Eric is working on proposals for most this week, and is still hard at work on Zone RFI Definitions.

Dave asked for hardware that he could use for BOINC. He has the BOINC alfa project running, and he’d like to have a separate server to run projects on. Eric said it would be fine if he ran it on marvin, it’s pretty stable, so it should be alright.

Recently, SETI@Home has been granting twice as much credit as it should have. This is happening because we used to give credits for the average of the integer performance and the floating point performance, and now we are shifting over to only giving credits for floating point performance. Usually the credits from integer performance are
What is the benefit of this shift?
Right now, the BOINC code doesn’t not provide a way to compute reasonable credit for GPU applications. Changing the way we give credits will We predict that people might be angry at this, but soon everyone will get used to it and it should be fine.

In case you haven’t noticed already, our server mork crashed yesterday. Matt suggested that we inact some sort of down time policy—where it is OK for SETI@Home and BOINC to be down for a certain amount of time. Matt recommended 72 hours, about the length of a weekend.

It should be kept in mind, however, that the longer we’re down, the harder it is to get it back up. Because of this, it was suggested that we should handle this problem on a case-by-case basis. People shouldn’t be expected to do it on the weekends, but if there is an easy, non-interfering thing you can do to fix a problem with SETI@Home, then it should be done. We should feel alright about leaving the project down on the weekend, and just put a note up on the website announcing the servers are down and we’ll get to it the next day.

But what if the BOINC server crashes? If this happens, all of the people downloading new projects would stop. This could be a much more serious problem. We should have a higher standard for BOINC—if BOINC crashes on the weekend, it should most definitely be fixed.

On that note, we are expecting a planned power outage up here at SSL in December between Christmas and New Year’s Day, when everybody is on furlough. The University has to replace a switch in the building. The power is going to go down for a day, then come back for a week, and then go down for another day. We are still unsure if we will have network coverage during that week, and we may have to plan on going down for that whole week.

Andrew announced that we had a couple of problems with the SERENDIP data recorder code. Thankfully, Terry caught the problem and fixed it. We recompiled the code and well, it’s fixed now. But we need to consider that in the current setup, if something goes wrong with SERENDIP and even just one beam goes down, it can dramatically increase our data rate. We need to implement some sort of mechanism to reject this before it goes in the database.

Dan mentioned that there is a little bit of cool work on using GPUs to do SETI spectroscopy going on right now, and that is something we might look for in the future.

Josh verified that his correction works, he is almost at the point where he can test the efficiency of his other RFI algorithms.

That is all this week from the Space Sciences Lab, Thanks for Reading!