Camelia, the Perl 6 bug

IRC log for #bioperl, 2009-08-20

| Channels | #bioperl index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
07:25 rbuels data::stream::bulk looks nice
07:25 * rbuels files it away in his tool box
07:32 Cryonic joined #bioperl
07:45 * faceface borrows rbuels tools and puts them back all out of order
07:45 rbuels gah!
07:46 faceface balin: rbuels thanks
07:46 faceface Its some wacky alignment format, so I figure I'll create Bio::SimpleAlign and output something sensible
07:46 faceface rbuels: ;-)
07:47 faceface rbuels: at one point you mentioned borrowing the compute resources at SGN
07:47 rbuels faceface: sure what do you need
07:47 faceface what specs do you have? I'm hitting 'problems' running some big assemblies... sadly the tools I'm using are not optimized for a cluster
07:47 faceface lots of memory!
07:48 rbuels we have a machine where we do assemblies like that
07:48 rbuels it's 128gb of ram
07:48 rbuels 16 proc
07:48 * rbuels checks to see how busy it is
07:48 faceface rbuels: I'm still in the early phases, so it may be that I can find a much better program to do what I need
07:48 faceface rbuels: thats 4 times bigger than we have (memory) and twice as large (processors)
07:48 faceface how much disk space ;-)
07:49 rbuels there's not that much available right now actually on it....
07:49 rbuels about 38G avail
07:50 rbuels but i bet some of that could be cleaned up
07:51 faceface rbuels: right
07:51 faceface its always the way for some reason
07:51 faceface disk = 98% full ... always
07:51 faceface I managed to generate 160 Gb using the Celera assembler (more memory efficient than Newbler).
07:52 rbuels yeah, this is becoming a problem for this machine
07:52 faceface people started to point at me (because the 'whole system' crashed).
07:52 rbuels i think we actually have some more disks on order for it
07:52 faceface how big a USB card can one machine have?
07:52 rbuels omg
07:52 rbuels haha
07:53 faceface I'm just wondering if I can plug in 255 0.5 Tb disks and raid 5 them ...
07:53 * rbuels sends email to sysadmin to make sure he's ordered the disk
07:53 faceface (usb disks)
07:53 rbuels hah, with the tweakyness of linux's usb support.....
07:53 rbuels no.
07:53 faceface hehe
07:54 faceface I'm just wondering what data rate you can get from motherboard (cpu) and usb card
07:54 faceface because I know individual usb connection is slow (hence need to raid)
07:55 rbuels well usb2 is not that slow
07:55 faceface cool
07:55 rbuels but when you start talking about having a lot of disks on it at the same time
07:55 rbuels that won't work
07:55 rbuels (well)
07:55 rbuels erm.
07:55 rbuels won't work well.
07:56 faceface not via a usb hub I guess, but I mean... how many usb cards can you slap on a mobo?
07:56 rbuels cause it can only talk to one drive at a time
07:56 faceface oh
07:56 rbuels hah, how many pci slots can you find on a mobo
07:56 rbuels 5 or 6 at most
07:56 faceface I need a hardware hacker ;-)
07:57 rbuels if you are thinking about 255 USB hard drives, probably what you are needing is a SAN
07:57 faceface I heard a nice talk from isilon
07:57 faceface very scalable storage solution... so I emailed them and asked for prices, but they blanked me
07:57 faceface tried a few times...
07:58 faceface perhaps the part where I said 'send me an evaluation node' marked me as a ...
07:58 faceface muppet?
07:58 rbuels lol
07:59 rbuels at sgn, we have several servers that we've gotten over the years
07:59 rbuels each of which has a few TB of disk in it
07:59 rbuels and shares things over nfs
07:59 rbuels but we're moving to something different
07:59 faceface its needed
08:00 rbuels what's needed?
08:00 faceface somethign different
08:01 * rbuels is losing mental steam
08:01 rbuels it's 2am for me
08:01 faceface night
08:02 rbuels night
08:02 faceface I'll let you know if I keep hitting problems, thanks for the offer
08:02 rbuels yeah, we can probably move most of the tempfiles off of tuber
08:03 rbuels to backup storage
08:03 rbuels (tuber is the name of the machine)
08:03 faceface nice
08:03 rbuels most of our servers are solanaceous things in one way or another
08:03 rbuels rubisco, solanine, hyoscine
08:03 rbuels tuber, eggplant
08:04 rbuels scopolamine
08:04 faceface name your archive 'storage vacuole'
08:04 faceface what?
08:04 rbuels eggplant is the current file server
08:04 rbuels cause it's....big and dark-colored?
08:04 rbuels lol ok sleep time
08:05 faceface *-)
08:05 rbuels night
08:30 faceface We are planning on some simple code refactoring for this module for the 1.7 release (minor cleanup issues, such as moving unnecessary methods to Bio::Align::Utilities
08:30 * faceface mimps
08:30 faceface as far as I'm concerned, 1.7 is already mythical.
