seti@home stats

Eugene Leitl (eugene.leitl@lrz.uni-muenchen.de)
Sat, 22 May 1999 16:31:59 -0700 (PDT)

Similiar to distributed.net, seti@home now allows creation of competing teams http://setiathome.ssl.berkeley.edu/team.html , a distinctly motivating factor.

Current stats:
http://setiathome.ssl.berkeley.edu/stats/totals.html

Users

                  287675
  Work units sent
                  771222
  Results received
                  239930
  Total CPU time
                  8023159 hr 53 min 06.4 sec 
                  ( 915.89 years)

Average CPU time
per work unit
                  33 hr 26 min 22.5 sec


The disparity between work units sent and results received indicates that the joins have not saturated yet. Somehow, I think the creators of the project have never anticipated such a resonance ;)

Though it is difficult to imagine a project more motivating to the public than SETI an ALife project akin to http://www.hip.atr.co.jp/~ray/tierra/netreport/netreport.html could come close, especially if providing stunning visuals.

As the rollout of xDSL continues, the local density of potential project participants connected with low-latency ~1 Mbps might become supercritical for really worthwhile projects. JITed Java performance is apparently at the threshold of becoming comparable to true compiled languages. Creating a distributed GA for the sake of finding a mutation function capable of mutating Java VM opcodes robustly would be obviously extremely worthile. Searches of integer 3d CA rule space would be that as well, and simultaneously guarantee a rich source of visuals.

Of course one could also think about writing a code GA searching for constructive IP stack buffer overruns, and applying just-discovered exploits immediately... >;)