1/2/2011 they reach 35 TFLOPS (-106 TFLOPS). If as you say, was due to lack of foresight of PG WUs, for any reason, it is a fatal error.
Join the above HPCS, could not process around 60 million points; day 12/31/2011 and 01/1/2012 and the situation worsens. Is team donor, becoming you large, at FAH?

And accepting that the great blackout, It was an accident (-160 TFLOPS). If I as a director of a private company; I accept such chain lost, shareholders would not be very happy!

Edit: Steve Jobs. Thank you for the correction

Jorge1950 wrote:
I like to think that the death of Steve Jocks; It was that brought about the fall of the collaboration; partners of "Mac os x/Intel".

Sorry, but you already twice wrong write: Steve Jobs

Sincerely: Arto (1953)
(I am also difficulty with English)

Now that HPCS (HP Cloud Service) reached a 1000 million points, from December 16 to January 12: 27 days. KakaoStats KakaoStats Radar

Edit:

Jorge- I suspect that this could cover private info that the donor[s] are entitled to , and probably should, keep to themselves.. :~/
Be Grateful and Multiply ! ;~)

PPWU changed recently.

Now does this cloud service process just a few WUs using many cores to pump them out rapidly or does each WU take about the time it would take to complete on a normal machine but runs many clients simultaneously?

690 clients sumultaneously. ~690 clients listed as active on the HPCS stats form, last time I looked. You look too.