Folding@Home

[ddrueding]

 

[Clocker]

:lol: :roll: :mrgrn: :wink: :excl:

 

[Handruin]

It's pretty sad for me that Sir ddrueding will have complete more total work points in the matter of a few weeks compared to what I've been able to accomplish in a couple years.

But hey, congrats! :rr:

 

[ddrueding]

I find it a funny coincidence that all of my machines came out to exactly the same WU per week as you guys...oh, well...just put another A64 3200 on 24/7

 

[The JoJo]

I finally upgraded my file server to a Asus p4c800-e with a PresHott 3000 (couldn't get a northwood), and I'm running 2 FAH instances on it (Thanks Mark!). Running a Fedora Core 2 SMP kernel with it, no problems.
YUM is recommended, seem to work nicely!

 

[ddrueding]

Another 2 years at this rate :roll:

 

[ddrueding]

Another A64 3000+ on the fire 24/7

 

[Clocker]

Finally started back up my 2nd folder/file server. ~175 per day wasn't too bad on just one machine...looking forward to producing more...

C

 

[Pradeep]

Added another Nocona to the fleet. So I have the following crunchers:

Two Nocona 2.8
One Pee4 2.8
One Pee4 2.4
One Pee4-M 1.6

There's a potential for another two Noconas in the coming days.....

 

[Clocker]

I have a work unit that is taking almost 150MB or RAM. Have you guys gotten some big ones like that? Or, is that just normal size these days?

C

 

[Pradeep]

That seems rather large. Did you select the "dl large work units" option?

 

[Handruin]

I got one running that's close to yours, right now:

 

[ddrueding]

Just sampled 3 of mine at random. All were given the OK to DL large WUs.

13.6MB
134.1MB
134.1MB

Looks like they are common, but not guaranteed.

 

[Pradeep]

Lost a Nocona 2.8, but added a Dual Xeon 2.4(533) and a P4 3.06.

 

[Clocker]

My fastest folder (see sig) will be down for a about week as I redo my den.....I'll just have a 3200+ running this week.

 

[P5-133XL]

At 6000 points per day we finally seem to have stopped our slow drop in the standings (from 39->94). At least for the next 100 days or so, we should stay at 94 or better. Congrats to all.

If we want to start steadily moving up in rank (at a reasonable rate) then we need another 2000 points per day. Any idea's on how to get them points?

 

[ddrueding]

I'll be getting another 6-7 2Ghz+ systems online in the next few days.

 

[blakerwry]

with winter coming up I will increase my folding output from nil to something... can't imagine it will be too great, but I'll try.

 

[Pradeep]

Added another three Nocona 2.8s, unfortunately they are heavily used and availability is varied.

 

[Bozo]

The pathetically slow box that I have folding managed to loose two days of work.

I'll keep plugging away at it.

Bozo :mrgrn:

 

[Handruin]

I'll be adding another machine under my account soon, once I find the right parts to get it going.

 

[i]

It's getting cold in here. Time to start up a client.

You should all be quivering with fear. My 300MHz Pentium II is now crunching away.

 

[Clocker]

My 2nd folder is back up online today after being down for about a week. Don't anticipate any downtime in the foreseeable future.

C

 

[blakerwry]

I've added 2 folders, a AthlonXP 1700+ w/ SSE and an AMD K6 @ unknown MHz. (how can I find out in linux?)



I might also add another AthlonXP and/or pee4

 

[i]

I've added 2 folders, a AthlonXP 1700+ w/ SSE and an AMD K6 @ unknown MHz. (how can I find out in linux?)

"dmesg" should be able to tell you. It prints out the messages produced back when the system booted up. There can be a lot of them, so you might try piping the output from dmesg through "less" (allows you to view text from any source one page at a time), like this:

me@nowhere:~$ dmesg | less

Or, if you're in a hurry, try filtering dmesg's output using "grep" (a pattern matching utility):

me@nowhere:~$ dmesg | grep processor

...or maybe...

me@nowhere:~$ dmesg | grep AMD

 

[i]

Oh, one good thing to know: dmesg's buffer is limited, and some parts of the OS will occasionally continue to send out messages to the system that dmesg picks up on. As a result, if your Linux system has been up for a while, the stuff you're interested in may have been bumped out of the dmesg buffer.

In that case, take a look in /var/log and see if you can fish out any useful information from the file called "syslog", or any of the multitude of sequential backups of previous syslogs you've probably got hanging around in there. (Note: a ".gz" extension means you'll need to run gunzip to decompress the file.)

