WU Evaluation times

[Ronin]

So, I went through about 25 of my crunchers, pulled the SetiSpy logs, and started filtering. I removed all WU that didn't show 100% complete (which bugs me, because with the 25 or so logs that I pulled, there were a LOT of WU that weren't 100%...~6-7k, I'd say), then did an average of the WU processing times to come up with a general idea as to what crunches how.

2x2.8GHz Xeon - 4h 44m
2x2400+ MP - 3h 39m
2xOpteron 240 - 2h 50m
4600+ X2 - 1h 55m
4800+ X2 - 1h 36m
A64 3000+ - 2h
FX53 - 1h 37m
FX55 - 1h 33m
P3 600 - 9h 27m
P4 1.7GHz - 5h 7m
P4 1.8GHz - 4h 3m
P4 2.4GHz - 3h 19m
P4 2.8GHz FNW - 2h 30m
P4 2.8GHz 1h 55m
P4 2.26GHz - 3h 6m
P4 2.66GHz - 2h 43
P4D 3GHz - 1h 56m
P4D 3GHz - 1h 58m
P4HT 3.6GHz Laptop - 2h 28m
P4HT 2.8GHz - 3h 25m
P4HT 2.8GHz - 2h 46m
P4HT 3.2GHz - 3h 3m
P4HT 3.2GHz - 2h 44m
P4HT 3.06GHz - 3h 11m
P4HT 3.0GHz - 2h 44m
P-M 1.6GHz - 2h 6m
2400+ XP - 2h 46m

Now, some of the WU times might be misleading (such as the Xeon box, which since it's started crunching on 2/5/04, has put out 10304WU). This is an example of a day for that box:

10/1/2005 12:40:21 AM
10/1/2005 1:07:56 AM
10/1/2005 2:48:59 AM
10/1/2005 4:21:48 AM
10/1/2005 4:47:23 AM
10/1/2005 6:00:09 AM
10/1/2005 7:40:07 AM
10/1/2005 8:18:52 AM
10/1/2005 9:38:18 AM
10/1/2005 10:44:41 AM
10/1/2005 12:28:03 PM
10/1/2005 1:06:57 PM
10/1/2005 2:02:54 PM
10/1/2005 3:17:54 PM
10/1/2005 5:07:26 PM
10/1/2005 6:01:14 PM
10/1/2005 6:34:18 PM
10/1/2005 7:17:51 PM
10/1/2005 9:59:05 PM
10/1/2005 10:48:20 PM
10/1/2005 11:18:55 PM

The 4800+ X2 at stock seems to be the front runner, averaging ~30 WU/day since I brought it online (I really need to get the Dual DC Opt275 back and running). If I've missed a CPU and I have it, let me know and I'll be happy to check the logs.

 

[Wolfsraider]

Wow very nice Thank you so much Ronin

 

[petrusbroder]

Wow, that is just great! Thanks!

 

[Hurricane Andrew]

Very nicely done. Thanks! :beer:

Is that A64 3000+ a 754 or 939? I have a 754, OC'ed to 2.2 GHz, and it's been taking just over 2:10 per WU on average.

 

[Smoke]

Your Xeon box is a good example of why I think the best measure is to divide 24 hours by the average daily output of the tested system. Then the real "strength" of the systems can then be sorted in a more realistic fashion.

For example, my primary work station is a 3.2 P4 HT system that averages 17.84 WUs @ day according to my SetiQueue: TeAm Enterprise Clients sorted by RPD

24 / 17.84 = 1.345 hours per WU or 1:20:42 (h:m:s) hours per WU

If you run them through a SetiQueue, just use the RPD number and forget about the FAST WUs for everything averages out in the long run.

 

[Ronin]

Originally posted by: Hurricane Andrew
Very nicely done. Thanks! :beer:

Is that A64 3000+ a 754 or 939? I have a 754, OC'ed to 2.2 GHz, and it's been taking just over 2:10 per WU on average.

