Welcome to the MacNN Forums.

If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below.

You are here: MacNN Forums > Community > Team MacNN > Enhanced Optimized

Enhanced Optimized (Page 23)
Thread Tools
QSilver
Dedicated MacNNer
Join Date: Jun 2006
Location: Chicago
Status: Offline
Reply With Quote
Nov 4, 2007, 11:13 PM
 
What am I doing wrong?

I finally got around to downloading the G4-v8 version and every time I try to download WUs, I get computation errors within a second or two on all of them. I've heard that this happens sometimes due to access rights or ownership issues. I checked the access on the Alex's app itself and it looks OK (755). The owner is me and the group was admin. I looked at the 7.1mb folder and everything was the same except for the group. I did a chgrp to 'staff,' but the next WU batch errored out the same way.

What am I missing?

QS
     
Bad to the bone
Fresh-Faced Recruit
Join Date: Aug 2006
Status: Offline
Reply With Quote
Nov 5, 2007, 04:27 AM
 
Happened to me a while ago but can't remember the exact circumstances. Couple of options:

Dump the contents of the S@H folder, put all files from Alex's folder again and see what happens.

If that doesn't help, re-install BOINC and proceed from there.

If all fails, dump the whole BOINC data folder and start all over from scratch.
     
jedimstr
Junior Member
Join Date: Nov 2003
Status: Offline
Reply With Quote
Nov 5, 2007, 09:29 AM
 
Originally Posted by QSilver View Post
What am I doing wrong?

I finally got around to downloading the G4-v8 version and every time I try to download WUs, I get computation errors within a second or two on all of them. I've heard that this happens sometimes due to access rights or ownership issues. I checked the access on the Alex's app itself and it looks OK (755). The owner is me and the group was admin. I looked at the 7.1mb folder and everything was the same except for the group. I did a chgrp to 'staff,' but the next WU batch errored out the same way.

What am I missing?

QS
The user should be boinc_master and the group should be boinc_project.

The easiest way to make sure all the permissions are set correctly in the Boinc Data folder is to reinstall Boinc Manager (there's a new dev version 5.10.28 available at Berkeley).
----------------------------------------------------
Jedi's Lair: Reviews, Tips, and the RickyCam
----------------------------------------------------
Jedi's Photos: Living life one shutter click at a time...
     
beadman
Dedicated MacNNer
Join Date: Nov 2004
Location: Virginia
Status: Offline
Reply With Quote
Nov 5, 2007, 10:08 AM
 
     
zombie67
Senior User
Join Date: Jun 2006
Location: Dublin, CA
Status: Offline
Reply With Quote
Nov 5, 2007, 10:47 AM
 
Originally Posted by jedimstr View Post
The user should be boinc_master and the group should be boinc_project.

The easiest way to make sure all the permissions are set correctly in the Boinc Data folder is to reinstall Boinc Manager (there's a new dev version 5.10.28 available at Berkeley).
I use this method:

(quit BOINC)
cd /Library/Application\ Support/BOINC\ Data/projects/setiathome.berkeley.edu/
sudo chown boinc_master:boinc_project *
(restart BOINC)
     
jedimstr
Junior Member
Join Date: Nov 2003
Status: Offline
Reply With Quote
Nov 5, 2007, 12:33 PM
 
Originally Posted by zombie67 View Post
I use this method:

(quit BOINC)
cd /Library/Application\ Support/BOINC\ Data/projects/setiathome.berkeley.edu/
sudo chown boinc_master:boinc_project *
(restart BOINC)
Yup, that works too
Easiest method, not the only method (easiest in terms of serving the command line challenged as well as the "vi-forever" contingent).
----------------------------------------------------
Jedi's Lair: Reviews, Tips, and the RickyCam
----------------------------------------------------
Jedi's Photos: Living life one shutter click at a time...
     
