The ***OFFICIAL*** F@H Harassment and discussion thread!!!

[Sparky]

I saw the harassment thread over at WCG and thought hey we should have one too So here it is. Use this thread to harass others to get them folding more, or to just have some general chat about F@H

[Sparky]

And I shall start

Jimwah, pick up the pace, c'mon! Your chart is downhill, and I'm right behind you in 24 hour average production. You aren't gonna let me produce more than you now are ya?

GAM, get a move on! Slacker, you used to be gunning for me, whatever happened to "sparkle day" as you liked to call it?

Angra, you aren't top producer anymore, and you aren't even red I'm guessing you had to get the computers running on other stuff for the time being....

[pythagoras]

The timing of this thread is so so bad for anyone who wants to harass me

hehehehehehe

[jimwah]

Yeah my output is down 'cos currently it's just from one box, the e6300 (apart from a random 396 pointer recently, I have no idea where that came from )

I've got machines spread thinly at the moment, but I can't have you overtake me in the dailies So maybe I'll have to get some boxen back on here soon Also, I've seen the parts for my new machine at work, e6400 due sometime this month

[jimwah]

By the way pythagoras I'm looking after 5th place for you Hope you don't mind The elusive Timarak better put his foot down too soon.... he's the next target...

[Bloody_Sorcerer]

I've only got my 3800x2 (1-2 cores) and x1900GT (if not running, 2 CPU cores, if running 1 CPU 1 GPU client). However, by F@H's stats, I have 4 active procs in the past week and 7 in the past 50 days... that 1 in the past week is totally unaccounted for and the 7 has me utterly baffled...
Shame the stats aren't more specific, ie which machine returned what and what that machine is hardware-wise...

[Sparky]

OK jimwah your 24 hour average is higher than mine again, happy?

Pythagoras, why is the timing of the thread bad? You are ahead of me by quite a bit but I am outproducing you so at the current rate I will pass you at some point

Sorcerer maybe some secret donor someplace

[STEvil]

http://www.xtremesystems.org/forums/...14#post1944614

I harass everyone, lets get some ideas to attract more memberships!!

[Sparky]

sounds good STEvil!

[texnofobix]

ooh ooh can i harass myself for only having P3s?

although I am carrying my 0.00904075731% share of our points.

[angra]

Angra, you aren't top producer anymore, and you aren't even red I'm guessing you had to get the computers running on other stuff for the time being....

yep yep there was a meeting for the main project that has users on my cluster this week. Normally they have a flurry of activity just before and after the meetings as people get their programs, benchmarks, and demos working. I am planning to try out the SMP client in a week or two, but I also was planning to pay back the WCG guys a bit for their help on our push against TPR.

I have a couple of surprises up my sleeve in the works, that won't come to fruition for a month or so...

[[XC] moddolicous]

yep yep there was a meeting for the main project that has users on my cluster this week. Normally they have a flurry of activity just before and after the meetings as people get their programs, benchmarks, and demos working. I am planning to try out the SMP client in a week or two, but I also was planning to pay back the WCG guys a bit for their help on our push against TPR.

I have a couple of surprises up my sleeve in the works, that won't come to fruition for a month or so...

Any help you can give to WCG would be great because DDTung is taking a DC break (well deserved may I add). That smp client should definitely boost your average quite a bit. All I've heard about it is how good it is. Plus, I don't think that those project 21xx are available under SMP are they?

[STEvil]

Out of curiosity, anyone tried putting "200%" as the desired cpu usage when installing?

[Sparky]

Out of curiosity, anyone tried putting "200%" as the desired cpu usage when installing?

no.... but the allowable range it gives is 5-100, so I have a feeling it wouldn't work. How could you run your CPU at 200% anway?

moddolicous, I have my newly acquired PC (repaired 3200+ winchester @ 2.3GHz) crunching for WCG for a spell as a thank-you to you guys

[jimwah]

Plus, I don't think that those project 21xx are available under SMP are they?

