TERAFLOPS

Message boards : Rosetta@home Science : TERAFLOPS

To post messages, you must log in.

1 · 2 · Next

AuthorMessage
Profile rochester new york
Avatar

Send message
Joined: 2 Jul 06
Posts: 2842
Credit: 2,020,043
RAC: 0
Message 23084 - Posted: 18 Aug 2006, 17:54:22 UTC

anyone know how rosetta went fron 37 tflops to 47 t flops in only 2 days //where there that many more computers on board????
ID: 23084 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile scsimodo

Send message
Joined: 17 Sep 05
Posts: 93
Credit: 946,359
RAC: 0
Message 23091 - Posted: 18 Aug 2006, 18:31:21 UTC - in response to Message 23084.  

anyone know how rosetta went fron 37 tflops to 47 t flops in only 2 days //where there that many more computers on board????


Maybe the "Predictor effect"? Predictor@home closes for an unknown time, maybe we see some switchers increasing the TFlops. Seems reasonable, the number of new users/hosts increased too.


ID: 23091 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Gerry Rough
Avatar

Send message
Joined: 2 Jan 06
Posts: 111
Credit: 1,389,340
RAC: 0
Message 23121 - Posted: 18 Aug 2006, 19:51:43 UTC

Myself and a couple of others went over and gave Rosetta a boost as a similar project for those who are interested in protein prediction. Perhaps it worked better than expected.

(Click for detailed stats)
ID: 23121 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Ethan
Volunteer moderator

Send message
Joined: 22 Aug 05
Posts: 286
Credit: 9,304,700
RAC: 0
Message 23124 - Posted: 18 Aug 2006, 19:58:59 UTC - in response to Message 23121.  

I'm waiting a few days to see if it levels out. The server was down for half a day earlier this week, it could be a combination of 36 hours worth of credits being reported in 24. . and boinc users who couldn't get new rosetta wu's had a defecit to fill if they run multiple projects.
ID: 23124 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Avi

Send message
Joined: 2 Aug 06
Posts: 58
Credit: 95,619
RAC: 0
Message 23130 - Posted: 18 Aug 2006, 20:12:19 UTC

Maybe its being reported off the new work unit stats? Both of my stock clients/machines are getting upto double credit on work units.
ID: 23130 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile scsimodo

Send message
Joined: 17 Sep 05
Posts: 93
Credit: 946,359
RAC: 0
Message 23140 - Posted: 18 Aug 2006, 20:31:18 UTC - in response to Message 23124.  
Last modified: 18 Aug 2006, 20:32:21 UTC

I'm waiting a few days to see if it levels out. The server was down for half a day earlier this week, it could be a combination of 36 hours worth of credits being reported in 24. . and boinc users who couldn't get new rosetta wu's had a defecit to fill if they run multiple projects.


Predictor@Home closes for an undetermined time, there may be a lot of switchers heading to Rosetta. Users/Hosts in the last 24h is raising too...


ID: 23140 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Charles Dennett
Avatar

Send message
Joined: 27 Sep 05
Posts: 102
Credit: 2,070,914
RAC: 0
Message 23152 - Posted: 18 Aug 2006, 20:56:14 UTC - in response to Message 23124.  

I'm waiting a few days to see if it levels out. The server was down for half a day earlier this week, it could be a combination of 36 hours worth of credits being reported in 24. . and boinc users who couldn't get new rosetta wu's had a defecit to fill if they run multiple projects.


In the past when the project has been offline for a few hours, the following day or so has seen as rather large jump in the TF rate. It has always taken a larger jump than one might expect. At least that's been my observation.

Charlie

-Charlie
ID: 23152 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Gerry Rough
Avatar

Send message
Joined: 2 Jan 06
Posts: 111
Credit: 1,389,340
RAC: 0
Message 23411 - Posted: 19 Aug 2006, 13:04:45 UTC

This thread reminds me that there has been a burning question that I keep forgetting to ask. What is the difference between teraflops and terahertz. Hertz are, of course, the standars measurement of cycles per second that a processor can perform. Is there a difference?