Lateralus
Moderator Emeritus
Join Date: Sep 2001
Location: Arizona
Status: Offline
Reply With Quote
Nov 6, 2007, 11:38 PM
 
Don't suppose anybody would be up to writing a 'n00b FAQ' on how to get going with SETI BOINC and how to install the optimized client?
I like chicken
I like liver
Meow Mix, Meow Mix
Please de-liv-er
     
reader50
Administrator
Join Date: Jun 2000
Location: California
Status: Offline
Reply With Quote
Nov 7, 2007, 12:26 AM
 
We wrote one awhile back. It's somewhat dated, in that it doesn't consider the modern SETI versions. It does cover most of the steps.

Note, if you use that guide, do NOT install the special BOINC compiles it links to. Those don't work right any more. Today, we use optimized workers instead of optimized BOINCs.
     
arkayn
Dedicated MacNNer
Join Date: Aug 2005
Location: Golden Valley, AZ
Status: Offline
Reply With Quote
Nov 7, 2007, 11:46 AM
 
I PM'd this to a cruncher on the main SETI site.

The easiest way I have found is to shut down the boincmanager, then download the optimized client for your processor and expand it.

Go into your hard drive/Library/Application Support/Boinc Data/Projects/setiathome.berkeley.edu folder and copy in all the files from the zip file.

Next step is to reinstall the boincmanager to repair permissions on the folder. It should take about 30 seconds.

It should start up and start crunching with the optimized worker.
     
Billy
Forum Regular
Join Date: Jun 2006
Location: Toronto, Canada
Status: Offline
Reply With Quote
Nov 7, 2007, 04:23 PM
 
Originally Posted by arkayn View Post
I PM'd this to a cruncher on the main SETI site.
It's not too easy to find his client on the Seti site.
     
Todd Madson
Mac Elite
Join Date: Apr 2000
Location: Minneapolis, MN USA
Status: Offline
Reply With Quote
Nov 13, 2007, 12:37 PM
 
More continued proof of Alex's excellent 8.0 client.

I hit 1105 rac this morning on the G5.

The increase is gradual but noticeable.

Cool.
     
arkayn
Dedicated MacNNer
Join Date: Aug 2005
Location: Golden Valley, AZ
Status: Offline
Reply With Quote
Nov 13, 2007, 10:38 PM
 
It would help if the pending was not so high as well, I figure that my iMac would be over 2000 if not for the 5300 in pending I have right now.
     
TiloProbst
Junior Member
Join Date: Jul 2005
Status: Offline
Reply With Quote
Nov 14, 2007, 08:36 AM
 
My iMac C2D 2,13 GHz hit 1600 RAC today. What can I expect when SETI is not running 100% of the time? I often quit it for gaming or video rendering.
     
beadman
Dedicated MacNNer
Join Date: Nov 2004
Location: Virginia
Status: Offline
Reply With Quote
Nov 14, 2007, 12:21 PM
 
Originally Posted by TiloProbst View Post
My iMac C2D 2,13 GHz hit 1600 RAC today. What can I expect when SETI is not running 100% of the time? I often quit it for gaming or video rendering.
The theoretical RAC is easy to calculate. Take the time in seconds that a given work unit uses, call it "S"; the credit earned for that work unit "C", the number of processors you are using "P".

Potential RAC = P * 86400 / (S/C)

For each hour you are NOT running SETI, subtract 1/24 of the potential RAC.
beadman
     
halimedia
Dedicated MacNNer
Join Date: Oct 2005
Location: Switzerland
Status: Offline
Reply With Quote
Nov 14, 2007, 02:34 PM
 
Originally Posted by beadman View Post
The theoretical RAC is easy to calculate. Take the time in seconds that a given work unit uses, call it "S"; the credit earned for that work unit "C", the number of processors you are using "P".

Potential RAC = P * 86400 / (S/C)

For each hour you are NOT running SETI, subtract 1/24 of the potential RAC.
beadman
Keep in mind that not all WUs are crunched with equal performance. Some have a higher yield in credits/second than others.

HTH,

Ron
     
beadman
Dedicated MacNNer
Join Date: Nov 2004
Location: Virginia
Status: Offline
Reply With Quote
Nov 15, 2007, 10:26 AM
 
Originally Posted by halimedia View Post
Keep in mind that not all WUs are crunched with equal performance. Some have a higher yield in credits/second than others.

