Discussion of the new credit systen (2)

Message boards : Number crunching : Discussion of the new credit systen (2)

To post messages, you must log in.

1 · 2 · 3 · 4 . . . 8 · Next

AuthorMessage
Mod.DE
Volunteer moderator

Send message
Joined: 23 Aug 06
Posts: 78
Credit: 0
RAC: 0
Message 25239 - Posted: 28 Aug 2006, 9:34:33 UTC

This thread is a continuation of the dicussion of the new credit system. All prior posts can be found here.
I am a forum moderator! Am I?
ID: 25239 · Rating: 0 · rate: Rate + / Rate - Report as offensive
NJMHoffmann

Send message
Joined: 17 Dec 05
Posts: 45
Credit: 45,891
RAC: 0
Message 25244 - Posted: 28 Aug 2006, 9:53:05 UTC - in response to Message 25239.  
Last modified: 28 Aug 2006, 9:53:33 UTC

This thread is a continuation of the dicussion of the new credit system. All prior posts can be found here.

@Mod.DE (OT):Why do you think, you must split a thread because of "modem users"? They will have their display settings in the "message board preferences" set to fit their needs.
ID: 25244 · Rating: 0 · rate: Rate + / Rate - Report as offensive
Mod.DE
Volunteer moderator

Send message
Joined: 23 Aug 06
Posts: 78
Credit: 0
RAC: 0
Message 25247 - Posted: 28 Aug 2006, 10:15:35 UTC - in response to Message 25244.  

This thread is a continuation of the dicussion of the new credit system. All prior posts can be found here.

@Mod.DE (OT):Why do you think, you must split a thread because of "modem users"? They will have their display settings in the "message board preferences" set to fit their needs.

Well that might be but I doubt many know about those settings. Mod9 used to split long threads in order to make them better organized and it was welcomed here. So I thought I'll do the same. If people feel we should not split I can revert that, but nothing is hidden or deleted it's just a continuation and all former posts stay visible.
I am a forum moderator! Am I?
ID: 25247 · Rating: 0 · rate: Rate + / Rate - Report as offensive
Astro
Avatar

Send message
Joined: 2 Oct 05
Posts: 987
Credit: 500,253
RAC: 0
Message 25248 - Posted: 28 Aug 2006, 10:17:49 UTC
Last modified: 28 Aug 2006, 10:21:33 UTC

Try being a dial up user. I was, the best you can do is set to display "most recent post first" and then hit the stop button on the browser to avoid waiting 3 minutes for all the old (already read) posts to load. And heaven forbid they click on one of my threads with all the pics. When posts counts exceed 300 it can take upwards of 4-5 minutes to load an entire thread.

Man, I'm glad I moved to a place with dsl. LOL

[edit]that said, I do think 170 is a bit low, especially in a thread without graphics and not many sigs, but it's his call
ID: 25248 · Rating: 0 · rate: Rate + / Rate - Report as offensive
NJMHoffmann

Send message
Joined: 17 Dec 05
Posts: 45
Credit: 45,891
RAC: 0
Message 25249 - Posted: 28 Aug 2006, 10:30:58 UTC - in response to Message 25248.  
Last modified: 28 Aug 2006, 10:37:19 UTC

mmciastro:
Try being a dial up user. I was, the best you can do is set to display "most recent post first" and then hit the stop button on the browser to avoid waiting 3 minutes for all the old (already read) posts to load.
There is the "if a thread contains more than ... display only the first and the last ... postings" setting.
[edit] Some projects, with a very old forum code, still don't have it.[/edit]
ID: 25249 · Rating: 0 · rate: Rate + / Rate - Report as offensive
Astro
Avatar

Send message
Joined: 2 Oct 05
Posts: 987
Credit: 500,253
RAC: 0
Message 25251 - Posted: 28 Aug 2006, 10:34:46 UTC - in response to Message 25249.  

