BioHPC ECCO HPC Resource

The ECCO high-performance computing system is deployed as a part of BioHPC. All the resources available at BioHPC are available on ECCO servers, including networked storage and software .

Each ECCO user must have BioHPC account (it can be requested online) and belong to one of ECCO user groups:

ecco01  general group for all ECCO users that do not belong to any of the groups below
ecco_pjb298 Panle Barwick Lab  
ecco_jh2535 Jean-Francois Houde Lab 
ecco_ckh55 Christopher Huckfeldt Lab
ecco_lv39  Lars Vilhuber Lab 

Reserving a machine

BioHPC machines are divided into two groups: rental (pay-per-use) and hosted (owned by a group who has continuous access). The machines in the ECCO HPC cluster are operated as rental machines, however they are a bit special: reservations are restricted to ECCO group members, and there is no cost per use. This system ensures that the ECCO machines can be effectively shared among the group members.

Importantly, as with all BioHPC rental machines, your files stored on the rental machine are subject to deletion when your reservation ends. Therefore, you must transfer files out of /workdir/$USER/ before your reservation ends! For more information about the file deletion policy on rental machines, see the section "Local storage" here.

To use an ECCO server, a user has to create a reservation on BioHPC website (login first, then go to "Reservations" and choose category "Restricted"), or a user can be added to an existing reservation (the reservation owner can do this through the webpage, in the My Reservations section).

Once your reservation begins, you can access the machine through ssh or VNC. More documentation about access can be found here.

Each group has its own access priorities summarized in the table below. L (low) priority means maximum single reservation length of 1 day (24 hours), M (medium) 3 days (72 hours), H (high) 7 days (168 hours).

Machine name previous name (CAC) ecco01 ecco_pjb298 ecco_jh2535 ecco_ckh55 ecco_lv39
cbsuecco01 L M H M M
cbsuecco02 ecco-0-4 L M M M H
cbsuecco03 ecco-0-5 L H M M M
cbsuecco04 ecco-0-6 L H M M M
cbsuecco05 ecco-0-7 L H M M M
cbsuecco06 ecco-0-8 L H M M M
cbsuecco07 ecco-0-10 L H M M M
cbsuecco08 ecco-0-9 L M M H M

Software

All software installed on the BioHPC environment is available on all ECCO machines. The current list of software can be browsed here (or see here for a more compact view). It is strongly recommended to check the BioHPC software page before using particular software - sometimes several versions are available, or environment variables must be set in a certain way - a good example is GCC .

In addition to BioHPC software (including GCC set, R, Julia,  Python, Perl) there are several customized software titles installed on the ECCO machines. Not all software is available on all servers; the availability on each server is shown in the table below. Also, note that tmux and vim versions on ECCO machines are different than on the rest of BioHPC servers.

Machine name previous name (CAC) tmux vim Matlab STATA OXMetrics SAS Intel Compilers NAG Fortran Lib NAG Fortran Compiler
R2018a R2019a R2020a
cbsuecco01 x x   x   x x x* x x
cbsuecco02 ecco-0-4 x x x x x x x x x* x* x*
cbsuecco03 ecco-0-5 x x x x x x x x* x* x*
cbsuecco04 ecco-0-6 x x x x x x x x* x* x*
cbsuecco05 ecco-0-7 x x x x x x x x* x* x*
cbsuecco06 ecco-0-8 x x x x x x x x* x* x*
cbsuecco07 ecco-0-10 x x   x   x x x* x* x*
cbsuecco08 ecco-0-9 x x   x   x x x* x* x*

* - installed but user-supplied license (in home directory) is required to use it

NOTE about MATLAB: MATLAB installations are now in the process of being updated. The goal is to have the latest 'a' version installed on all machines, and additionally up to three previous 'a' versions available on cbsuecco02-06.

At present, the default MATLAB version is R2019a on all machines except cbsuecco06 where R2020a is now the default. To use a specific (non-default) version of  MATLAB, modify your search PATH by running

export PATH=/local/opt/MATLAB/RXXXXa:$PATH

replacing 'XXXX' with the desired year. You can add the export command above to your .bashrc file.