Teraflops seem to be sliding ....

Message boards : Cafe Rosetta : Teraflops seem to be sliding ....

To post messages, you must log in.

AuthorMessage
Brandon

Send message
Joined: 7 Jun 06
Posts: 10
Credit: 756,972
RAC: 0
Message 18885 - Posted: 18 Jun 2006, 14:43:50 UTC

in spite of the increase of users? Although I personally LOVE BOINC and all the options it offers ... I wonder if a number of people find it too confusing and abandon quickly after they join ....
ID: 18885 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Tarx

Send message
Joined: 2 Apr 06
Posts: 42
Credit: 103,468
RAC: 0
Message 18888 - Posted: 18 Jun 2006, 15:52:00 UTC - in response to Message 18885.  
Last modified: 18 Jun 2006, 15:52:49 UTC

in spite of the increase of users? Although I personally LOVE BOINC and all the options it offers ... I wonder if a number of people find it too confusing and abandon quickly after they join ....

I suspect that many are scaling back due to the summer heat. I've seen that in other projects where many teams have a drop for summer production and then return back to full power in the fall.
At Boinc they posted "There may also be an environmental cost. If your electricity is produced by burning fossil fuels, the extra electricity usage produces greenhouse gases that contribute to global warming. If this is the case, we recommend that you not leave your computer on just to run BOINC, or that you reduce your overall energy use to compensate." http://boinc.berkeley.edu/energy.php
The one thing they should have mentioned is that in the summer time, the cost goes more than double if the computer is in an airconditioned area! And it also might make a room too warm if on all the time.
On the other hand, if need to heat a room (e.g. in the winter, or in a cold basement), then the power cost is just a fraction of the electricity used as the heat is desired (and most of the power a PC uses, excluding the monitor, turns to heat).
ID: 18888 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Feet1st
Avatar

Send message
Joined: 30 Dec 05
Posts: 1755
Credit: 4,690,520
RAC: 0
Message 18894 - Posted: 18 Jun 2006, 16:59:23 UTC
Last modified: 18 Jun 2006, 16:59:45 UTC

There are always many variables at work. Adding a host doesn't necessarily add 100% resoruce share of a machine, and the machine isn't necessarily on 24hrs a day.

You will note that for most of the projects, there are about twice as many users as there are ACTIVE users. This implies that there is a high fallout rate. Either to other projects, or away from BOINC complexities.

That is why it is important to keep coming up with new ways to get the word out.

Can you help Rosetta?
Reasons some people avoid BOINC projects
Rosetta@home needs more publicity
Rosetta Associated Press article
Public Radio Journalist Looking for Participants
Rosetta@home Web Site Translations
Rosetta video/outreach project
How can we bring more users to the Rosetta project
Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/
ID: 18894 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile John Hunt
Avatar

Send message
Joined: 18 Sep 05
Posts: 446
Credit: 200,755
RAC: 0
Message 18895 - Posted: 18 Jun 2006, 17:27:27 UTC


Just a quick observation (for what it's worth).
Over on LHC, there is a guy who claims that everytime he connects with the project, another host ID is created.

http://lhcathome.cern.ch/forum_thread.php?id=2177

Could something similar be happening here?
ID: 18895 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Feet1st
Avatar

Send message
Joined: 30 Dec 05
Posts: 1755
Credit: 4,690,520
RAC: 0
Message 18915 - Posted: 19 Jun 2006, 4:00:40 UTC

Yes, several folks have described that as well. Seems generally to be related to BAM. It would reflect as a host add (not a user)... and then, eventually, the original would reflect as a host that is no longer active.

I believe this is likely the cause of some of the steep dips you see in the charts of active hosts.

But my point was that there is indeed a high fallout rate, and it is pretty consistent with all of the projects. If you feel you have an angle on WHY folks drop out, see if it might be something that the project could improve upon and if it's already been discussed in this thread.
Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/
ID: 18915 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Winkle

Send message
Joined: 22 May 06
Posts: 88
Credit: 1,354,930
RAC: 0
Message 18961 - Posted: 20 Jun 2006, 6:21:43 UTC

I believe that a lot of people drop out after a while simply because of apathy. Once the initial excitement of watching the graphics and seeing just how 'cool' it is wears off, the interest is gone and they drop out.

I dropped out of Predictor several months ago simply because there was no real project updates (or updates that were readily available). I never knew if I was helping the project or not, and whether I was doing any good at all. As far as other projects, I dropped out of Seti because of all the infighting and 'bitchiness' between its members. Einstein... I just didn't see the point and its relevance to humanity. Overall I found that the above distracted from the real project goals and the good which could come from them.

Don't get me wrong. The projects have their own merits in their own right. I just found they were not for me.

Instead I turned my CPU cycles to FAD, and since its demise looked around and found Rosetta. I only hope that all the work done there wasn't wasted.

Here I find that there is no 'clickiness' within the group and the majority are here only for one thing... and that is for the combined benefit of humanity, and there are project updates and descriptions readily available. I would hate for that to change in any way.

One thing I have noticed is that although my preferences state that I am willing to receive Rosetta Newsletters, I have yet to receive one. Maybe I haven't been here long enough, but I believe that would help keep people interested as they then know that they are making a difference.

Ian
ID: 18961 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Feet1st
Avatar

Send message
Joined: 30 Dec 05
Posts: 1755
Credit: 4,690,520
RAC: 0
Message 18985 - Posted: 20 Jun 2006, 16:29:31 UTC - in response to Message 18961.  
Last modified: 20 Jun 2006, 16:32:05 UTC

...my preferences state that I am willing to receive Rosetta Newsletters, I have yet to receive one.


Glad to hear that others appreciate the weather here as much as I do. We did send a "newsletter" of sorts out to participants back in May. In it we explained some of the recent science and project changes that people may not have been aware of if they weren't monitoring the msg boards or had dropped out. We also explained that CASP and the need for more TFLOPs. During the month of May we went from about 30 to 36 TFLOPs, so you are right on. Keeping people informed did indeed help bring more interest back to the project. That was certainly not the only factor for the increase in power, but everything helps.

You personally may not be able to crunch a TeraFLOP... but if you implement an idea!...

I wanted to take the opportunity to point out that the letter was written by the project participants, not the project team. And you can help with such things as well. There are presently people working on website translations. Adding a tell-a-friend option to the website. Creating a more friendly landing page. And ultimately, we'd like to produce an easy to install CD with project information, a user's guide, the videos being produced, and all the BOINC and Rosetta software. We'd also like to put together some materials for a "Teach the Teachers" campaign".

And you can always share an idea in the "How can we bring more users to the Rosetta project" thread.

We are always looking for more volunteers to help with projects like these.
Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/
ID: 18985 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Brandon

Send message
Joined: 7 Jun 06
Posts: 10
Credit: 756,972
RAC: 0
Message 19051 - Posted: 21 Jun 2006, 12:04:40 UTC
Last modified: 21 Jun 2006, 12:06:36 UTC

I'd like to back that up. I started on SETI years ago and then moved to United Devices for the cancer/anthrax/genome projects. I stayed there for several years because I love the idea of being able to aid in a direct application of science for disease research. I found that the United Devices web page got to where it wasn't even updated at all with project information (except for active user counts). I went to the web site in England for info but they seemed to stop updating about the project as well though they had static information on the science. I stayed with it because I want to help contribute but when I discovered Rossetta and saw how active it was I just thought "Yes!!!!". I personally don't care about credits .... but I LOVE to see the constant updates about how much power is being generated, etc ... I love just knowing you guys aren't ignoring us. It does make a difference. I don't anything about the science ... but I love that you guys are all excited about it and about how this computing power is allowing to look at more cases, and that you do inform us about it.
ID: 19051 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
FluffyChicken
Avatar

Send message
Joined: 1 Nov 05
Posts: 1260
Credit: 369,635
RAC: 0
Message 19285 - Posted: 25 Jun 2006, 15:31:23 UTC

Given a lot of the power increase will have been from DPC, XtremeSystems and Free-DC challenge between them and the move to the 7 day tasks here, a lot may now be repaying debt to other projects (part of the BOINC system) or they have moved back to their original projects of choice. Since the 7day task cause rosetta to give move crunching instantly it also means it reaches it's debt limit quicker.

Another note is Summer holidays at Universities and soon (here in the UK anyway) Schools will/are on holiday now. so any comuter there that run boinc will not be on or on as often.

The it's summer for the Nothern Hem., the World Cup is on so people are getting out ;-)
Team mauisun.org
ID: 19285 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : Cafe Rosetta : Teraflops seem to be sliding ....



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