HTH,

Ron
Well, if you're going to be picky about it <grin>, Ron, then a sample on work units, say 100 so as to be statistically significant, should be picked, each one run through the formula, and then determine the mean and standard deviation of the result.

Having said that, yes of course. For example, here's a sample of 5 of my WU crunched on my MBP:

rac sec cred
1266 7400 54.23
1588 8226 75.61
1072 4818 29.89
1328 2516 19.33
1087 2634 16.57

This small sample has a mean of 1268 and a standard deviation of 210. So, for the non-statistical person, this just indicates that the best average RAC I can expect is around 1268, plus or minus 210. So if I was really lucky and received mostly the 75 credit WU, I'd have a really high RAC, and if I'm unlucky and receive mostly the 29 credit or 16 credit WU, I'll have really low RAC. Again, this assumes uninterrupted crunching 24/7. Since I carry my MBP back and forth between home and work, and on frequent travel, my current actual RAC of 1121 says I'm below the theoretical average, part of which is due to only running 22 hours per day (reduction of 106) , and part of which is a slightly higher count of low-credit WU.

Now aren't you glad you asked the question? <grin>
beadman
     
Billy
Forum Regular
Join Date: Jun 2006
Location: Toronto, Canada
Status: Offline
Reply With Quote
Nov 15, 2007, 04:24 PM
 
Originally Posted by beadman View Post
Well, if you're going to be picky about it <grin>, Ron, then a sample on work units, say 100 so as to be statistically significant, should be picked, each one run through the formula, and then determine the mean and standard deviation of the result.

Having said that, yes of course. For example, here's a sample of 5 of my WU crunched on my MBP:

rac sec cred
1266 7400 54.23
1588 8226 75.61
1072 4818 29.89
1328 2516 19.33
1087 2634 16.57

This small sample has a mean of 1268 and a standard deviation of 210.
beadman
Just for fun, I deleted A Kan's v8 and ran the "5.28 enhanced" version from the seti site. On my iMac Core Duo the time for a standard 63.98 credit WU went from a time of 9,711.88 to 13,477.11 sec., an increase of 39%. On a non-standard 52.25 credit WU, the "5.28 enhanced" application took an astounding 15,863.07 sec. to complete. Isn't it time that A. Kan's application became the "standard" Macintosh application on Seti? The "result" says something about the application coming from Apple? How does that work?
     
Todd Madson
Mac Elite
Join Date: Apr 2000
Location: Minneapolis, MN USA
Status: Offline
Reply With Quote
Nov 16, 2007, 12:09 PM
 
In the past the Seti folk were not necessarily happy about huge amounts of fast
enhanced crunchers because they only had so many work units to distribute and
thus some work units were crunched multiple times by multiple parties.

With multibeam, I'm not entirely sure if this is an issue anymore. The stock clients
have always been slower/less efficient.

That being said, I'm sure there are people out there who could be achieving
significantly better performance than they are but are unaware that there are
enhanced clients to do so.
     
TiloProbst
Junior Member
Join Date: Jul 2005
Status: Offline
Reply With Quote
Nov 21, 2007, 05:37 PM
 
Sorry, but I couldn't care less about mathematical estimation. I want real world facts. Topped 1702 RAC today, which I think is good because I sometimes kill SETI to do some renderings etc.
     
halimedia
Dedicated MacNNer
Join Date: Oct 2005
Location: Switzerland
Status: Offline
Reply With Quote
Nov 22, 2007, 03:46 AM
 
Originally Posted by TiloProbst View Post
Sorry, but I couldn't care less about mathematical estimation. I want real world facts. Topped 1702 RAC today, which I think is good because I sometimes kill SETI to do some renderings etc.
In my experience, a T7400 can achieve a RAC of slightly over 2000 when crunching full-time.