The SMP has its own set of WU, have a look down the current WU list here: http://fah-web.stanford.edu/psummary.html for anything with a GRO-SMP as the code (right at the bottom of the page). Very nice units

@ angra I've noticed a lot more reliability in the SMP client now, I literally haven't touched my C2D running it for weeks now, whereas in the first few weeks it came out, I had to check it constantly & delete cores & all sorts of annoying stuff. Although I vaguely remember the start of februaury being the date for a newer client, or the next stage of beta, can't remeber but something was happening...

Timarak has upped his output in the last few days, I might have to add another machine or two too keep him in sight

[[XC] moddolicous]

no.... but the allowable range it gives is 5-100, so I have a feeling it wouldn't work. How could you run your CPU at 200% anyway?

moddolicous, I have my newly acquired PC (repaired 3200+ winchester @ 2.3GHz) crunching for WCG for a spell as a thank-you to you guys

Thanks you also Sparky. I'm still 100% over here as I will switch back over to WCG when I am top50 here. Currently I am 61.

[angra]

Any help you can give to WCG would be great because DDTung is taking a DC break (well deserved may I add). That smp client should definitely boost your average quite a bit. All I've heard about it is how good it is. Plus, I don't think that those project 21xx are available under SMP are they?

so.

I tried halfheartedly to get the WCG client running on my cluster. Didn't work out so well. Got lots of error messages and stuff in the log like "if this happens repeatedly you need to reset the project".

I noticed that the WCG Forum is mostly Windows focused. I didn't see a ton of linux discussion. Am I missing it?

Anyway, the SMP client is very nice. I have it running on 16 nodes, and my PPD seems to be about 3/4 of what my whole cluster (50 nodes) was with the old client.

[Sparky]

I'm using BOINC on two systems (fold6 and fold7) to help them out, windows based though. Each one gave me an error when setting up but they seem to be working fine and I am getting credit for the points.

Glad to know the SMP client is working well for you angra, nice to see your stats climbing again If I can get my hands on a cheap dual core to plop in fold7 I'll run the SMP client there - also I heard that I could even run the SMP client on it - a single core A64 3200+ - because even though it wouldn't be faster it gets higher point WUs and doesn't have those horrid 212x WUs. I don't know, I'm toying with the idea...

[angra]

I'm using BOINC on two systems (fold6 and fold7) to help them out, windows based though. Each one gave me an error when setting up but they seem to be working fine and I am getting credit for the points.

OK, I checked and I have some points, so maybe it is working

I am not sure what the "systems" are, or what fold6 and fold7 are (are those your computer names maybe?). I did not follow the instructions they have on the WCG forum, because the client I downloaded via the instructions refused to unpack (reported a corrupt archive) - I went to the FTP site and downloaded the latest or close to latest linux client and used that. I probably screwed up a lot along the way . here is a dump of my initial output - running on a test system before I deploy to the nodes:

2007-01-17 10:59:04 [---] Starting BOINC client version 5.8.2 for i686-pc-linux-gnu
2007-01-17 10:59:04 [---] log flags: task, file_xfer, sched_ops
2007-01-17 10:59:04 [---] Libraries: libcurl/7.16.0 OpenSSL/0.9.8d zlib/1.2.3
2007-01-17 10:59:04 [---] Data directory: /XXXXXXX/BOINC_debug
2007-01-17 10:59:04 [---] Processor: 1 GenuineIntel Intel(R) Xeon(TM) CPU 3.06GHz
2007-01-17 10:59:04 [---] Memory: 3.84 GB physical, 11.18 GB virtual
2007-01-17 10:59:04 [---] Disk: 1.35 TB total, 1.18 TB free
2007-01-17 10:59:04 [World Community Grid] URL: http://www.worldcommunitygrid.org/; Computer ID: 101948; location: (none); project prefs: default
2007-01-17 10:59:04 [---] General prefs: from World Community Grid (last modified 1969-12-31 18:00:01)
2007-01-17 10:59:04 [---] Host location: none
2007-01-17 10:59:04 [---] General prefs: using your defaults
2007-01-17 10:59:04 [World Community Grid] Restarting task faah1238_d118n383_x2BPZ_00_2 using faah version 528

absolutely no friggin clue if this is even remotely correct . Also, the compute nodes can not see the outside world except via HTTP proxy (this is what I use for FAH) - will the BOINC client use a http proxy in the same way?