(Click for detailed stats)
ID: 23411 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Vester
Avatar

Send message
Joined: 2 Nov 05
Posts: 257
Credit: 2,520,276
RAC: 6
Message 23421 - Posted: 19 Aug 2006, 14:04:53 UTC
Last modified: 19 Aug 2006, 14:05:23 UTC

Gerry Rough, see http://en.wikipedia.org/wiki/FLOPS. The clock speed of a processor is not an accurate indicator of floating point operation performance.
ID: 23421 · 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 23438 - Posted: 19 Aug 2006, 15:34:37 UTC

looks like its down to only 31 flops now .few weeks ago it held to about 36-38
ID: 23438 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
R.L. Casey

Send message
Joined: 7 Jun 06
Posts: 91
Credit: 2,728,885
RAC: 0
Message 24884 - Posted: 25 Aug 2006, 18:32:58 UTC - in response to Message 23438.  

looks like its down to only 31 flops now .few weeks ago it held to about 36-38


It's been climbing steadily, and today is running at between 41 and 43 Tflops! :-)
ID: 24884 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
tralala

Send message
Joined: 8 Apr 06
Posts: 376
Credit: 581,806
RAC: 0
Message 24954 - Posted: 26 Aug 2006, 9:02:33 UTC

The server status on the main page of Rosetta is not reliable. It fluctuates a lot. I think the best overview and the most accurate and up-to-date information is delivered by free-dc:

http://stats.free-dc.org/new/allproj.php?proj=rah

This shows the average was about 37-38 before the credit change and now seems to stabilize around 35-36 with the new credit system.
ID: 24954 · Rating: 2 · rate: Rate + / Rate - Report as offensive    Reply Quote
STE\/E

Send message
Joined: 17 Sep 05
Posts: 125
Credit: 3,226,630
RAC: 915
Message 24963 - Posted: 26 Aug 2006, 10:24:21 UTC - in response to Message 23140.  
Last modified: 26 Aug 2006, 10:24:33 UTC

Predictor@Home closes for an undetermined time, there may be a lot of switchers heading to Rosetta. Users/Hosts in the last 24h is raising too...

HashClash has not given out any new work for almost 1 month now either. Some of the new users could have come over here waiting for HC to come back online & give out more work. There are several other Projects that are slow right now also ...
ID: 24963 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
R.L. Casey

Send message
Joined: 7 Jun 06
Posts: 91
Credit: 2,728,885
RAC: 0
Message 24988 - Posted: 26 Aug 2006, 14:37:01 UTC - in response to Message 24954.  

The server status on the main page of Rosetta is not reliable. It fluctuates a lot. I think the best overview and the most accurate and up-to-date information is delivered by free-dc:

http://stats.free-dc.org/new/allproj.php?proj=rah

This shows the average was about 37-38 before the credit change and now seems to stabilize around 35-36 with the new credit system.

Thnnks for the link, tralala. Very nice graphics! I must be missing something, however: all I was able to find were daily, accumulative, 7-day average, and 28-day average for points, users, teams, and hosts. The same data for Tflops would be great, but so far I haven't found it. Could you give me another hint? Much appreciated!

In fact, I can't offhand recall finding such data over time anywhere, yet, but it would be great to see, along with a description of how it's derived. Thanks again!
ID: 24988 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
tralala

Send message
Joined: 8 Apr 06
Posts: 376
Credit: 581,806
RAC: 0
Message 24995 - Posted: 26 Aug 2006, 15:20:54 UTC - in response to Message 24988.  

The server status on the main page of Rosetta is not reliable. It fluctuates a lot. I think the best overview and the most accurate and up-to-date information is delivered by free-dc:

http://stats.free-dc.org/new/allproj.php?proj=rah

This shows the average was about 37-38 before the credit change and now seems to stabilize around 35-36 with the new credit system.

Thnnks for the link, tralala. Very nice graphics! I must be missing something, however: all I was able to find were daily, accumulative, 7-day average, and 28-day average for points, users, teams, and hosts. The same data for Tflops would be great, but so far I haven't found it. Could you give me another hint? Much appreciated!

In fact, I can't offhand recall finding such data over time anywhere, yet, but it would be great to see, along with a description of how it's derived. Thanks again!


Points convert nicely into TFLOPS if you dived them by 100.000. :-)
ID: 24995 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
R.L. Casey

Send message
Joined: 7 Jun 06
Posts: 91
Credit: 2,728,885
RAC: 0
Message 25025 - Posted: 26 Aug 2006, 20:50:18 UTC - in response to Message 24995.  

The server status on the main page of Rosetta is not reliable. It fluctuates a lot. I think the best overview and the most accurate and up-to-date information is delivered by free-dc:

http://stats.free-dc.org/new/allproj.php?proj=rah

This shows the average was about 37-38 before the credit change and now seems to stabilize around 35-36 with the new credit system.

Thnnks for the link, tralala. Very nice graphics! I must be missing something, however: all I was able to find were daily, accumulative, 7-day average, and 28-day average for points, users, teams, and hosts. The same data for Tflops would be great, but so far I haven't found it. Could you give me another hint? Much appreciated!

In fact, I can't offhand recall finding such data over time anywhere, yet, but it would be great to see, along with a description of how it's derived. Thanks again!


Points convert nicely into TFLOPS if you dived them by 100.000. :-)


Ok, so using that rule of thumb(?) then Rosetta's points per day graph implies that their Tflops rate has been generally between ~35 Tflops and 39+ Tflops over the last 60 days. Is that the estimate you consider unreliable? I'm curious. Thanks.
ID: 25025 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
tralala

Send message
Joined: 8 Apr 06
Posts: 376
Credit: 581,806
RAC: 0
Message 25027 - Posted: 26 Aug 2006, 21:00:14 UTC - in response to Message 25025.  


Ok, so using that rule of thumb(?) then Rosetta's points per day graph implies that their Tflops rate has been generally between ~35 Tflops and 39+ Tflops over the last 60 days. Is that the estimate you consider unreliable? I'm curious. Thanks.


No I didn't say they are unreliable, they are reliable. And as I said in another posts it seems with the old credit system Rosetta had about 37 TFLOPS on average and with the new one about 35. The 2 TFLOPS loss is probably mainly due to the fact that Xtremesystems pulled out (which means the produced about 2 TFLOPS of new credits). Not such an extreme loss, right?
ID: 25027 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Angus

Send message
Joined: 17 Sep 05
Posts: 412
Credit: 321,053
RAC: 0
Message 25028 - Posted: 26 Aug 2006, 21:04:12 UTC - in response to Message 25027.  


Ok, so using that rule of thumb(?) then Rosetta's points per day graph implies that their Tflops rate has been generally between ~35 Tflops and 39+ Tflops over the last 60 days. Is that the estimate you consider unreliable? I'm curious. Thanks.


No I didn't say they are unreliable, they are reliable. And as I said in another posts it seems with the old credit system Rosetta had about 37 TFLOPS on average and with the new one about 35. The 2 TFLOPS loss is probably mainly due to the fact that Xtremesystems pulled out (which means the produced about 2 TFLOPS of new credits). Not such an extreme loss, right?


Well, considering they produced about 500,000 of the 3,500,000 credits, that's about a 15% loss, since you're basing TFLOPS on credits (which is non-sense, by the way).

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: 25028 · 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 25031 - Posted: 26 Aug 2006, 21:14:48 UTC

it would be nice to know the teraflops we are putting out every day as a group to see how we are doing
ID: 25031 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Angus

Send message
Joined: 17 Sep 05
Posts: 412
Credit: 321,053
RAC: 0
Message 25033 - Posted: 26 Aug 2006, 21:18:58 UTC - in response to Message 25031.  

it would be nice to know the teraflops we are putting out every day as a group to see how we are doing


There is NO way to tell, currently. Anything anyone puts out is only a SWAG.

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: 25033 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
1 · 2 · Next

Message boards : Rosetta@home Science : TERAFLOPS



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