A Tour of our Computer Hardware, Part II

A Tour of our Computer Hardware, Part II



my mom just sent me this old test I had from the seventh grade that showed I got hundred-percent on my quiz and stron immediately back when we had nine planets seventh grade science teacher was all about astronomy so that was his big thing and that's probably where I had started strong serve the room now this is a room that has a lot of our computers that we have running all the time they're a little bit noisy so we don't have them up here I'm an R&D engineer here at UC Berkeley Radio Astronomy lab and I get to work on all kinds of radio astronomy instrumentation projects currently mostly SETI related why don't look at that end is for the astronomy department itself this little one you think all this our head node this guy has this property system on it he runs Linux this guy's really the brain of everything and these guys just sort of wake up and run everything on him simply signifies our assistant administration because one system one place to install software worries about like what version of programs we have on this computer versus a computer so the compute node we saw in there is useful this project called breakthrough listen which is a different kind of SETI experiment and this commensal observing with breakthrough listen we actually get to point to telescope where we want to point it this is our compute node you can actually decide this with dual CPU system memory we got a lot of disk drives here 24 disk drives in the front Titan X GPU there this box basically reports data while we observe a breakthrough listen we actually get to tell us what we want to point it into that we see where we wanted to be and so you're charged with the driver's seat and we get to the plaintiff data while we are observing identity you can go back and crunch numbers further the ideas we've like the data every day we picked off the food stuff we started over again so raid 0 gives us a lot of disks in parallel so to get a lot of food but because we have favorite party needs most of what we do is recorded to disk while we're observing because we have so much data coming in this is currently running on the greenback telescope and it is very wide bandwidth and so we get a lot of data coming in record to disk we use a raid 0 array which is basically a bunch of disk drives in parallel a lot of these telescopes tend to be in remote locations which have limited throughput to the internet and so we can't just take all the data from the telescope and send it to the cloud we have to do other data reduction techniques to make it more manageable besides that we said this extended the voltage day that's just too much descent so you have to do something with it I was doing some work with embedded programming I ended up connecting with some people who worked at settings – – on the Allen telescope array with this little tiny embedded Java device called the tiny and then a few years later I saw a job posting for software and hardware development on the Helen telescope array here at the Freid astronomy lab and I applied the rest is history survey did I have a messy bedroom growing up I did actually yeah yeah yeah in fact my whole life has been kind of just organized chaos I guess you'd call it maybe and so this is kind of more of the same

About the author

Comments

  1. I actually expected more of a "beast" setup. I stopped running BOINC on my home pc because all i use it for now is gaming, but I have my old LG G4 running BOINC 24/7 now, haha. Every little bit helps, right?

  2. wow! give these guys some equipment!!! Commun Cisco, IBM and NVidia!!! where I work we are trashing equipment that are not even half as old as this!!! They deserve Cisco UCS servers!!! and a proper clean room.

  3. I have a shed and I have installed 2500 watt-hour solar on it and I power 3 computers that crunch [email protected] I hope solar can drop in price and GPUs can become even more efficient so others can do the same.

  4. Would recommend that you contact ComputerCorps out in Carson City, NV to see if you can upgrade your systems at at fraction of the cost. Netgear? Really, you should be running Cisco.

  5. It would be WONDERFUL if they had used quality equipment to record this interview. Can anybody understand but a few words said? ^..^~~

  6. Could be courting controversy, but maybe the conversation will do some good.
    Interesting statistics from boincstats.com;
    It looks like over 90% of [email protected] users are not active and there are only 139,000 active users worldwide?
    It looks like in the single month before the 'signal' from HD164595 over 2% of active users had dropped off in that month alone (lucky there was a signal, huh!).

    As you can see above, BerkeleySETI only has 1,748 subscribers – and how many of them are active followers?
    (I can give a flavour of that perhaps: a video from only 4 months ago amassed a massive 251 views).
    The video above has attracted – up until this point – only 16 comments. That is not good – it suggests no real interest in the video, no conversation about what is happening with SETI and no passion or engagement for/with the project.

    In any case, it's seems like pretty poor performance overall and perhaps somebody should be looking at getting the subscribers and active users increased and getting a bit of life back into the project.
    (Note: I'm no expert in stats, so if I'm wrong about any of the information I'm happy to retract the comment).

  7. It's sad how low tech this datacenter is, as it's doing the universal search for other intelligence.  Fund Science!

Leave a Reply

Your email address will not be published. Required fields are marked *