 

[blakerwry]

Detected 200457 kHz processor.
Console: colour VGA+ 80x25
Calibrating delay loop... 399.77 BogoMIPS
Memory: 95308k/98304k available (1112k kernel code, 412k reserved, 1408k data, 6
4k init, 0k bigmem)
Dentry hash table entries: 262144 (order 9, 2048k)
Buffer cache hash table entries: 131072 (order 7, 512k)
Page cache hash table entries: 32768 (order 5, 128k)
VFS: Diskquotas version dquot_6.4.0 initialized
CPU: L1 I Cache: 32K L1 D Cache: 32K
CPU: AMD AMD-K6tm w/ multimedia extensions stepping 02
Checking 386/387 coupling... OK, FPU using exception 16 error reporting.


---

md driver 0.90.0 MAX_MD_DEVS=256, MAX_REAL=12
raid5: measuring checksumming speed
raid5: MMX detected, trying high-speed MMX checksum routines
pII_mmx : 314.706 MB/sec
p5_mmx : 290.322 MB/sec
8regs : 280.035 MB/sec
32regs : 201.168 MB/sec
using fastest function: pII_mmx (314.706 MB/sec)



ooohh la la.. what a crappy folder... it should be replaced shortly with a pII of some sort... probably twice as fast.

 

[The JoJo]

Uh, how about just "cat /proc/cpuinfo" ? No need for searching that way.

 

[i]

Uh, how about just "cat /proc/cpuinfo" ? No need for searching that way.

Neat! I've yet to come across a *NIX system that didn't store useful info in /var/log and via dmesg. Is having helpful info stored in /proc/cpuinfo relatively universal? I hope so ... that's going to be useful. Thanks JoJo.

 

[blakerwry]

i wonder if this machine is even fast enough.... 200MHz, but it sits mostly idle and is running 24/7... I remember the baseline for FAH being a pII/Celeron 400MHz that is on 8 hours per day or more. Hope my units aren't expiring...

 

[i]

Hope my units aren't expiring...

How can you tell if that happens?

 

[blakerwry]

I know you can get a mesage that the unit is expired and a new one will be re-downloaded. I've seen this a few times when restarting a folder after having not folded for a season. But I have no idea if there are any checks in place during a single WU to ensure that it isn't going to be expired when it is completed.

 

[Clocker]

deadlines are given here:

http://vsp27.stanford.edu/cgi-bin/allprojects

 

[blakerwry]

ugh, I jsut found out that my p4 has been folding towards group 0, under the anonymous name.

 

[ddrueding]

Well, the majority of my folders were just taken offline. There is an incompatability between it and Star Wars: Battlefront. Levels take 5-10x longer to load with the client running in the background. It seems some of the portions of the app use "Idle" priority. :evil:

 

[Handruin]

Wow, that's a shame... I wonder if there is a way to set F@H to a sub-idle priority? I can't figure why a game would want to use idle priority during a map load. Only guess is some type of cut-scene, or music perhaps.

 

[ddrueding]

It's a real shame consiering how little that game is played in the first place. I'm thinking f seting up some way to stop/start the service when the game is launched. I already launch the game via .CMD file (that also handles Daemon tools). I just need to look up the syntax and figure out how to restart the service after the game is closed. Currently it looks like this:

@echo off
"\Program Files\D-Tools\daemon.exe" -unmount 0
cd "\Program Files\LucasArts\Star Wars Battlefront\GameData\"
"\Program Files\LucasArts\Star Wars Battlefront\GameData\Battlefront.exe" %1 %2 %3 %4 %5 %6 %7 %8 %9

It currently unmounts the drive due to copy protection issues (no-cd patch is installed). Any suggestions on how to modify this?

 

[blakerwry]

perhaps...

@echo off
"\Program Files\D-Tools\daemon.exe" -unmount 0
cd "\Program Files\LucasArts\Star Wars Battlefront\GameData\"
start "BattleFront" /HIGH c:\progra~1\lucasa~1\Starwa~1\GameData\Battle~1.exe %1 %2 %3 %4 %5 %6 %7 %8 %9

 

[Pradeep]

Any news on why memory utilisation has gone up on certain units (100MB+) when the "dl large units" option hasn't been selected? I had to take the client off a bunch of machines due to this.

 

[Pradeep]

But on a brighter note I added three Nocona 2.8s for "burn-in" testing