There is the "if a thread contains more than ... display only the first and the last ... postings" setting.

eek, forgot about that.
ID: 25251 · Rating: -1 · rate: Rate + / Rate - Report as offensive
Profile Angus

Send message
Joined: 17 Sep 05
Posts: 412
Credit: 321,053
RAC: 0
Message 25267 - Posted: 28 Aug 2006, 14:15:47 UTC

Completely unnecessary -

Learn to use the forum preferences:

Display and Behaviour section -

100 If a thread contains more than this number of posts
75 only display the first one and this many of the last ones

And modem users should also have avatars and signatures turned off, and use the 'Show images as links' option.
Proudly Banned from Predictator@Home and now Cosmology@home as well. Added SETI to the list today. Temporary ban only - so need to work harder :)



"You can't fix stupid" (Ron White)
ID: 25267 · Rating: 0 · rate: Rate + / Rate - Report as offensive
zombie67 [MM]
Avatar

Send message
Joined: 11 Feb 06
Posts: 316
Credit: 6,621,003
RAC: 0
Message 25274 - Posted: 28 Aug 2006, 15:12:32 UTC - in response to Message 25232.  
Last modified: 28 Aug 2006, 15:12:57 UTC

1) wonder woman has been consistantly (slightly) faster than batman crunching SETI. Yet it is earning almost HALF the credit of slower batman.
This looks odd but I don't know nothing about OSX. Perhaps the difference in the OS version is responsible?


Nope. Like I said, with SETI, wonder woman was slightly faster. Same versions of OS as now. This change in performance started with the switch to the new credit system, as far as I can tell.

3) What the heck is rogue doing below *all* the P4s??? Something is not right here. I thought athlons were supposed to stomp all over P4s.
This is really strange and points to a problem with your machine. You should get twice the number. As others have pointed it it looks if two instances would run on the same core thus each getting only half the CPU power. Then your numbers make sense. Please check, whether it might be both instances uses the same core, or try to install two instances of BOINC, each bound ot one core. But this does not indicate any problem with the new credit system, since similar X2 4200+ get exactly twice the credits (under Linux and Windows).


If I had two instances running, I would be getting half the credits per job, but twice the number of jobs in a day. That is not happening. I just added up the past 4 full days: 314, 312, 322, 345. That averages 13.5 credits/hour, which is slightly better than the last 5 jobs alone indicated. But it still puts is slower than all but the slowest pentium 4.

I'm stumped.



Reno, NV
Team: SETI.USA
ID: 25274 · Rating: 0 · rate: Rate + / Rate - Report as offensive
tralala

Send message
Joined: 8 Apr 06
Posts: 376
Credit: 581,806
RAC: 0
Message 25283 - Posted: 28 Aug 2006, 16:04:50 UTC - in response to Message 25274.  


If I had two instances running, I would be getting half the credits per job, but twice the number of jobs in a day. That is not happening. I just added up the past 4 full days: 314, 312, 322, 345. That averages 13.5 credits/hour, which is slightly better than the last 5 jobs alone indicated. But it still puts is slower than all but the slowest pentium 4.
I'm stumped.

It looks like you are getting twice the number of jobs (8 six hour jobs per day) for which you get half the credit of comparable machines.

Look at this host. It has the same specs and gets about 100 credits per 6 hour job (might be slightly overclocked).
ID: 25283 · Rating: -1 · rate: Rate + / Rate - Report as offensive
zombie67 [MM]
Avatar

Send message
Joined: 11 Feb 06
Posts: 316
Credit: 6,621,003
RAC: 0
Message 25293 - Posted: 28 Aug 2006, 17:45:25 UTC - in response to Message 25283.  


It looks like you are getting twice the number of jobs (8 six hour jobs per day) for which you get half the credit of comparable machines.

Look at this host. It has the same specs and gets about 100 credits per 6 hour job (might be slightly overclocked).


I'm not following. For the 27th, I returned 7 jobs (I was just finishing transitioning from 8 to 6 hour jobs). I looked at the machine you referenced. It had 8 jobs on the 27th too. Seems to me they pretty much mach with regard to number of jobs.

Besides the difference in credits, I am seeing no other difference....except for the OS of course.

I checked the processes. There are only the two rosetta application running, each at about 100%, which is right for a dual core processor. For what it's worth, it looks identical to the ps output from the Pentium D machines.

Reno, NV
Team: SETI.USA
ID: 25293 · Rating: 0 · rate: Rate + / Rate - Report as offensive
tralala

Send message
Joined: 8 Apr 06
Posts: 376
Credit: 581,806
RAC: 0
Message 25295 - Posted: 28 Aug 2006, 17:49:39 UTC - in response to Message 25293.  