HTH,

Ron
     
Karl Schimanek
Junior Member
Join Date: Oct 2004
Location: Germany
Status: Offline
Reply With Quote
Nov 24, 2007, 05:34 PM
 
Hello :-)
Is any RAC known of dual processor G4s? Maybe 7448? Dual 1.8GHz?
     
halimedia
Dedicated MacNNer
Join Date: Oct 2005
Location: Switzerland
Status: Offline
Reply With Quote
Nov 25, 2007, 05:27 AM
 
Originally Posted by Karl Schimanek View Post
Hello :-)
Is any RAC known of dual processor G4s? Maybe 7448? Dual 1.8GHz?
I have a G4 Digital Audio with a GigaDesigns Dual 1.33 GHz PPC 7455 upgrade card installed (2x 2MB L3), crunching full-time. With Alex' current worker (and MB-data), it struggles to reach a RAC of over 450. Previously (i.e. with v8 pre-release and LF-data), it hovered around 600. I'd be curious to hear results from any 7448 owners, too...

Cheers,

Ron
     
arkayn
Dedicated MacNNer
Join Date: Aug 2005
Location: Golden Valley, AZ
Status: Offline
Reply With Quote
Dec 16, 2007, 11:43 AM
 
Message from a user

From: Keith (ID 59417)
Subject: Repetitions of part of stderr on Alex's Apps

Arkayn

I don't know whether a Personal Message to you is appropriate, but I know Alex has plenty on his plate to sift through incoming messages.

I have heard from him that the -5000 error will be rectified for G5s on the next update, and assume that this problem is a similar type of problem affecting all the V8 apps by Alex.

There are random (but frequent) duplications in the Task Details of part of stderr text (from start to flopcount), and occasionally triplications (or even more). An example of a triplication follows. I have experienced these problems in tasks for both G4 and G5 but am not sure about MacBookPro. If you can pass on the relevant information to Alex, this might be useful to him before the next update. Don't hesitate to get back to me if you want further information that I may be able to provide. I can test MacBookPro if required, but at present it's only running ClimatePrediction.

Keith

------------------------------------------------------------
Task ID 679588173
Name 23mr07ab.7347.14387.8.6.207_0
Workunit 188959738
Created 11 Dec 2007 10:26:49 UTC
Sent 11 Dec 2007 16:49:26 UTC
Received 15 Dec 2007 18:37:16 UTC
Server state Over
Outcome Success
Client state Done
Exit status 0 (0x0)
Computer ID 4003300
Report deadline 3 Feb 2008 11:16:40 UTC
CPU time 38188.454398
stderr out
<core_client_version>5.10.30</core_client_version>
<![CDATA[
<stderr_txt>
OS X optimized S@H Enhanced application by Alex Kan
Version info: OS X Altivec (G4/G5, G4-optimized v8-nographics) V5.13 by Alex Kan

Work Unit Info:
...............
Credit multiplier is : 2.85
WU true angle range is : 0.400473
OS X optimized S@H Enhanced application by Alex Kan
Version info: OS X Altivec (G4/G5, G4-optimized v8-nographics) V5.13 by Alex Kan

Work Unit Info:
...............
Credit multiplier is : 2.85
WU true angle range is : 0.400473
OS X optimized S@H Enhanced application by Alex Kan
Version info: OS X Altivec (G4/G5, G4-optimized v8-nographics) V5.13 by Alex Kan

Work Unit Info:
...............
Credit multiplier is : 2.85
WU true angle range is : 0.400473

Flopcounter: 16442033648156.503906

Spike count: 7
Pulse count: 0
Triplet count: 0
Gaussian count: 2
</stderr_txt>
]]>
Validate state Valid
Claimed credit 54.2523081632438
Granted credit 54.2523081632438
application version 5.27
I am posting it here for Alex.
     
Gecko_r7
Forum Regular
Join Date: Oct 2005
Location: Las Vegas, NV
Status: Offline
Reply With Quote
Jan 10, 2008, 01:23 AM
 
Ok.....time to fess-up. Who's getting a new V8 Mac Pro?

I know someone is jones'n for one now.....