[Sparky]

OK, I checked and I have some points, so maybe it is working

I am not sure what the "systems" are, or what fold6 and fold7 are (are those your computer names maybe?).

Yes those are the computer names. They are listed as to the specs in my sig, near the bottom in that list.

I did not follow the instructions they have on the WCG forum, because the client I downloaded via the instructions refused to unpack (reported a corrupt archive) - I went to the FTP site and downloaded the latest or close to latest linux client and used that. I probably screwed up a lot along the way . here is a dump of my initial output - running on a test system before I deploy to the nodes:

2007-01-17 10:59:04 [---] Starting BOINC client version 5.8.2 for i686-pc-linux-gnu
2007-01-17 10:59:04 [---] log flags: task, file_xfer, sched_ops
2007-01-17 10:59:04 [---] Libraries: libcurl/7.16.0 OpenSSL/0.9.8d zlib/1.2.3
2007-01-17 10:59:04 [---] Data directory: /XXXXXXX/BOINC_debug
2007-01-17 10:59:04 [---] Processor: 1 GenuineIntel Intel(R) Xeon(TM) CPU 3.06GHz
2007-01-17 10:59:04 [---] Memory: 3.84 GB physical, 11.18 GB virtual
2007-01-17 10:59:04 [---] Disk: 1.35 TB total, 1.18 TB free
2007-01-17 10:59:04 [World Community Grid] URL: http://www.worldcommunitygrid.org/; Computer ID: 101948; location: (none); project prefs: default
2007-01-17 10:59:04 [---] General prefs: from World Community Grid (last modified 1969-12-31 18:00:01)
2007-01-17 10:59:04 [---] Host location: none
2007-01-17 10:59:04 [---] General prefs: using your defaults
2007-01-17 10:59:04 [World Community Grid] Restarting task faah1238_d118n383_x2BPZ_00_2 using faah version 528

absolutely no friggin clue if this is even remotely correct . Also, the compute nodes can not see the outside world except via HTTP proxy (this is what I use for FAH) - will the BOINC client use a http proxy in the same way?

Well if you are getting points I'm guessing it is working. I don't really know what is supposed to be in the log and what shouldn't. BOINC does support HTTP proxy, it is under options (at least in the windows BOINC client).

[angra]

Yes those are the computer names. They are listed as to the specs in my sig, near the bottom in that list.

oh man, duh! I am sorry. It has been long enough since I was doing system tweaking that my brain just kind of edits out the tiny print on the bottom of people's sigs now . It all makes sense now. Coolio. I guess a good question on my config is "how much PPD should a 3GHz Pentium-4 era Xeon get". I can run for a few days to see if it looks about right before deploying to the compute nodes...

[Drunner611]

Hey, so I finally figured out how to get F@H working past my firewall and now I want to post stats here on my forum page. How do I get those cool tickers like Angra and the rest of you have?

[Sparky]

It is a stats image from folding.extremeoverclocking.com

Just copy this into your sig, except remove the * in the IMG tags
[IMG*]http://folding.extremeoverclocking.com/sigs/sigimage.php?un=Drunner611&t=36362[/IMG*]

and you get this

[Drunner611]

Thanks much. Anyone know if it would be difficult to hijack like 30 school computers in the computer lab?

[Scimitar]

You need administator privaleges to install it as a service. An admin would notice the normal console or GUI version. You could also get into serious trouble doing that. Also, Stanford's policy requests that permission be obtained before running clients on computers that aren't yours.