It looks like you are getting twice the number of jobs (8 six hour jobs per day) for which you get half the credit of comparable machines.

Look at this host. It has the same specs and gets about 100 credits per 6 hour job (might be slightly overclocked).


I'm not following. For the 27th, I returned 7 jobs (I was just finishing transitioning from 8 to 6 hour jobs). I looked at the machine you referenced. It had 8 jobs on the 27th too. Seems to me they pretty much mach with regard to number of jobs.

Besides the difference in credits, I am seeing no other difference....except for the OS of course.

I checked the processes. There are only the two rosetta application running, each at about 100%, which is right for a dual core processor. For what it's worth, it looks identical to the ps output from the Pentium D machines.



Try to set this host to a new venue, let's say "school" and set number of processors to use to 1 so that only one instance is running. Then let it crunch 24 hours and see how much credit you receive for 6hour WU. I'm sure we will narrow this down in the end.
ID: 25295 · Rating: 0 · rate: Rate + / Rate - Report as offensive
Marky-UK

Send message
Joined: 1 Nov 05
Posts: 73
Credit: 1,689,495
RAC: 0
Message 25297 - Posted: 28 Aug 2006, 18:18:50 UTC - in response to Message 25295.  
Last modified: 28 Aug 2006, 18:20:33 UTC

Try to set this host to a new venue, let's say "school" and set number of processors to use to 1 so that only one instance is running. Then let it crunch 24 hours and see how much credit you receive for 6hour WU. I'm sure we will narrow this down in the end.

If it's BOINC 5.4+ you don't even need do to that - you can override individual global preferences at the host.

Create a file called global_prefs_override.xml in the BOINC directory with the following contents:

<global_preferences>
<max_cpus>1</max_cpus>
</global_preferences>

Then stop & start BOINC (or use 'boinccmd --read_global_prefs_override' or whatever the equivalent is on your OS).

To reverse this, delete the file and restart BOINC.
ID: 25297 · Rating: 0 · rate: Rate + / Rate - Report as offensive
BennyRop

Send message
Joined: 17 Dec 05
Posts: 555
Credit: 140,800
RAC: 0
Message 25298 - Posted: 28 Aug 2006, 18:19:04 UTC

Zombie: You're getting 1716.12 new credits for 888,579.83 seconds; or 166.86 credits per 86,400 seconds. You should be getting the 332 (perhaps a little more) per 86,400 seconds, and producing 172,800 seconds of Rosetta work a day while using 2 cores at 100%.

How does Linux (Ubuntu in particular) treat 24/7 tasks running side by side with other tasks of equal priority in the Linux version of Task manager? Do you get 86,400 seconds a day with 50% of the cpu power, or 43,200 seconds a day with 100% of the cpu power?

Where/how do you check which cpu core an application has been bound to?
And are there multiple locations for viewing low power mode settings?


ID: 25298 · Rating: 0 · rate: Rate + / Rate - Report as offensive
zombie67 [MM]
Avatar

Send message
Joined: 11 Feb 06
Posts: 316
Credit: 6,621,003
RAC: 0
Message 25299 - Posted: 28 Aug 2006, 18:29:11 UTC - in response to Message 25295.  

Try to set this host to a new venue, let's say "school" and set number of processors to use to 1 so that only one instance is running. Then let it crunch 24 hours and see how much credit you receive for 6hour WU. I'm sure we will narrow this down in the end.


Thanks.

I'll give that a try tomorrow. For now, I went into grub and made sure that it was using the SMP kernel to boot, rather than some of the others listed. We'll see if that makes a difference.

I'm wondering if there is some additional linux package I need to add? I have the SMP kernel added, of course. Otherwise it would be running only one rosetta application.

Reno, NV
Team: SETI.USA
ID: 25299 · Rating: 0 · rate: Rate + / Rate - Report as offensive
zombie67 [MM]
Avatar

Send message
Joined: 11 Feb 06
Posts: 316
Credit: 6,621,003
RAC: 0
Message 25304 - Posted: 28 Aug 2006, 18:49:52 UTC - in response to Message 25298.  

