MBB other services

In addition to the online resources presented in this web portal, the MBB platform also offers the following services to the CeMEB member units :

Hosting

Requests to those services are accessible on our tracking system. For example, if you want to add your program as an online tool or as a downloadable binary/source package on this platform website.

Development

The development service is focused on sharing our experience about scientific program implementation techniques. A regular call for project allows you to submit a request if you need development assistance about parallelizing / optimizing a program, memory cleaning, porting code to other platforms, creating/modifying a WebUI / GUI, packaging, compilation routines... Those projects must not exceed a total of 3 months / 1 FTE (Full Time Equivalent).

Software forge

We are hosting a software forge here. A MBB account is necessary in order to access to it. Otherwise, you will need to do a specific request to access to it.

MBBWorkflows

This new service (2019 Dec.) allows user who are not really comfortable with command line interface, to launch jobs with a WebUI (through a browser). It replaces our old galaxy instance. Moreover, if you have any workflow project, it can be discussed with our team to develop a specific workflow and then deliver and deploy it in a reliable way (container), via a reproducible package containing all the workflow steps. If you want to access this platform, please submit a ticket here. Developpers can also send us their workflows on our forge.

SSH cluster access

The MBB computation cluster behind the web service is also reachable by SSH connections. It's possible to request an SSH account on our tracking system. That account gives you access to a Linux shell on the cluster master node to submit your jobs to the computing nodes through Open Grid Scheduler queue manager.

RStudio

This user-friendly interface is available to setup R scripts and run them on a dedicated machine. Use the tracking system to ask for an access to this web-based service. However, for a long term usage or if you plan to use much resources, please consider to use our cluster or book a "Standalone multicores servers" (see bellow).

GPU computations

3 GPU machines are equiped with :
These nodes can be reserved with the tracking system to run/training or even compile GPGPU codes. These machines are preinstalled with nvidia-docker and the latest version of CUDA.
A docker of Nvidia DIGITS (the *D*eep Learning *G*PU *T*raining *S*ystem) for training deep learning models is also available. Supported frameworks are: Caffe, Torch, and Tensorflow.
The booking of these machines are moderated by our team.

Standalone multicores servers :

Four specific servers equiped with 64 cores and 512GB of memory are available for big demanding projects. These nodes can be reserved with the tracking system to run multithreaded or High-memory softwares. Users are also provided with root privileges allowing them to install all needed programs. The booking of these machines are moderated by our team. By default, we allow a booking for a maximum period of 3 weeks. If the current planning allows it, this period can be extent to 3 additionnal weeks. However, the overall time of reservation within a period of 3 months must not exceed 6 weeks.

Helpfull Web applications :

  • SubWaw: help building reproducible BioInformatic workflows
  • Wicopa: help to create container recipes
  • RPACIB: help to create biocontainer recipes
  • SAGE: help to generate simple shiny interfaces


All these services are reserved to CeMEB LabEx members and associates. v