Monterey
From FarmShare
Directions
http://www-marine.stanford.edu/find.html
Schedule
10:30-11:30 a pre-tutorial on Linux for beginners finalized to the process of getting access to and use of Stanford clusters: Alex/Gunnar.
- All attendees will have Stanford IDs, so recommend login to Farmshare?
- Probably will need to go over the basics of SSH and scp.
- Need to have basic Linux introduction, commands, examples for them to try
- Whether there are handouts or not, need to be sure that examples are documented on a website they can access after the fact.
Some Intro to Linux materials to consider pulling from if we don’t have things already:
- TACC course materials: http://www.tacc.utexas.edu/user-services/training/course-materials
- UNC RC Linux intro PPT (will put in GoogleDocs, will share if interested)
- OpenClassroom intro unix: http://openclassroom.stanford.edu/MainFolder/CoursePage.php?course=PracticalUnix
Use of clusters at Stanford (how to login to Farmshare and run jobs)
- Alex has been doing tech talk sessions on this. Also Steve Jones’ info is pretty good (something like this would prob be helpful to do with them, hands on): https://www.stanford.edu/group/hpcc/cgi-bin/mediawiki/index.php/I%27m_new_%26_don%27t_know_what_to_do
12:00-13:00 the main seminar on resource available for cluster computing at Stanford and at a national level, general administrative conditions to access them, etc.:
- Scientists are hired to do science, not run systems. So leverage the services offered around systems & HPC.
- Please also let them know we and ICME will have regular schedule of classes at Stanford on Linux, Matlab, Python, GPUs/CUDA, R. Goal is to video these and make available. Other topics of interest?
14:00-15:30 a final module on practical examples using R and C for parallel computing on the Farmshare. Giulio expects more attendees here than at other sessions. Jason mostly has the lead for this section. Perhaps you and Gunnar can demo the iPython work on Proclus? Good to review here the Farmshare resources (corns vs. barleys) . Whether R or C or whatever, the key is to show them how cluster computing can help them get more work done: job arrays, parallelization, GPUs, more efficient coding …. Only so much we can cover in 90 minutes. Maybe some hands on examples to try? Or examples of code before and after GPU-enablement (speed up)? Even if we don’t have on-site examples, there should be plenty online.
XSEDE examples
- TACC: C - http://www.tacc.utexas.edu/user-services/training/course-materials
- http://math.acadiau.ca/ACMMaC/Rmpi/index.html
- https://www.xsede.org/web/xup/online-training
- http://trg.apbionet.org/euasiagrid/docs/parallelR.notes.pdf
- http://cran.r-project.org/doc/contrib/Seefeld_StatsRBio.pdf (mostly the basics of R for Bio, they prob already use R)
- http://cran.r-project.org/doc/contrib/Martinez-RforBiologistv1.1.pdf
- http://www.nimbios.org/tutorials/TT_hpc2011 -- this looks promising
