Distributed computing team?

Which distributed computing project?


  • Total voters
    15
  • Poll closed .
glussier

glussier

Active Member
I am going away to Sacramento CA for 3 days next week, it might strech to 4 days if I decide to drive to Napa Valley to get a few cases of wine. Hopefully my folding rigs will take care of themselves.
 
Leoslocks

Leoslocks

_ _ _ _ _ _ _ _
Project: 5788 (Run 14, Clone 34, Gen 10)
Project: 5789 (Run 2, Clone 63, Gen 3)
Project: 5795 (Run 0, Clone 14, Gen 10)

Got a few working now. All my cards are nVidia now.
I have Big WU set as well as the Advanced Option.

They are nVidia specific. Here is the link to the WU anouncement.

http://foldingforum.org/viewtopic.php?f=52&t=11134
 
Last edited:
Burnt Powder

Burnt Powder

New Member
I have battled thru these recent 1920s on my SMP clients. About 1 in 3 failed and luckily after they failed I got 1760s in their place.

I would change over to the standard clients but I only get to check my clients about 3 times a day and my cores would chew thru the standard clients and be sitting idle because of Websense (need to use asknet=yes because of this crap). Two of my machines have 16 cores (HT) and I would have to run 32 standard clients on them (is that even possible?) to max all of the cores at 100%.........I can do this now with 4 SMP clients on each machine.

I am looking forward to a new/more stable SMP client...........anyone read an ETA on it?
Ok are there any client settings that will prevent my SMP's from getting 1920s and only 1760s?............the 1920s are killing me!!!! :(

I should be easily doing 40K a day with the hardware I am throwing at them.
 
Leoslocks

Leoslocks

_ _ _ _ _ _ _ _
The A2 core on the NotFreds usb sticks get about 4K per day running the 1920 PPD WUs. Hardly a hickup with three of these machines. Intel DG33TL Motherboard, Q6600, 2 GB ram and that is all the hardware.

I got my gaming card back from RMA. VisionTek 4870 X2. It too would not boot into Windows, installed and ran Driver Sweeper and cleaned up the old software, installed Cat 9.8 drivers and all is well with the gaming machine. That freeded up the GTX260 which replaced a 9800GTX+.
The client went from 5500 ppd to 7500 ppd. This is a BFG GTX260 55nm card. Now I have two 9800GTX+ cards waiting for a motherboard and processor.

Glussier, do you want to fool with this MSI board? Perhaps try running a NotFreds usb stick on a dual core.

Back on topic;
Acceptable size of work assignment and work result packets (bigger units
may have large memory demands) -- 'small' is <5MB, 'normal' is <10MB, and
'big' is >10MB (small/normal/big) [normal]?
Try Small setting

Set -advmethods flag always, requesting new advanced
scientific cores and/or work units if available (no/yes) [no]?
Try No to setting Advanced Methods.

Disable highly optimized assembly code (no/yes) [no]?

If you are having issues running the client, disable this. For 99.9% of the users, leave this at the default setting since this reduces the client performance a lot.
I don't know if I would disable this as I am sure it would cripple the client.
 
Last edited:
glussier

glussier

Active Member
Glussier, do you want to fool with this MSI board? Perhaps try running a NotFreds usb stick on a dual core.
If it's possible to get it working stably with an E7300 or an E8400, I can make it fold in my shack under your folding userid.
 
Leoslocks

Leoslocks

_ _ _ _ _ _ _ _
If I had a power supply, processor and another two gb of ram, I would try to get it working here. I would also need another ethernet port on the router.
I will send it to you with my blessings. Perhaps add it as Clairs id.
 
glussier

glussier

Active Member
Perhaps add it as Clairs id.
I can do that, she will have 2 comps in my shack. She's also suppose to send a 9800gt my way, so that she can increase her folding points and my electric bill. :)

There's also another of my neighbours who will send a q8200 and a gts260, as soon as Intel I5 is available, which should be around september. So, I'm looking forward to our folding production in the months to come.

