Last updated Apr 24, 2020 1:50 PM

Fiji

Overview and documentation

Overview

The Fiji Cluster resource, funded by NIH 1S10OD01230, is a 70 node cluster consisting of 60 compute nodes (4 x Intel Xeon E5-4620v2, 2.6GHz 8-Core, 512GB with hyperthreading). Four high-memory nodes(4 x Intel Xeon E7-8857v2, 3.0GHz 12-Core, 21.3GB RAM / Core). 6 management nodes (2 x Intel Xeon E5-2650v2, 2.6 GHz 8-Core, 64GB). Fiji also offers diverse GPU resources with nVidia Titan, Tesla k20, and Tesla k40 accelerator cards and an interactive visualization node (4x Intel Xeon E7-8867v4, 2.40GHz 18-core) . All of the nodes are connected together via bonded 10G network using a Brocade network fabric for redundancy. Fiji is currently connected to a 1PB BeeGFS scratch file system, with an upgraded 1.5PB system currently being acquired for speed and feature enhancements. Fiji also offers 1.5 PB of cluster available home and project storage using iXsystems TrueNAS product with a redundant offsite replication and disaster recovery target.

To Request an Account

Submit a ticket with your PI's name and email address included. We'll contact you both to get billing information and get your account added to the compute cluster.

Documentation

  • Fiji User Guide

    Fiji User Guide

    General usage instructions and tips for logging into Fiji and interacting with the scheduler.

  • Fiji FAQ

    Fiji FAQ

    Frequently asked questions about Fiji

  • Fiji Visualization Resources

    Fiji Visualization Resources

    Information about x2go, rstudio, and jupyterhub on fiji-viz

  • Python Virtual Environments

    Python Virtual Environments

    Information regarding the creation and management of Python virtual environments

  • Containers on Fiji

    Containers on Fiji

    Information about available containerization on Fiji and documentation for using Singularity


  • Posted in Category: Cluster Computing