T-FLOPS jump??

Message boards : Number crunching : T-FLOPS jump??

To post messages, you must log in.

AuthorMessage
FoldingSolutions
Avatar

Send message
Joined: 2 Apr 06
Posts: 129
Credit: 3,506,690
RAC: 0
Message 45935 - Posted: 10 Sep 2007, 15:11:19 UTC

Is it just me or has Rosetta T-FLOPS estimate jumped from ~50 before the crash to about 85 afterwards! could this be due to the influx of results once it came back online or have people re-doubled their efforts?
ID: 45935 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Christoph Jansen
Avatar

Send message
Joined: 6 Jun 06
Posts: 248
Credit: 267,153
RAC: 0
Message 45937 - Posted: 10 Sep 2007, 15:25:35 UTC - in response to Message 45935.  

Is it just me or has Rosetta T-FLOPS estimate jumped from ~50 before the crash to about 85 afterwards! could this be due to the influx of results once it came back online or have people re-doubled their efforts?


It's due to incoming results genearaed during the project's downtime.

The Teraflops estimate on the main page is not a long-time moving average like RAC but rather a daily value. It always correlates directly to the value given under "Credits last 24h". Right now it says:

Credits last 24h : 8,446,053

TeraFLOPS estimate: 84.461

In other words: Credits during last 24 hours divided by 100.000 gives you the current TFlops.
ID: 45937 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Mod.Sense
Volunteer moderator

Send message
Joined: 22 Aug 06
Posts: 4018
Credit: 0
RAC: 0
Message 45944 - Posted: 10 Sep 2007, 17:23:23 UTC

Users that keep more then a day of tasks on hand, crunched through all of them and then reported them all back on the same day.

Unfortunately, it will probably drop back to around 60 again (where it was before the outage) very soon.
Rosetta Moderator: Mod.Sense
ID: 45944 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
FoldingSolutions
Avatar

Send message
Joined: 2 Apr 06
Posts: 129
Credit: 3,506,690
RAC: 0
Message 45949 - Posted: 10 Sep 2007, 19:05:14 UTC - in response to Message 45944.  

Unless someone figures out a way to utilize ATI's R600 boards. At 475 G-FLOPs a hit a 2 GPU system could knock out the best part of a T-FLOP. Get a hundred of them online you have yorselves a tripling in power!
ID: 45949 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile JChojnacki
Avatar

Send message
Joined: 17 Sep 05
Posts: 71
Credit: 9,921,016
RAC: 2,359
Message 45962 - Posted: 10 Sep 2007, 22:10:05 UTC
Last modified: 10 Sep 2007, 22:11:45 UTC

Still, it's kind of cool to see this number, as of 6:08 PM EDT.
TeraFLOPS estimate: 99.998

Hopefully we will be able to get it to stay at that level someday soon.

~Joel

ID: 45962 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
[RKN] schatten1411 , Mitglied des Teams und des VEREINS Rechenkraft.net

Send message
Joined: 25 Apr 07
Posts: 12
Credit: 441,995
RAC: 0
Message 46132 - Posted: 13 Sep 2007, 16:12:06 UTC - in response to Message 45937.  


Credits last 24h : 8,446,053

TeraFLOPS estimate: 84.461



Realy ?? Claimed or granted ?
ID: 46132 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Evan

Send message
Joined: 23 Dec 05
Posts: 268
Credit: 402,585
RAC: 0
Message 46175 - Posted: 14 Sep 2007, 7:57:46 UTC

At what time of the day does the last 24 hrs begin. It always seems to be rising rather than falling and restarting.
ID: 46175 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Betting Slip

Send message
Joined: 26 Sep 05
Posts: 71
Credit: 5,702,246
RAC: 0
Message 46514 - Posted: 18 Sep 2007, 12:02:59 UTC - in response to Message 46175.  

At what time of the day does the last 24 hrs begin. It always seems to be rising rather than falling and restarting.


It's a rolling 24hrs so the figure you see is always the last 24hrs
ID: 46514 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile rochester new york
Avatar

Send message
Joined: 2 Jul 06
Posts: 2842
Credit: 2,020,043
RAC: 0
Message 46521 - Posted: 18 Sep 2007, 14:20:11 UTC

looks like the t flops are going backwards :[
ID: 46521 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Greg_BE
Avatar

Send message
Joined: 30 May 06
Posts: 5659
Credit: 5,694,012
RAC: 1,911
Message 46535 - Posted: 18 Sep 2007, 17:25:01 UTC - in response to Message 46521.  

probably due to the CAPRI failure and 5.78 hanging up with CAPRI.
looks like the t flops are going backwards :[


ID: 46535 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Azurrio

Send message
Joined: 20 Feb 06
Posts: 8
Credit: 237,979
RAC: 0
Message 46834 - Posted: 23 Sep 2007, 8:59:52 UTC
Last modified: 23 Sep 2007, 9:02:59 UTC

Did a little research and didn't find anything related to this. I also didn't want to make a own topic for this one so going to post it here:

Is it possible to make somekind of graphic about the T-flops? I know that the amount of credits is straightly(?) related to the amount of t-flops but it would be nice to see how the calculation speed has developed say, in the last 6 months or in the past year or so...
I got the idea from einstein@home where they already have this kind of graph-> http://homepage.hispeed.ch/einstein/flops.png

Einstein@home t-flops graph
ID: 46834 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Mod.Sense
Volunteer moderator

Send message
Joined: 22 Aug 06
Posts: 4018
Credit: 0
RAC: 0
Message 46850 - Posted: 23 Sep 2007, 12:42:43 UTC

Take your pick, there are many statistics sites that present this information.
http://boinc.netsoft-online.com/e107_plugins/boinc/bp.php?project=6
http://www.boincstats.com/charts/chart_uk_rosetta_project_new_credits.gif

You can always go to the Rosetta homepage and click the little bar chart icons over on the top right in the server status box.
Rosetta Moderator: Mod.Sense
ID: 46850 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : Number crunching : T-FLOPS jump??



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