How does Linux (Ubuntu in particular) treat 24/7 tasks running side by side with other tasks of equal priority in the Linux version of Task manager? Do you get 86,400 seconds a day with 50% of the cpu power, or 43,200 seconds a day with 100% of the cpu power?


I am running nothing else on that machine, so other than the OS, there should be nothing with equal or higher priority. Assording to "ps" there is nothing else consuming any of the CPU%.

Where/how do you check which cpu core an application has been bound to?
And are there multiple locations for viewing low power mode settings?


I don't know how to tell which application is bound to which core.

I will double check the BIOS for low power settings. Not sure where that would be in the OS. I'll look around.

Thanks,
Erik


Reno, NV
Team: SETI.USA
ID: 25304 · Rating: 0 · rate: Rate + / Rate - Report as offensive
zombie67 [MM]
Avatar

Send message
Joined: 11 Feb 06
Posts: 316
Credit: 6,621,003
RAC: 0
Message 25315 - Posted: 28 Aug 2006, 19:29:16 UTC

DING! Found it. There is a gnome tool called CPU Frequency Scaling Monitor. I fired it up and, sure enough, both CPUs are running at 50%. What the...? The help document says I should be able to adjust this in the preferences, but that options isn't there. So I dug around on google, and found the answer here.

Turns out that this feature is turned off by default. The command to enable that deature is "sudo dpkg-reconfigure gnome-applets".

I'm running both CPUs at 2.2ghz now! I set the run time to one hour to quickly see the results.

I wonder how many AMD/linux machines are sitting there running at 50% without the owners realizing? You can't tell by looking at the process monitor or "ps".

Anyway, that for the help everyone!
Reno, NV
Team: SETI.USA
ID: 25315 · Rating: 0 · rate: Rate + / Rate - Report as offensive
Profile dag
Avatar

Send message
Joined: 16 Dec 05
Posts: 106
Credit: 1,000,020
RAC: 0
Message 25323 - Posted: 28 Aug 2006, 20:10:00 UTC - in response to Message 25315.  

DING! ...
I wonder how many AMD/linux machines are sitting there running at 50% without the owners realizing? You can't tell by looking at the process monitor or "ps".

Anyway, that for the help everyone!

Try running top().
dag
--Finding aliens is cool, but understanding the structure of proteins is useful.
ID: 25323 · Rating: 0 · rate: Rate + / Rate - Report as offensive
zombie67 [MM]
Avatar

Send message
Joined: 11 Feb 06
Posts: 316
Credit: 6,621,003
RAC: 0
Message 25324 - Posted: 28 Aug 2006, 20:34:02 UTC - in response to Message 25323.  

DING! ...
I wonder how many AMD/linux machines are sitting there running at 50% without the owners realizing? You can't tell by looking at the process monitor or "ps".

Anyway, that for the help everyone!

Try running top().


Like "ps", it shows the two processes, at 99-100% each, even when the CPUs are running at half speed. Or am I missing something?


Reno, NV
Team: SETI.USA
ID: 25324 · Rating: 0 · rate: Rate + / Rate - Report as offensive
Profile [B^S] thierry@home
Avatar

Send message
Joined: 17 Sep 05
Posts: 182
Credit: 281,902
RAC: 0
Message 25329 - Posted: 28 Aug 2006, 20:50:50 UTC

I have a Pentium D for a few weeks. Is it the same problem? The Task Manager shows both CPUs at 50%. But I said one chip, two cores, 50% each ....seems OK. But is it OK?
Thanks
ID: 25329 · Rating: 0 · rate: Rate + / Rate - Report as offensive
Ethan
Volunteer moderator

Send message
Joined: 22 Aug 05
Posts: 286
Credit: 9,304,700
RAC: 0
Message 25330 - Posted: 28 Aug 2006, 20:52:48 UTC - in response to Message 25329.  

It's not the same. . windows goes by total cpu available. If you max out one cpu in a dual system it will show 50% cpu utilization. The same is true with hyperthreaded cpu's.
ID: 25330 · Rating: 0 · rate: Rate + / Rate - Report as offensive
1 · 2 · 3 · 4 . . . 8 · Next

Message boards : Number crunching : Discussion of the new credit systen (2)



©2024 University of Washington
https://www.bakerlab.org