08:42 faceface hmm... so if I add a seq with a gap... does it open a gap? I guess not...
08:42 faceface I have a reference sequence and a bunch of pairwise alignments
12:15 nErVe joined #bioperl
12:25 ChanServ joined #bioperl
12:25 driley joined #bioperl
12:25 faceface joined #bioperl
12:26 nErVe joined #bioperl
12:28 bioperl-bot joined #bioperl
12:29 ChanServ joined #bioperl
12:29 bioperl-bot joined #bioperl
12:29 nErVe joined #bioperl
12:29 tarbo2_ joined #bioperl
12:29 balin joined #bioperl
12:29 Cryonic joined #bioperl
12:29 driley joined #bioperl
12:29 faceface joined #bioperl
13:00 faceface joined #bioperl
14:40 nErVe joined #bioperl
16:08 _tristan__ joined #bioperl
17:13 nErVe joined #bioperl
17:21 _tristan_ joined #bioperl
18:06 tarbo2_ joined #bioperl
19:28 pyrimidine joined #bioperl
19:31 pyrimidine faceface: re: '1.7 is already mythical', it's not that far off, actually.  Any code refactoring or restructuring will have to go into 1.7.  rbuels is already helping on that front, so feel free to chip in.
19:31 * rbuels gets out his press-gang noose
19:31 * pyrimidine thought you said 'moose' at first
19:31 rbuels lol
20:52 vein joined #bioperl
20:57 pyrimidine rbuels: hah, beat ya to it!
20:57 rbuels lol, sure did
20:57 * rbuels sets his mail client to check for new mail EVERY 30 SECONDS
20:58 rbuels pyrimidine: so you're a postdoc at u-ill-u-c right now?
20:58 * rbuels seems to recall hearing that
20:58 pyrimidine yup
20:58 rbuels pyrimidine: how much longer you plan on being there?
20:59 pyrimidine dunno.  Maybe another year or two, tops
20:59 pyrimidine then out into the real world
21:00 rbuels does that mean industry?  cause academia isn't the real world  :-)
21:00 pyrimidine (or academia, depending on how the job hunt goes ;)
21:00 rbuels ah.
21:01 pyrimidine fac. jobs are few and far between (though Jason and Aaron Mackey both are faculty now)
21:01 rbuels man i'm already frustrated at how few good developers there are in academia....don't go into industry
21:01 rbuels lol
21:01 rbuels yeah, there are hardly any faculty jobs
21:02 pyrimidine that'll change but it will take a few years, economy willing
21:03 rbuels i'm trying to decide what i want to do with my career
21:04 rbuels i'm actually not that interested in science, but i AM interested in the hard software problems posed in bioinformatics
21:04 pyrimidine that's not  uncommon in the field
21:04 pyrimidine several of the guys here are the same way
21:05 rbuels i'm trying to decide if i really need to get a phd
21:05 rbuels i'm kind of leaning toward yes
21:05 rbuels but it almost seems like trying to fit a square peg in a round hole
21:06 rbuels but the way academia is structured......
21:06 rbuels seems like you need a phd to be taken seriously
21:06 pyrimidine depends
21:06 pyrimidine publications talk in academia
21:06 pyrimidine degrees do help consderably, though
21:07 pyrimidine *considerably
21:07 pyrimidine The CS dept here is top-notch BTW
21:07 rbuels yeah i know
21:08 rbuels i'm pretty interested in the data-modeling side of things actually
21:08 rbuels maybe a CS PhD working on that aspect.....
21:08 pyrimidine for instance
21:08 pyrimidine ?
21:08 rbuels for instance, chado as a database needs a lot of work
21:08 rbuels but there are a lot of good ideas there
21:09 pyrimidine yes
21:09 rbuels and in other database schemas, like that one from SRI
21:09 * rbuels forgets what it's called
21:09 rbuels and GUS
21:09 rbuels not really that interested in algorithm development....
21:10 rbuels would be bored to tears working on a gene finder or something like that
21:10 pyrimidine that area's done to death, yes
21:10 rbuels so, i dunno
21:11 pyrimidine the 'hot' area here appears to be GNR (genetic reg. networks) and how to process nextgen seq data
21:11 vein Three-dimensional chromatin modeling
21:12 rbuels what's the town like where you are pyrimidine
21:12 rbuels fun place to live?
21:12 rbuels i'm pretty spoiled now, living in portland
21:12 pyrimidine um, corn fields and soybean?
21:12 rbuels :-{
21:12 pyrimidine long winters?
21:12 rbuels hehehe
21:12 pyrimidine um, stay in portland
21:12 rbuels there are no good schools here though
21:13 rbuels good meaning top-notch
21:13 vein Or, come down to Irvine California o UCI
21:13 rbuels is irvine cool?
21:13 vein Its a pretty happening place as the saying goes
21:13 vein lots to do in California
21:13 * pyrimidine lives in the middle of vast swathes of corn and soy
21:13 vein er, sourthen anyway
21:14 vein *southern
21:14 pyrimidine yes, SoCal is nice.
21:29 vein It is a bit more expensive to live out this way though

| Channels | #bioperl index | Today | | Search | Google Search | Plain-Text | summary