Originally posted by: Sukhoi
Last time I checked in here a month or two ago it seemed normal. Now I don't see any of the oldtimers that I recognize and it seems that SETI is dead. What happened? 😕 I was just going to install BOINC but now I'm not sure what to do.
Well, I'm still doing the daily SETI@home-stats, and as long as SetiatWork is down the weekly stats instead of Assimilator1, but Berkeley had a small backend-server-problem, and for this reason there haven't been any stats for 4 days before now they're back.
As for my crunching, still my main project is SETI@home, but is normally also crunching some wu for any new BOINC-project... and with BBC/CPDN's 5 months-wu even with only one running my seti-production is down... Oh, and when new beta-builds is released, has a tendency to run seti-beta instead of "main".
😉
As for "what happened", Seti_Enhanced was released in May, for majority of wu this means 2x-3x longer crunch-times compared to "normal" seti-v4.18, and some users thinks this is too long...
Another change is that before optimized seti-applications was 3x-4x times faster than the "standard" application, and this on average also gave 3x-4x more credit/hour. Since most of the optimizations is already in Seti_Enhanced, optimized Seti_Enhanced is more along the lines of 1.2x-1.5x, and for many the biggest complaint is they're not getting the same credit/hour as before any longer...
Seti_Enhanced also "counts flops" to decide claimed credit, instead of relying on the BOINC-benchmark, and apparently the claims from this is a little low, so for next release they'll be roughly 5% higher. Another problem is that not all types of calculations is as effective on all computers, and different angle-ranges has different amounts of the various calculations, meaning number of flops isn't linear with crunch-time. For this reason, VLAR is again having a disadvantage over other angle-ranges, and while haven't seen any advantage on VHAR on my Amd's, Intel and especially Intel-M seems to have an advantage on high angle-ranges...
There was also a bug in the splitter, giving some wu that crunched for many hours but only gave 0.1 credits or something, and others there application crashed. The splitter-bug has been fixed. There still seems to be an application-bug there some wu can get "stuck", but the upcoming beta will hopefully fix this...
Lastly, there's been many over-heated discussions on the seti-forums, and this topped itself when Crunch3r, the one releasing the fastest optimized applications for windows, left SETI@home partially due to cheating-accussations by one user on the seti-forums, but also partially since he was feeling ignored by Eric Korpela since didn't answer an email-message that hind-sight has shown didn't make it past the spam-filter...
Anyway, SETI@home has now finally started distribution of the last box of tapes that was collected at Arecibo when they installed the new recorder earlier this month, so anything from 2005/2006 showing-up is new tapes not crunched before.
😎
Hopefully by the time these tapes is depleted, Astropulse is ready to take over, or all the neccessary changes to start distributing data from the new recorder is finished...