Community  
BROWSE SEARCH
Search for:

LOGIN
Email:
Password:
If you don't have an account,
click here to sign up.
DOCUMENTATION DISCUSSIONS FILE ROOM
Computing power comparison: BOINC vs supercomputers

Computing power comparison: BOINC vs supercomputers

 
ID: 60194
Posts: 3

Hello,

I am a PhD student writing a review of online citizen science projects.  I want to include a comparison of the computing power of distributed computing vs the world's top supercomputers, but I have a question about how to do this comparison to make sure I'm making an accurate comparison.

The Gridrepublic Stats page (http://www.gridrepublic.org/index.php?page=stats) shows two estimates of distributed computing power.  The first estimate appears in the top-right of the page as "Recent Avg" (currently listed as  9812.15 Tflops).  The second estimate appears in the list of the world's top supercomputers, above the first supercomputer listed as "All projects available through GridRepublic, in aggregate: 10494403 Gflops".  These estimates are similar but not equal and I want to know why.

I appreciate whatever help people can offer me.  Thank you in advance for your assistance.

Jay

ID: 15261
Posts: 141

Did you ever recieve an answer?  

You should be able to get a technical answer from contacting the GridRepublic developers directly http://www.gridrepublic.org/contact

ID: 60194
Posts: 3

Thank you, Spark, for your reply.  I never did get an answer to my question, but decided not to pursue it because in my research I decided to just give a rough summary of computing power comparison between distributed computing and supercomputers, rather than use precise numbers.  Thank you for the suggestion though.

Jay

ID: 15261
Posts: 141

Jay I understand your need to move forward with your thesis.  Though as a note I do not belive you could directly compare the numbers of volunteer computing and supercomputers as the benchmark program used for testing the supercomputer rating is not run on the volunteer computing newtork.  The values are more a rough comparision that you can go by.  The supercomputer benchmark program is: http://www.top500.org/project/linpack.  While in volunteer compututing it is an aggrigate of the performance of individual computers based on the BOINC benchmark calculations.  Off the top of my head I am not sure how those two benchmarks compare.  BOINC Benchmarks non technical: http://boinc.berkeley.edu/wiki/Computation_credit
ID: 60194
Posts: 3

Thank you Spark.  I used the Top500 data for supercomputers.  It is  similar (but more recent) than the supercomputer data provided on the  GridComputing stats page.  My goal is to provide a rough comparison  between the aggregate power of the distributed computing network and the  most powerful supercomputer in the world, to convey the potential power  of distributed computing to a non-technical audience.  I used  supercomputer data from TOP500 (www.top500.org with Nov. 2010 data set).  Computing power refers to maximal LINPACK performance achieved.   Currently I have the following statement, and I welcome any feedback or  suggestions you have on it.  I know there are caveats galore in  comparing supercomputer and distributed computing power, but from what I  understand my statement is accurate albeit imprecise.

"At  the time of writing, the distributed computing network is approximately  four times as powerful as the most powerful supercomputer in the world  (www.gridrepublic.org/index.php?page=stats accessed Jan. 14, 2011)."

ID: 15261
Posts: 141

"At  the time of writing, the distributed computing network is approximately  four times as powerful as the most powerful supercomputer in the world  (www.gridrepublic.org/index.php?page=stats accessed Jan. 14, 2011)."

Your statement is fairly accurate I would say to get the point across.  

I would be very much interested in reading your thesis when you are done.  I a Masters Student at the Univeristy of Michigan and have been invovled in volunteer computing since 1999 and currently helping GridRepubic. You can reach me through my Google Profile: http://www.google.com/profiles/brierjon 

ID: 66126
Posts: 1

thanks for the links
ID: 124084
Posts: 2

My goal is to provide a rough comparison  between the aggregate power of  the distributed computing network and the  most powerful supercomputer  in the world, to convey the potential power  of distributed computing to  a non-technical audience

dvla contact phone number

Post   Previous   Page 1 of 1   Next
Please enable JavaScript support in Your web browser to display this page property.