754, on an AMD-8151 reference board (it's actually a flubbed 3400+, as it's not running at 2.2GHz)

 

[Ronin]

Originally posted by: Smoke
Your Xeon box is a good example of why I think the best measure is to divide 24 hours by the average daily output of the tested system. Then the real "strength" of the systems can then be sorted in a more realistic fashion.

For example, my primary work station is a 3.2 P4 HT system that averages 17.84 WUs @ day according to my SetiQueue: TeAm Enterprise Clients sorted by RPD

24 / 17.84 = 1.345 hours per WU or 1:20:42 (h:m:s) hours per WU

If you run them through a SetiQueue, just use the RPD number and forget about the FAST WUs for everything averages out in the long run.

I don't...all my queueing comes from SetiDriver. I suppose it might be easier to track that way, though, eh?

 

[Assimilator1]

Hey great stats Ronin

I have just got to assimilate my bosses new Pentium M 1.6GHz!

 

[Smoke]

Originally posted by: Ronin

Originally posted by: Smoke
Your Xeon box is a good example of why I think the best measure is to divide 24 hours by the average daily output of the tested system. Then the real "strength" of the systems can then be sorted in a more realistic fashion.

For example, my primary work station is a 3.2 P4 HT system that averages 17.84 WUs @ day according to my SetiQueue: TeAm Enterprise Clients sorted by RPD

24 / 17.84 = 1.345 hours per WU or 1:20:42 (h:m:s) hours per WU

If you run them through a SetiQueue, just use the RPD number and forget about the FAST WUs for everything averages out in the long run.

I don't...all my queueing comes from SetiDriver. I suppose it might be easier to track that way, though, eh?

It is the only way I know of to monitor a wide-spread mini-team (not on the same LAN).

By giving each computer(client) its very own PORT NUMBER and using a properly organized naming convention, you can tell in a flash which system is not functioning as it should.

 

[Ronin]

Smoke, your queue have room for 700-800WU/day from me (pulling down, that is)? If so, maybe I can hook in and just abuse yours.

 

[Smoke]

Originally posted by: Ronin
Smoke, your queue have room for 700-800WU/day from me (pulling down, that is)? If so, maybe I can hook in and just abuse yours.

You'd like it better if you had the control switch so I suggest you just install one of your own.

It might very well choke my little dsl account.

But we can try if you want.

 

[amdxborg]

Wow thanx Ronin!

 

[TurtleBlue]

Ronin
:thumbsup: This...is a A-1 effort! :thumbsup:

:beer: My top 2 of 4 crunchers are p4 3.06 hyper & p4 2.4 and have similar results :beer:

:shocked: I plan on O/C the p4 2.4 CAREFULLY & see how much more I can get!

That Xeon box results is astounding...did you hook up you car's Nitro system to it? :Q

 

[Spacehead]

Ronin, you might want to check out SETIWatch. It'll sort out your average times for you(among other things) as long as you have SETI Driver set to create a 'SETILog.csv' file for each comp.
Depends on how you have things set up though i suppose as to how easy it would be for you.

Thanks for the info

 

[Dalephi]

Nice. I guess I will have to start saving for an X2.

Thanks Ronin.

 

[Ronin]

Originally posted by: Spacehead
Ronin, you might want to check out SETIWatch. It'll sort out your average times for you(among other things) as long as you have SETI Driver set to create a 'SETILog.csv' file for each comp.
Depends on how you have things set up though i suppose as to how easy it would be for you.

Thanks for the info

SetiWatch proved hit and miss (because scanning across the network for the amount of PCs I had crunching would kill system resources, AND take a while just to compile the information).

 

[Wolfsraider]

Ronin,

your xeon scores are a tad low. are you running naconas?

here is one day on mine:

2005 Nov 13 11:34pm 2.321,9.03 0.676 borg-bowser.hsd1.wa.comcast.net. 3:01

2005 Nov 13 11:33pm 2.581,9.67 0.627 borg-bowser.hsd1.wa.comcast.net. 3:12

2005 Nov 13 10:42pm 2.549,9.59 0.631 borg-bowser.hsd1.wa.comcast.net. 3:16

2005 Nov 13 09:13pm 2.549,9.59 0.631 borg-bowser.hsd1.wa.comcast.net. 3:17

2005 Nov 13 08:33pm 2.518,9.5 0.638 borg-bowser.hsd1.wa.comcast.net. 3:16

2005 Nov 13 08:23pm 2.486,9.42 0.644 borg-bowser.hsd1.wa.comcast.net. 3:11

2005 Nov 13 07:27pm 11.175,18.08 0.426 borg-bowser.hsd1.wa.comcast.net. 3:14

2005 Nov 13 05:53pm 2.388,9.18 0.663 borg-bowser.hsd1.wa.comcast.net. 2:58

2005 Nov 13 05:12pm 11.175,18.08 0.426 borg-bowser.hsd1.wa.comcast.net. 3:13

2005 Nov 13 05:05pm 22.055,18.08 0.426 borg-bowser.hsd1.wa.comcast.net. 3:15

2005 Nov 13 04:12pm 2.216,8.8 0.695 borg-bowser.hsd1.wa.comcast.net. 3:07

2005 Nov 13 02:54pm 2.181,8.73 0.701 borg-bowser.hsd1.wa.comcast.net. 3:10

2005 Nov 13 01:57pm 2.25,8.87 0.689 borg-bowser.hsd1.wa.comcast.net. 3:04

2005 Nov 13 01:48pm 2.216,8.8 0.695 borg-bowser.hsd1.wa.comcast.net. 3:10

2005 Nov 13 01:02pm 2.181,8.73 0.701 borg-bowser.hsd1.wa.comcast.net. 3:10

2005 Nov 13 11:40am 2.181,8.73 0.701 borg-bowser.hsd1.wa.comcast.net. 3:04

2005 Nov 13 10:49am 2.15,8.68 0.658 borg-bowser.hsd1.wa.comcast.net. 2:59

2005 Nov 13 10:36am 2.114,8.61 0.711 borg-bowser.hsd1.wa.comcast.net. 3:05

2005 Nov 13 09:49am 2.006,8.43 0.727 borg-bowser.hsd1.wa.comcast.net. 3:05

2005 Nov 13 08:33am 2.006,8.43 0.727 borg-bowser.hsd1.wa.comcast.net. 3:05

2005 Nov 13 07:42am 12.431,18.08 0.426 borg-bowser.hsd1.wa.comcast.net. 3:11

2005 Nov 13 07:20am 1.93,8.33 0.737 borg-bowser.hsd1.wa.comcast.net. 3:08

2005 Nov 13 06:36am 1.893,8.28 0.742 borg-bowser.hsd1.wa.comcast.net. 3:07

2005 Nov 13 05:23am 1.856,8.24 0.746 borg-bowser.hsd1.wa.comcast.net. 3:07

2005 Nov 13 04:30am 1.781,8.17 0.752 borg-bowser.hsd1.wa.comcast.net. 3:07

2005 Nov 13 04:11am 1.818,8.2 0.749 borg-bowser.hsd1.wa.comcast.net. 3:08

2005 Nov 13 03:29am 0.396,19.46 0.426 borg-bowser.hsd1.wa.comcast.net. 3:11

2005 Nov 13 02:15am 1.658,8.1 0.757 borg-bowser.hsd1.wa.comcast.net. 3:07

2005 Nov 13 01:22am 1.658,8.1 0.757 borg-bowser.hsd1.wa.comcast.net. 3:07

2005 Nov 13 01:02am 0.882,8.86 0.694 borg-bowser.hsd1.wa.comcast.net. 3:04

