A3 units: Project 6013

miniyazz · 15 Feb 2010 at 22:35

Moan, moan, moan

Project ID: 6013
Core: GRO-A3
Credit: 380
Frames: 100

Name: SMP2 - Windows
Path: C:\FAHSMP\
Number of Frames Observed: 243

Min. Time / Frame : 00:02:56 - 13,122.5 PPD
Avg. Time / Frame : 00:03:07 - 11,981.8 PPD
Cur. Time / Frame : 00:03:22 - 10,601.2 PPD
R3F. Time / Frame : 00:03:23 - 10,549.0 PPD
All Time / Frame : 00:03:23 - 10,549.0 PPD
Eff. Time / Frame : 00:03:27 - 10,345.2 PPD

Anyone else noticed how rubbish these units are, pointswise? Or is it just me?

And, yes I know we all do this for the good of mankind, it's not about the points at all

Biffa · 16 Feb 2010 at 00:42

Bloody whinging i7 owners.

Project ID: 6013
Core: GRO-A3
Credit: 380
Frames: 100

Name: Q6600_2400
Path: c:\Program Files (x86)\Folding@Home Windows SMP Client V1.01
Number of Frames Observed: 100

Min. Time / Frame : 00:09:23 - 2,293.6 PPD
Avg. Time / Frame : 00:09:36 - 2,216.4 PPD

Name: Q6600_3200
Path: d:\fah\cpu\
Number of Frames Observed: 300

Min. Time / Frame : 00:05:46 - 4,760.7 PPD
Avg. Time / Frame : 00:06:25 - 4,056.0 PPD
Cur. Time / Frame : 00:06:34 - 3,927.6 PPD
R3F. Time / Frame : 00:00:00 - 0.0 PPD
All Time / Frame : 00:06:34 - 3,927.6 PPD
Eff. Time / Frame : 00:07:06 - 3,632.6 PPD

verbal · 16 Feb 2010 at 03:02

Biffa said:
Bloody whinging i7 owners.

Hehe

They are slower than the rest though.

Project ID: 6013
Core: GRO-A3
Credit: 380
Frames: 100

Name: Q9300 SMP
Path: G:\apps\Folding@Home Windows SMP Client V1.01\
Number of Frames Observed: 100

Min. Time / Frame : 00:05:10 - 5,613.6 PPD
Avg. Time / Frame : 00:05:40 - 4,887.3 PPD

I remember someone here complaining about the 6015's........ see below for comparison (maybe these 6013s weren't about then..)

Project ID: 6015
Core: GRO-A3
Credit: 484
Frames: 100

Name: Q9300 SMP
Path: G:\apps\Folding@Home Windows SMP Client V1.01\
Number of Frames Observed: 100

Min. Time / Frame : 00:05:06 - 8,151.2 PPD
Avg. Time / Frame : 00:05:20 - 7,622.1 PPD

miniyazz · 16 Feb 2010 at 07:41

Biffa said:
Bloody whinging i7 owners.

lol

That's odd, on mine the 6015's are amongst the highest scoring.

Project ID: 6015
Core: GRO-A3
Credit: 484
Frames: 100

Name: SMP2 - Windows
Path: C:\FAHSMP\
Number of Frames Observed: 300

Min. Time / Frame : 00:02:50 - 19,684.7 PPD
Avg. Time / Frame : 00:03:03 - 17,624.8 PPD
Cur. Time / Frame : 00:03:25 - 15,173.8 PPD
R3F. Time / Frame : 00:03:17 - 15,790.0 PPD
All Time / Frame : 00:03:11 - 16,286.0 PPD
Eff. Time / Frame : 00:03:14 - 16,034.2 PPD

JonJ678 · 16 Feb 2010 at 18:11

I'm thoroughly confused by all these new units, seem to have missed something important here. This thread seems as good a one as any to trash a little.

Under linux, with 12gb of ram available, what cpu client do I want to be running on a 4ghz i7? With no flags in particular it was reporting about 7k ppd, until I broke the work directory by accident. I'm now trying to coax it into bigadv units, but I can't work out if these still exist, or if it's now better to run normal units since some form of quick completion bonus has been put in place. Bit lost really, what should it be doing?

SiriusB · 16 Feb 2010 at 18:59

How much do you use your computer and what for?

If you do a lot of RAM intensive tasks that uses most of your RAM and CPU then you probably shouldn't run bigadv. If you choose to do so it will cost you a massive amount of RAM and you really won't be able to do anything CPU intensive.

You have to crunch those huge WUs almost non-stop to make the deadlines and get the points.

If you actually want to use your computer just go for the SMP client. Much "safer" for the science as an i7 will rip through anything, plus you can actually use your machine for uhm.. stuff

JonJ678 · 17 Feb 2010 at 13:14

Cheers Sirius, clear summary there. The computer goes through stages of being thrashed by cad work and then sitting basically idle for weeks, so the answer is probably to swap between bigadv and normal depending on what it's doing at the time. Given I'm yet to successfully reboot (windows for cad) without breaking the cpu work unit I think this is the most sensible approach.

Thanks for your reply

verbal · 17 Feb 2010 at 13:35

Make sure to use "ctrl-c" to close the cpu client before you reboot.

SiriusB · 17 Feb 2010 at 13:58

Switching between the two is a good idea if you know it will idle for a long periods. When you know you are about to switch, it is a good idea to close the client [Ctrl + C as above] then from the command line restart the client with

Code:

.\path\to\fah -oneunit

This way the client will finish the current WU and won't download anymore once it is done. This avoids you having to wait for the right second to close the client or closing the client on a WU you will never finish - which is bad for the science.

JonJ678 · 17 Feb 2010 at 22:26

Using ctrl+c on the terminal it's running in works brilliantly for gpu clients under wine, but has a pretty poor success rate with cpu units. I'm not sure why, possible that I need to leave it longer between killing the process and rebooting. If I can't work it out over the next half dozen or so reboots I'll post the various errors up on here.

Cheers