Link to the TAMNUN temporary homepage for details about TAMNUN

These notes are a preparation for the courses to be given by SGI about TAMNUN. The first 3 links, which should be viewed in order, may also be useful for other first time LINUX users who need to learn to run programs on local and remote UNIX computers, such as PHELAFEL or the PHCLASS computers.

The fourth link describes the PBS queueing system used on TAMNUN and ALUF etc. This is also needed by NANCO users who have not used ALUF, since NANCO had a different queueing system.

NANCO/TAMNUN users and Technion Physics Department graduate students may contact Dr Joan Adler (phr76ja_at_tx.technion.ac.il) with further questions, OTHERS ARE WELCOME TO USE THE SITE ``AS IS''.

    Course links:
  1. How working on a remote LINUX machine differs from using your desktop - compile/link/run in LINUX
  2. Editing and managing files in a LINUX environment
  3. Accessing a remote LINUX machine from the windows environment
  4. PBS - by Julia Halupovich
The diagram below shows how to get the most benefit from these pages by first reading thru the first three links and then repeating them while taking the recommended actions.

The TAMLEARN page is: here .

QUESTION - What is PBS and why do you need it?
ANSWER - If we all ran our jobs interactively at the same time, the computer would spend all its time sharing out the resources and jumping from job to job. A queuing system shares the load out to different nodes so that a sensible number of jobs run at once and others wait their turn. PBS is the queuing system selected for use at the TECHNION. NANCO had a different system called SGE but for the end user they are similar.

Experienced NANCO users should go straight to the PBS link, but note that some more assistance for PBS with examples is given on the page about accessing a remote LINUX machine from the windows environment.

If you need the LINUX environment part of this course, proceed now to the ``How working on a remote LINUX machine differs from using your desktop'' - compile/link/run in LINUX

TAMNUN, is a cluster computer, funded by the Technion, via the Minerva Foundation and RBNI. TNN/SGI (Silicon Graphics) is the vendor and the cluster configuration includes 88 compute nodes with 12 Intel cores, which is 1056 cores with 96GB RAM per node, plus a masternode. 4 of the compute nodes have GPUs. MELLANOX infiniband switches with 2:1 blocking connect all nodes. The software includes RedHat LINUX, the PBS queueing system and Intel compilers. A large variety of codes, including MATLAB, fluent, Gaussian, VASP, LAMMPS will be installed, with usage subject to license restrictions.

TAMNUN is a contraction of Technion-Minerva-NANO and also alludes to an oCtoPUS with many CPUs. It will replace NANCO as well as providing additional computer support to Minerva and general Technion users. NANCO users will need to re-register using the link: TAMNUN account registration. Your NANCO files will not be moved automatically to TAMNUN, but they are saved and can be restored upon request. NANOCLUSTER-L subscribers have already been moved to TAMNUN-L. New users and other interested researchers may join the TAMNUN-L directly by pointing their browser to this link.

For the moment, until the computer becomes too full, usage will be free for everyone. When the load becomes too large users who are not partial owners will be charged. System and user support will be provided by the Technion Division for Computing and Information Systems, with first contact for users being the HPC consultant Ms Yulia Halopovich, hpc-support-l@listserv.technion.ac.il

See this page for more details including a video of TAMNUN's arrival.

Some more details about TAMNUN and its policies are on the mailing list archive which is linked below.

[Join or leave the TAMNUN list] [List archives]