2005 Nov 13 12:17am 1.542,8.09 0.759 borg-bowser.hsd1.wa.comcast.net. 3:06

another one but this one was during a few updates and a bit of gaming


2005 Nov 12 11:07pm 1.542,8.09 0.759 borg-bowser.hsd1.wa.comcast.net. 3:07

2005 Nov 12 10:15pm 12.408,18.08 0.426 borg-bowser.hsd1.wa.comcast.net. 3:12

2005 Nov 12 09:58pm 12.408,18.08 0.426 borg-bowser.hsd1.wa.comcast.net. 3:12

2005 Nov 12 09:11pm 1.428,8.12 0.756 borg-bowser.hsd1.wa.comcast.net. 3:07

2005 Nov 12 08:00pm 1.39,8.14 0.755 borg-bowser.hsd1.wa.comcast.net. 3:06

2005 Nov 12 07:01pm 1.352,8.17 0.753 borg-bowser.hsd1.wa.comcast.net. 3:07

2005 Nov 12 06:45pm 1.277,8.24 0.747 borg-bowser.hsd1.wa.comcast.net. 3:08


2005 Nov 12 06:02pm 1.169,8.36 0.736 borg-bowser.hsd1.wa.comcast.net. 3:08

2005 Nov 12 04:53pm 0.481,9.83 0.62 borg-bowser.hsd1.wa.comcast.net. 3:12

2005 Nov 12 03:53pm 1.132,8.41 0.732 borg-bowser.hsd1.wa.comcast.net. 3:13

2005 Nov 12 03:37pm 1.096,8.46 0.727 borg-bowser.hsd1.wa.comcast.net. 3:11

2005 Nov 12 02:52pm 1.06,8.52 0.722 borg-bowser.hsd1.wa.comcast.net. 3:09

2005 Nov 12 01:29pm 12.385,18.08 0.426 borg-bowser.hsd1.wa.comcast.net. 3:19

2005 Nov 12 12:32pm 1.06,8.52 0.722 borg-bowser.hsd1.wa.comcast.net. 3:08

2005 Nov 12 12:17pm 0.989,8.64 0.712 borg-bowser.hsd1.wa.comcast.net. 3:10

2005 Nov 12 11:31am 0.954,8.71 0.706 borg-bowser.hsd1.wa.comcast.net. 3:08

2005 Nov 12 09:57am 0.917,8.78 0.7 borg-bowser.hsd1.wa.comcast.net. 3:07

2005 Nov 12 09:18am 0.814,9 0.682 borg-bowser.hsd1.wa.comcast.net. 3:01

2005 Nov 12 08:58am 0.848,8.93 0.688 borg-bowser.hsd1.wa.comcast.net. 3:01

2005 Nov 12 08:14am 0.78,9.08 0.676 borg-bowser.hsd1.wa.comcast.net. 3:01

2005 Nov 12 06:44am 12.363,18.08 0.426 borg-bowser.hsd1.wa.comcast.net. 3:13

2005 Nov 12 06:09am 0.669,9.34 0.656 borg-bowser.hsd1.wa.comcast.net. 3:04

2005 Nov 12 05:49am 0.574,9.59 0.637 borg-bowser.hsd1.wa.comcast.net. 3:10

2005 Nov 12 05:12am 0.574,9.59 0.637 borg-bowser.hsd1.wa.comcast.net. 3:11

2005 Nov 12 03:31am 0.45,9.91 0.614 borg-bowser.hsd1.wa.comcast.net. 3:10

2005 Nov 12 03:05am 0.511,9.75 0.626 borg-bowser.hsd1.wa.comcast.net. 3:09

2005 Nov 12 02:39am 0.481,9.83 0.62 borg-bowser.hsd1.wa.comcast.net. 3:09

2005 Nov 12 02:00am 12.341,18.08 0.426 borg-bowser.hsd1.wa.comcast.net. 3:13

2005 Nov 12 12:21am 12.341,18.08 0.426 borg-bowser.hsd1.wa.comcast.net. 3:13


I wonder why the big difference?