Skip to end of metadata
Go to start of metadata

You are viewing an old version of this content. View the current version.

Compare with Current View Version History

« Previous Version 5 Next »

Overview

What is HPC? - or - What is a Super Computer?

The term high performance computing (HPC) refers to any computational activity requiring more than a single computer to execute a task. Super computers and computer clusters are used to solve advanced computation problems.

How many cores does the Super Computer have? (Storage, RAM...)

48 compute nodes at 12 cores each = 576 cores

48 compute nodes at 64GB of RAM each = 768GB or roughly 3/4 of TB

48 compute nodes at 500GB local storage each = 23 TB

108 TB of GPFS storage. General Parallel File System, now called Spectrum Scale, is a high-performance clustered file system developed by IBM.

How fast is it?

  • 10 TFlops as reported by LINPACK. LINPACK is the benchmarking software used to rate and rank super computers.

  • The network for intranode communication is based on Infiniband and is rated for 56Gbits/second.

  • Comparison: A CPU currently found in some gaming PCs is the Intel I7 3930K which normally runs in the 3GHz range overclocked to 5GHz will net about 100 GFlops.

  • Florida Poly's Supercomputer is getting 10TFlops without overclocking and with plenty of room to grow in capacity and speed.

Who made it?

  • This HPC platform was made by IBM. The same people who made Watson the Jeopardy playing Super Computer!

What is it for? - or - What can it be used for?

  • The platform is very flexible and can be used for a number of things modeling and simulation mathematics and chemistry to assisting in video rendering or "big-data".

What Operating System (OS) does it use?

  • At this time, Florida Poly is using Red Hat Enterprise Linux.

What else is in the racks?

  • In addition to the HPC cluster Florida Poly has a Virtualization suite.
    Florida Poly is currently running VMware's vSphere 5.5 update 1
    Storage for the suite consists of 21 TB SAS and 2 TB of SSD
    Hypervisors and storage are connected via 10GB links

Requesting Access How do I package?

  • Professors

    • Please contact helpdesk to request access to the supercomputer or to request additional applications.

  • Students

    • A professor must request access to the HPC for the student. (see above)

Useful Information and commands.

  • Most users will use Putty to ssh into the HPC using their Poly email address and password. Usually going to Login.hpc.lab

  • To upload files to the HPC please use Filezilla or a similar ftp client. The host is login.hpc.lab, username/password is your FLPoly username/password and port is 22

  • Submitting jobs to the LSF (Use bsub)

    1. Compile the hello_world example code provided by default

      1. /opt/ibm/platform_mpi/bin/mpicc -o hello_world.exe /opt/ibm/platform_mpi/help/hello_world.c

    2. Submit a job through LSF to test the message passing

      1. bsub -n 10 -R "span[ptile=1]" -o %J.out "/opt/ibm/platform_mpi/bin/mpirun -lsf -vapi /home/(Username)/hello_world.exe; wait"

    3. Check the output of the %J.out file to verify results

  • If you request a Virtual Machine, you will be given the address in which you can connect either by ssh or remote desktop depending on the setup.

  • To upload files please use Filezilla or a similar ftp client. The host is the name of the Virtual Machine you were given, username/password is your FLPoly username/password and port is 22

  • For the EmberDB cluster please ssh into ember.hpc.lab using your FLPoly username and password.

    • Once connected use

      • Mysql –u (your username) -p –h ember-db

        • It will then ask for your FLPoly password.

Applications

  • Spack is a package manager for supercomputers, Linux, and macOS. It makes installing scientific software easy. With Spack, you can build a package with multiple versions, configurations, platforms, and compilers, and all of these builds can coexist on the same machine.

    • spack find

      • This lists all available packages.

    • spack load

      • This loads the package into your environment. For example spack load python@3.7.3 this loads Python 3.7.3 into your environment.

  • You can then use PIP to install necessary python modules

    • python3 -mpip install matplotlib for example

  • Apache Hadoop 2.6.0

  • Apache Spark 1.3.1

HPC Components

IBM nx360 M4

Model:

Dual Intel Xeon Processor E5-2620 v2 6C

Total cores per node:

12 core with 64GB RAM

Hardware threads per core:

12

Hardware threads per node

12 x 12 = 144

Clock rate:

2.1GHz

RAM:

8 8GB (1x8GB, 2Rx8, 1.35V) PC3L-12800 CL11 ECC DDR3 1600MHz LP RDIMM

Cache:

15MB Cache 1600MHz 80W

Local storage:

108 TB of GPFS storage

Management & Login Nodes

IBM System x3550 M4

The x3550 M4 is a cost- and density-balanced 1U, 2-socket server.

Intel Xeon processor E5-2600 v2 product family.

Supports up to 1866 MHz memory speeds.

Supports up to 768 GB memory with 32 GB LRDIMMs.

Network

The RackSwitch G8124E has support for 1G or 10G, this switch is designed for leveraging 10G Ethernet or have plans to in the future.

  • No labels