If you like better, I could send a dual core cpu your way and get you a few gigs of ram. The ethernet port is easy to solve, just use a crossover cable to 1 of your p5q-e which already has 2 ethernet ports.
 
Last edited:
Leoslocks

Leoslocks

_ _ _ _ _ _ _ _
I will try the cross over cable on the next machine. I imagine I will have to enable Internet connection sharing on the P5Q-E machine. I will not have much time to fool with this for a few weeks. By then I should be getting close to having another processor and board on the way.

We are working toward 900K per week. Maintaining that will be an achievement.
 
Leoslocks

Leoslocks

_ _ _ _ _ _ _ _
I had an SMP client and the GTX260 working for a few days on the Intel DG33TL board. Soon the 1920 WUs started killing the SMP client. I went back to 4 classic clients and the GPU client.

I need a good FLAC player before I get another board and processor. I am looking at the Samsung P3
 
Burnt Powder

Burnt Powder

New Member
Leoslocks said:
Back on topic;
Acceptable size of work assignment and work result packets (bigger units
may have large memory demands) -- 'small' is <5MB, 'normal' is <10MB, and
'big' is >10MB (small/normal/big) [normal]?
Try Small setting

Set -advmethods flag always, requesting new advanced
scientific cores and/or work units if available (no/yes) [no]?
Try No to setting Advanced Methods.

Disable highly optimized assembly code (no/yes) [no]?

If you are having issues running the client, disable this. For 99.9% of the users, leave this at the default setting since this reduces the client performance a lot.
I don't know if I would disable this as I am sure it would cripple the client.
No go.....still getting the 1920s and liking them like a plague.
 
Bink

Bink

Will moderate for food
My ppd dropped in the last month because I've reimaged all the work systems. I now have a lot more systems folding (with the classic client). My last estimate has me at about 38 computers with at least 1 client running and another 25 with 2 clients running.

For some reason though, the Folding stats are only registering 10 processors and my ppd seems to reflect that. I'll need to check that things this end are behaving.

I believe I've located the source of the problem. All the clients must communicate with the internet via proxy (Squid). The clients are attempting this connection, but Squid is logging these errors (in cache.log) :

Code:
2009/08/27 15:04:23| ctx: enter level  0: 'http://assign2.stanford.edu/'
2009/08/27 15:04:23| HttpMsg.cc(157) first line of HTTP message is invalid

If I try to access the above URL "http://assign2.stanford.edu" using Firefox, via the proxy, I get this error:

ERROR
The requested URL could not be retrieved

Invalid Response
error was encountered while trying to process the request:

GET / HTTP/1.1
Host: assign.stanford.edu
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.2; en-GB; rv:1.9.1.2) Gecko/20090729 Firefox/3.5.2 (.NET CLR 3.5.30729)
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
Accept-Language: en-gb,en;q=0.5
Accept-Encoding: gzip,deflate
Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.7
Keep-Alive: 300
Proxy-Connection: keep-alive
Pragma: no-cache
Cache-Control: no-cache

The HTTP Response message received from the contacted server could not be understood or was otherwise malformed. Please contact the site operator.

Your cache administrator may be able to provide you with more details about the exact nature of the problem if needed.

Your cache administrator is webmaster.

Generated Thu, 27 Aug 2009 05:05:06 GMT by Squid (squid/3.0.STABLE17)

I've researched this a bit and it appears to be an issue with the folding webserver. The HTTP header seems to be invalid and Squid is unable to handle this. I'm not sure where to go from here.
 
Leoslocks

Leoslocks

_ _ _ _ _ _ _ _
2009/08/27 15:04:23| ctx: enter level 0: 'http://assign2.stanford.edu/'
2009/08/27 15:04:23| HttpMsg.cc(157) first line of HTTP message is invalid
The address ; http://assign2.standord.edu is used to check if the server is receiving. One method of determining is a server is down is to put the server IP in a web browser. One should see the OK message when the window opens.
I set the same OK message when putting the assing2.stanford.edu address in a browser.
 
glussier

glussier

Active Member
Just back from Sacramento, seems like my folding was doing better without me around. I should have stayed there for a few more days. :)
 

Associates