She's sure gonna be a crunchin' beastie!
     
gorbag
Fresh-Faced Recruit
Join Date: Nov 2001
Location: Milky Way
Status: Offline
Reply With Quote
Jan 10, 2008, 07:09 AM
 
Dude, ordered one yesterday....

Been waiting for new pros for about a year now...

It's the default clock speed; I couldn't justify the extra $ for the slightly faster clock. Now my wife will get my G5 and she can finally retire her old dual usb ibook.
--
Gorbag ("Beren" in Boinc)
     
SciFrog
Dedicated MacNNer
Join Date: Jun 2003
Location: NY
Status: Offline
Reply With Quote
Jan 10, 2008, 11:33 AM
 
I am close to, but I need a display and don't want to buy the current Apple ones...
Please join Team MacNN BOINC and Folding teams!
No more wasted CPU cycles!
     
gorbag
Fresh-Faced Recruit
Join Date: Nov 2001
Location: Milky Way
Status: Offline
Reply With Quote
Jan 11, 2008, 12:05 PM
 
Originally Posted by SciFrog View Post
I am close to, but I need a display and don't want to buy the current Apple ones...
I've been using an HP LP2465 with a MacPro at work for about a year now and have been very happy with it (my employer is pretty much an HP shop, but I was free to buy what I needed; my lab also has some DELL 20" screens bought in 05). You should be able to find it under $600. It has two DVI-I interfaces, and none of the analogue stuff you'd get with the DELL (which is also well-rated - the ultrasharp not the cheaper one they introduced a few months ago). With DELL you have to wait for a sale, though. Note our usage is primarily software development, not color critical work, so I can't say how accurate these monitors are with respect to, say Pantone colors.

There's some pretty good monitor reviews at anandtech.com that might help.
--
Gorbag ("Beren" in Boinc)
     
Thanar
Junior Member
Join Date: Feb 2002
Location: Kozani, Greece, EU
Status: Offline
Reply With Quote
Jan 13, 2008, 03:27 PM
 
First off, Happy New Year, everybody!

I just came back from the states, with a new MacBook in hand. However, no matter how hard I try to setup BOINC on this machine, it doesn't seem to work as it should. As soon as I install the Manager, it periodically hangs for a few minutes, even before attaching to any projects.

I have noticed that these hangs happen each time BOINC Manager tries to make a network connection, be it project attachment, WU download, result upload, or even when I manually ask for a Project Update. This happens to both SETI@home and Rosetta@home, that I've tried so far... After each hang, there is an "exit with zero state" message in the Mesenger's log, and some finished work goes down the drain.

I also noticed that, during those hangs, the kernel_task goes over 5% which seems like a bit suspicious to my eyes. I also notice that in just 15 minutes of uptime, kernel_tasks faults have already build up to 40k! Does this looks right? I mean, on my old iBook it's under 20k after a week's uptime...

Also tried installing 10.5.1 from scratch, to no avail.
Any ideas would be greatly appreciated...

EDIT (new info):
I am also seeing this in the Console:
Jan 13 23:29:13 Thanars-MacBook /usr/sbin/spindump[579]: process 522 is being monitored
Jan 13 23:29:58 Thanars-MacBook /usr/sbin/spindump[579]: process 522 is being no longer being monitored
( Last edited by Thanar; Jan 13, 2008 at 05:32 PM. )
     
Billy
Forum Regular
Join Date: Jun 2006
Location: Toronto, Canada
Status: Offline
Reply With Quote
Jan 14, 2008, 03:25 PM
 
Try repairing your permissions with Disk Utility.
     
Thanar
Junior Member
Join Date: Feb 2002
Location: Kozani, Greece, EU
Status: Offline
Reply With Quote
Jan 14, 2008, 03:35 PM
 
I've been a Mac user for over 10 years. I've done extensive debugging. Already tried lots of possible solutions, e.g. disk/permissions repair, fresh OS install, 10.5 / 10.5.1, on airport / on ethernet, new user, defferent memory, different router. Still no luck at all.

