Presentation Supercomputing Zulip @ IAC SIEpedia IT News

Index of all issues November 2021 June 2021 March 2021 December 2020 Summer 2018 March / April 2018 January / February 2018 November / December 2017 September / October 2017 May / June 2017 January / February 2017 November / December 2016 July / August 2016
For earlier issues, please browse the Index

SIENews

The newsletter of the SIE de Investigación y Enseñanza N. 51 - May/June 2014

Teide HPC

Our researchers can now get access to the Teide-HPC. This supercomputer has more than 17000 Intel Sandy Bridge cores (the same type as those that you can find in MareNostrum), it is the second most powerful supercomputer in Spain and one of the most powerful computers in the World (169th position in June 2014, according to the Top500 Supercomputer Sites). If you are interested in executing your applications there, please get in touch with us by sending an email to res_support@iac.es. We have also written a Tutorial to help users who wish to run their programs in the Teide-HPC. We'll be also happy to give direct support to those users who need it.

IACPublications presented at the "LISA VII" Conference

The IACPubs application, a regularly updated database containing all papers written by IAC researchers, was presented in a contributed talk, delivered by Monique (general overview) and Jorge (technical aspects), at the 7th Library and Information Services in Astronomy (LISA VII) conference in Naples last June. IACPubs raised the interest of several astronomical institutions, which share the same concerns and needs in terms of managing and measuring the scientific output of their research staff. Among other interesting news presented at LISA VII, we'd like to highlight ADSLabs, which will eventually replace the "old tried-and-true ADS Abstract Databases", and ORCID, whose main purpose is to assign a persistent digital identifier to all researchers, thus eliminating the present confusion about authors with multiple name variants or same or look-alike names (we encourage everybody who hasn't already done so to register in ORCID and get their ID).

"Tele" command removed

With the new state-of-the-art VoIP telephone system, and its integrated phone directory, the command "tele" (used to look for phone numbers from the command line) has become out of date. Thus, after 13 years of hard and honest work, "tele" has been retired from service.

HTCondor + LaPalma Hall of Fame (1st semester 2014)

The first semester of 2014 is over and we have updated the HTCondor Stats in the Supercomputing Hall of Fame forum thread. HTCondor has been very busy in these 6 months, with more than 25 users consuming in total about 1.300.000 CPU hours (one and a half century in just 6 months!). As usual, we encourage our researchers to contact us or check our HTCondor documentation to see whether it is suitable for their computational needs. As for the LaPalma supercomputer, in the March-June period users consumed about 901350 CPU hours out of 1156710 CPU hours available (about 78%).

HTCondor Control Panel

Recently we have been developing some additional tools to help us managing in a efficient way the whole HTCondor pool, and to let our users get more detailed and complete information about HTCondor usage. In the last issue we introduced an online tool to get statistics data about HTCondor executions for each machine (http://carlota:81/condor_stats/). Now we are pleased to announce a new tool that offers a panoramic view of all our machines in a glance. You can use it to see on what PCs HTCondor is executing jobs, which slots are free, to check the load of any machine, their hardware and OS specifications, the HTCondor queue status, and many further parameters. These data are presented in a easy-to-interpret graphical panel, with additional graphs summarizing the most important HTCondor usage data. This Control Panel is still in beta version, but you can check it at http://carlota:81/condor_stats/condor_status.php; any feedback will be welcomed.

Standalone Checkpointing Mechanism

Have you ever been disappointed because you are running a C, C++ or Fortran program, and all of a sudden something happens and the execution is abruptly stopped (for instance, your PC crashes)? When that happens, most times you lose all the work and have to run again your program again from scratch, wasting a lot of time and resources ... It would be wonderful if we could "freeze" the execution at some points and then be able to run your application again from any of those points at any time, wouldn't it? That is perfectly possible with HTCondor's "Standalone Checkpointing Mechanism": you do NOT need to use HTCondor at all, only compile your program with a special command that links it with some libraries that allow you to set checkpoints for your execution at any time you want (for instance at regular intervals). This feature has been available since some time ago, but now more compilers are supported, after we upgraded HTCondor to v8.0.6 a few weeks ago. Please check the usage instructions, and get in touch with us if you have any doubt.
SIE de Investigación y Enseñanza :: N. 51 - May/June 2014 - Contact: