Wild fluctuations in project's TeraFLOPS

Message boards : Number crunching : Wild fluctuations in project's TeraFLOPS

To post messages, you must log in.

AuthorMessage
Profile Dimitris Hatzopoulos

Send message
Joined: 5 Jan 06
Posts: 336
Credit: 80,939
RAC: 0
Message 11693 - Posted: 6 Mar 2006, 0:06:21 UTC

I like to watch the TeraFLOPS info on R@h's homepage, which sadly has been "stuck" at 19-21 TeraFLOPS for the past 2 months.

Last Friday I noticed it jump to 26TF and right now I see it's back to 18TF.

Considering the rollercoaster, I wonder how this statistic is calculated.

Also, maybe applying some moving average over hourly stats would smooth things out?


Best UFO Resources
Wikipedia R@h
How-To: Join Distributed Computing projects that benefit humanity
ID: 11693 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
David Baker
Volunteer moderator
Project administrator
Project developer
Project scientist

Send message
Joined: 17 Sep 05
Posts: 705
Credit: 559,847
RAC: 0
Message 11697 - Posted: 6 Mar 2006, 2:58:41 UTC - in response to Message 11693.  

I like to watch the TeraFLOPS info on R@h's homepage, which sadly has been "stuck" at 19-21 TeraFLOPS for the past 2 months.

Last Friday I noticed it jump to 26TF and right now I see it's back to 18TF.

Considering the rollercoaster, I wonder how this statistic is calculated.

Also, maybe applying some moving average over hourly stats would smooth things out?



I've noticed this as well. I was excited about the jump to 26TF and thought it was because of going to shorter wu's, but apparently not. could there be a slow down on the weekend for some reason?
ID: 11697 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Moderator9
Volunteer moderator

Send message
Joined: 22 Jan 06
Posts: 1014
Credit: 0
RAC: 0
Message 11699 - Posted: 6 Mar 2006, 3:39:44 UTC - in response to Message 11697.  

I like to watch the TeraFLOPS info on R@h's homepage, which sadly has been "stuck" at 19-21 TeraFLOPS for the past 2 months.

Last Friday I noticed it jump to 26TF and right now I see it's back to 18TF.

Considering the rollercoaster, I wonder how this statistic is calculated.

Also, maybe applying some moving average over hourly stats would smooth things out?



I've noticed this as well. I was excited about the jump to 26TF and thought it was because of going to shorter wu's, but apparently not. could there be a slow down on the weekend for some reason?


The server was down for a while today (Sunday). However, depending on how these figures are calculated this should not have reduced the total flops available. It will of course reduce the results figures.

Moderator9
ROSETTA@home FAQ
Moderator Contact
ID: 11699 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile anders n

Send message
Joined: 19 Sep 05
Posts: 403
Credit: 537,991
RAC: 0
Message 11708 - Posted: 6 Mar 2006, 12:34:30 UTC

Seti was down for 3-4 days last week.

Is it possibel that it had some effect on Rosetta?

Anders n
ID: 11708 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
doc :)

Send message
Joined: 4 Oct 05
Posts: 47
Credit: 1,106,102
RAC: 0
Message 11729 - Posted: 6 Mar 2006, 18:14:11 UTC

if you take a close look youll see that the TeraFLOPS estimate is allways exactly the same as Credits last 24h, just with the decimal point (or whatever the right word is, cant remember :D) moved some places.
ID: 11729 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : Number crunching : Wild fluctuations in project's TeraFLOPS



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