I would be very interested to hear from someone who's installed the latest BOINC Manager on the latest MacBook, set up from scratch directly on 10.5.1. I am suspecting that if there was an older BOINC Manager installed and gets upgraded, there would be no problem.
     
Thanar
Junior Member
Join Date: Feb 2002
Location: Kozani, Greece, EU
Status: Offline
Reply With Quote
Jan 14, 2008, 07:02 PM
 
As a matter of fact, I have opened a thread discussing the possible bug I am facing, in BOINC's dev forums, here. You may want to have a look, there's a little more information...
     
Billy
Forum Regular
Join Date: Jun 2006
Location: Toronto, Canada
Status: Offline
Reply With Quote
Jan 15, 2008, 03:49 PM
 
Try the alpha test Boinc Manager at BOINC alpha test

There was, I believe, something about a memory leak with a previous version.

Bill
     
Billy
Forum Regular
Join Date: Jun 2006
Location: Toronto, Canada
Status: Offline
Reply With Quote
Jan 15, 2008, 03:55 PM
 
Originally Posted by Thanar View Post
As a matter of fact, I have opened a thread discussing the possible bug I am facing, in BOINC's dev forums, here. You may want to have a look, there's a little more information...
I got the same "exit with no result" when I ran Disk Utility, Verify Disk, as I believe that it "dismounts" the disk to run the check. Nothing "bad" happened to those Boinc work units and they continued on without problem. I am assuming that it was the Verify Disk. I also did a Repair Permissions, but I don't think that would dismount the disk.
     
sancio
Fresh-Faced Recruit
Join Date: Dec 2006
Location: Italy
Status: Offline
Reply With Quote
Feb 9, 2008, 10:51 AM
 
Originally Posted by alexkan View Post
O ye of little faith!

No, I didn't forget about all of you, but as time went on, it became easier to just do small-scale (rather than real-world) testing. Also, it eventually became clear to me that if I was going to be a perfectionist about this, I would never wind up releasing anything. So, without further ado...here's v8!

Core Duo
Hi,
I tried just for fun CoreDuo version on an Athlon64 X2 4400+ and it seems to boost respect to the stock client

See here.
Only one error here.
     
arkayn
Dedicated MacNNer
Join Date: Aug 2005
Location: Golden Valley, AZ
Status: Offline
Reply With Quote
Feb 9, 2008, 11:25 AM
 
Got Leopard running on a PC huh.

How much work did it take.
     
sancio
Fresh-Faced Recruit
Join Date: Dec 2006
Location: Italy
Status: Offline
Reply With Quote
Feb 13, 2008, 05:21 PM
 
Originally Posted by arkayn View Post
Got Leopard running on a PC huh.

How much work did it take.
Yes, only a test to see Leopard (on my G3 it doesn't run).
Now there is Linux Slack on the AMD.

It hasn't required a lot of work, about a half day:
I tried 2 "versions",
iATKOS, even if patched for AMD, doesn't work.

However Leopard worked well,
the only system applications that didn't work were X11 and Front Row.
GoogleEarth had some problems, VLC worked well only with Rosetta, BOINC/SETI seemed to running not NICE although ps showed it ok.

But it was far from perfect:
sound card not recognized (audio via Bluetooth headset ok)
video card not recognized: 1024x768 32bit fixed.

To have a truly functional system: EFI emulation with stock kernel and the possibility to perform updates, it need an Intel Core2Duo and at least supported video/sound cards ...
Or better a true MAC
( Last edited by sancio; Feb 13, 2008 at 05:29 PM. )
     
Billy
Forum Regular
Join Date: Jun 2006
Location: Toronto, Canada
Status: Offline
Reply With Quote
Apr 30, 2008, 11:55 AM
 
Originally Posted by Todd Madson View Post
In the past the Seti folk were not necessarily happy about huge amounts of fast
enhanced crunchers because they only had so many work units to distribute and
thus some work units were crunched multiple times by multiple parties.

With multibeam, I'm not entirely sure if this is an issue anymore. The stock clients
have always been slower/less efficient.

