How Often Are Stats Updated?

[TC]

I signed up yesterday and this morning I had 15 units completed according to the Stanford stats list. I checked it again tonight approx 13 hours later and it was still at 15 - but I know my machines have turned in 9 more results. I know I saw 3 hours somewhere, but it's been more than 12 and nothing has been updated. Is this just a temporary server problem such as I'm used to with Seti?

[Punk]

ya, stats are supposed to be updated every three hours, but they haven't been updated since yesterday afternoon for some reason - think they r having server problems or something. they r also upgrading to a newer version of F@H (2.0), so it might be becuase of that.

[TC]

Originally posted by Would71
Heh.. I was just on my way in here to mention the 2.0 thing.. I think that's the hold-up on the stats this weekend. I hope they release it this week. Hopefully I should have my GHz Duron in the mail by the end of the week, and what a way to welcome it to the household.. with a new Folding client. :D

Is this the version that hopefully supports multithreading?

[TC]

Originally posted by Would71
It is indeed. :cool:

Sweet - guess I'll finally devote my two dual boxes to folding.

[thefly]

Ya! I've noticed that the stats are stuck too.

Ah well, it will be worth it once version 2 is out!

[SickBoy]

So TC which of those boxes are you Folding with? :)

SickBoy

[Lews]

The current client supports multithreading. I have been using 1.34 on my linux box for quite awhile. Its the dual 1ghz Pentium III. I just have the client running 2 times. Each instance shows up on top using individual cpu's.

[SickBoy]

I think what they mean by "multithreading" is the ability of the client to use both processors at 100% utilization with just one client running.... theoretically this would halve the time needed for a protein for SMP machines.

Running multiplie clients is something a lot of people have been doing for now, including people that use just single processor machines.

SickBoy

[Lews]

Running just one client does not use up 100% of both processors. On my machine it goes about 55 one 45 the other. give or take a percentage point for other processess.

Firing up the second client does use up 100% of both processors which in effect double the amount of wu's processed.

That is one thing I haven't understood is doing more than one client on single processor machines as it just takes longer to process each client. I suppose it would be beneficial if your clients aren't connected to the internet 24/7

[GuNRocK]

yah the stats have sucked latly...i finshed like 3 WU and dident get credit for any....i have seen taht i have been getting the fast units now....but i am not sure how much there worth and how they affected my total WU yet... here is apic c:\my documents\wufast.bmp

[Ploaf]

Originally posted by Lews
That is one thing I haven't understood is doing more than one client on single processor machines as it just takes longer to process each client. I suppose it would be beneficial if your clients aren't connected to the internet 24/7
I do it only so that my pc's still have something to fold when the servers go down. There is no benefit to it when things are running smoothly. Caching should fix this problem too.

[TC]

Originally posted by SickBoy
So TC which of those boxes are you Folding with? :)

SickBoy

Right now I have one folding client and one seti client running on each of the dual boxes. If the new folding client supports multithreading (by which I mean one client would use both cpu's to halve the time - in the way photoshop runs on a dual system) I will devote those two boxes entirely to folding. I have a 3rd client running on one the tualatin boxes. Even if the new client is not multithreaded I might take Seti off altogether and just run two clients on each box. Seti isn't multithreaded either so the only thing you can do with a dual box is run two clients, one on each cpu.

[robjustice]

Welcome to the folding team. From your list of computers, you have alot of firepower. Here are your stats as of 7:30

55522 TC (www.overclockers.com) 86 7 2


Greatwork, and glad to have you on the team.

Fold on Brother...

[TC]

Originally posted by Would71
I think we're making a slight error here.. equating "multithreading" only with SMP support. If memory serves, (and often it doesn't) multithreading should even be of some benefit on single-CPU machines.

You may be right - but I was thinking that a program that is multithreaded can make use of SMP, whereas a program that is not multithreaded can only make use of one cpu at a time. In other words SMP is using more than one cpu at a time, and multithreading is the means by which a program does this. Heh, well now I may be confusing myself :)