That being said, I'm sure there are people out there who could be achieving
significantly better performance than they are but are unaware that there are
enhanced clients to do so.
Seems like there is a bottleneck as you correctly predicted; more work units being generated than the present CPUs can process.

29 Apr 2008 22:08:03 UTC
During today's outage, Jeff and I did yet more reorganization of room 329, culminating in finally, for the first time ever, putting sidious in a rack. This was a major step in filling this particular rack, which will hopefully replace one of the three racks in the closet sooner than later. We also did the steps to rebuild the replica database, which is happening in the background now. May complete tonight or tomorrow, and then it shall "catch up" quickly after that and we'll be back in business on that front.

Clarifying the bottleneck I mentioned yesterday - this is strictly due to our current data processing rate. Drives with raw data come in, which we always archive to off site storage as well as copy into our processing directory (where the splitters read them to make workunits). In a perfect world, we'd be processing data as fast as we archive them, but to do so would require a lot more active users. So frequently our 8 terabyte processing directory fills up with unsplit data, and everything logjams. So this isn't a database bottleneck - it's a data bottleneck. More people/computers is the solution.
     
QSilver
Dedicated MacNNer
Join Date: Jun 2006
Location: Chicago
Status: Offline
Reply With Quote
Apr 30, 2008, 04:48 PM
 
So if the statement "More people/computers is the solution" is true, then having a bunch of people use optimzed apps is also part of the solution.

And, all of the power users with relatively recent Windows rigs (with a minimum of SSE3) are all excited about the port of Alex Kan's v8 app. See announcement.

QS
     
arkayn
Dedicated MacNNer
Join Date: Aug 2005
Location: Golden Valley, AZ
Status: Offline
Reply With Quote
Apr 30, 2008, 06:49 PM
 
Running the new app now.
     
QSilver
Dedicated MacNNer
Join Date: Jun 2006
Location: Chicago
Status: Offline
Reply With Quote
May 1, 2008, 10:19 AM
 
Originally Posted by arkayn View Post
Running the new app now.
I assume you're running that only on your C2D laptop?
     
arkayn
Dedicated MacNNer
Join Date: Aug 2005
Location: Golden Valley, AZ
Status: Offline
Reply With Quote
May 1, 2008, 03:47 PM
 
Originally Posted by QSilver View Post
I assume you're running that only on your C2D laptop?
Actually it is a Inspiron 530S, which is a slim desktop. The laptop just got retired from SETI.
     
Todd Madson
Mac Elite
Join Date: Apr 2000
Location: Minneapolis, MN USA
Status: Offline
Reply With Quote
May 21, 2008, 04:47 PM
 
The G5 did a recent precipitous drop in seti performance using the v8 Alex client.

It's acting as if somehow it uninstalled the Alex client and went back to the stock client.

Is this even possible?

Will research it tonight.
     
BTBlomberg
Forum Regular
Join Date: Sep 2005
Location: Chicago Suburbs
Status: Offline
Reply With Quote
May 22, 2008, 01:42 PM
 
Todd,

Usually that won't happen. Only way would likely be if you are working the stock app already.

Other factors that can do that are the WUs themselves. If they have a bunch of high work/low credit WUs they send out you could get hosed by that. There have always been some of those floating out there. THey may be noisy or some other problem and they take longer to process that they are worth is relative Points. On the other hand there are also WUs that hit the sweet spot and are easy to crunch, but have high points.

From memory only, I think I have seen WUs that get scored with points in the 50s often can take longer to process than ones in the 70s. That may not be an goods analysis of the situation, but there is some WUs that through things off balance and Noisy WUs are often the culprits.

There is a local radar system that often interferes with signal that they are trying to mask/blank out, if they are off on this we could get some bad WUs. Maybe larger patches of them. There can also be other interference they can not account for when batch splitting the tape data.

Check you message area right after startup of Boinc and it should say:
"Found app_info.xml: using anonymous platform"
and the Starting WU lines should end with "using setiathome_enhanced version 527"

I hope this is helpful.
     
Todd Madson
Mac Elite
Join Date: Apr 2000
Location: Minneapolis, MN USA
Status: Offline
Reply With Quote
May 28, 2008, 03:42 PM
 
Presently my G5 2.5 dual is at 690 rac after starting out somewhere over 1100 earlier
this month...the 2.2 ghz core2duo Vista laptop is RAC 1248.

The latter has tasks 368, the former has 66 so doing less too. Most of the C2D
workunits are in the 2000-4000 seconds range. Credit ranges from 50-80.

On the G5:
Looks like they are all 2000-9000 seconds in length, most though are 9000
claiming 75.52 credit. And getting it too. Many are in the 50 range and a few
are in the 17 range.

Here's some text from one of the finished workunits:

<core_client_version>5.5.0</core_client_version>
<stderr_txt>
OS X optimized S@H Enhanced application by Alex Kan
Version info: OS X Altivec (G4/G5, G5-optimized v8-nographics) V5.13 by Alex Kan

Work Unit Info:
...............
Credit multiplier is : 2.85
WU true angle range is : 0.447893

Flopcounter: 15324957707094.111328

Spike count: 4
Pulse count: 0
Triplet count: 5
Gaussian count: 0
</stderr_txt>

I know the larger caches in the C2D help but I wonder if I have a roadblock on
the G5 somewhere slowing things down.
     
Todd Madson
Mac Elite
Join Date: Apr 2000
Location: Minneapolis, MN USA
Status: Offline
Reply With Quote
May 28, 2008, 03:44 PM
 
Oh - links...

Powermac: Computer summary

C2D Vista laptop: Computer summary
     
Gareth Johnston
Junior Member
Join Date: Jan 2008
Status: Offline
Reply With Quote
May 29, 2008, 10:06 AM
 
There is no bottle neck, that is about right, the difference in architecture is what has made the difference. the Core 2 chips are just stunning, the wipe the floor with anything we have from the old G4/G5 days.
     
Todd Madson
Mac Elite
Join Date: Apr 2000
Location: Minneapolis, MN USA
Status: Offline
Reply With Quote
May 29, 2008, 11:14 AM
 
For quite some time the G5 and the C2D machine were running neck and neck both around 1100-1200
rac, then the G5 hit something and just dropped.

I agree with the Core2 architecture assessment. Apple definetely made the right decision on
switching to Intel. Insanity.
     
Gareth Johnston
Junior Member
Join Date: Jan 2008
Status: Offline
Reply With Quote
Jun 2, 2008, 05:46 AM
 
When the G5's were new we had a different credit multiplier for Seti, not sure if it was 3.25 or 3.15 but that has gone down to 2.85 at present. If your using the new PC optimised apps then you are taking full advantage of what the new chips can offer.

Alex's PPC app is stunning but the instruction set's and Alti-vec cannot compete in any way shape or form with the C2 chips.

Such a significant drop in RAC is a little odd though.

Did you make any changes to the system, is it running any different?
     
BTBlomberg
Forum Regular
Join Date: Sep 2005
Location: Chicago Suburbs
Status: Offline
Reply With Quote
Jun 18, 2008, 11:55 AM
 
Todd,

I just wanted to check in on something that just recently happened with the Mac Pro 2.8 Dual Quad. For some unknown reason it went from around 9500 credits average to 7500 credits average with no changes to the machine. It did have a mess of WUs that were getting low scores though. I have seen this before at different times with SETI, but like I said before it seams to be the WUs and much less about the machine.

Do you still have the hit to your G5s RAC?
     
arkayn
Dedicated MacNNer
Join Date: Aug 2005
Location: Golden Valley, AZ
Status: Offline
Reply With Quote
Jun 19, 2008, 12:14 AM
 
I always keep an eye on my efficiency score as that can throw the RAC for a loop as well.
     
 
Thread Tools
 
Forum Links
Forum Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Top
Privacy Policy
All times are GMT -4. The time now is 05:50 PM.
All contents of these forums © 1995-2017 MacNN. All rights reserved.
Branding + Design: www.gesamtbild.com
vBulletin v.3.8.8 © 2000-2017, Jelsoft Enterprises Ltd.,