2017-05 introduction to acenet€¦ · big data and data analytics tools and environments leading...
TRANSCRIPT
regional partner
Introduction to ACENETwww.acceleratediscovery.ca
regional partner
What is ACENET?
We accelerate discovery in Atlantic Canada through leadership and innovation in Advanced Research Computing infrastructure, expertise and training.
⚪Free of charge for researchers and industry-researcher collaborations.
⚪Funding - federal and provincial governments, some member institutions.
⚪Regional partner of Compute Canada.
regional partner
Overview
Began in 2003.CFI funding $25 million – 2006. Similar organizations developed in other regions. The four regions have come together with a national umbrella organization called Compute Canada.Compute Canada operates as a federated model; ACENET is Compute Canada’s regional partner in Atlantic Canada.
Compute Canada helps to: • better allocate scarce
funding dollars;• provide researchers with
seamless access to infrastructure and resources nationally; and
• enable the advanced computing community to have a single, strong voice in the national forum.
regional partner
Where is ACENET?
⚪ Support any researcher in region, not just member institutions
⚪ 18 staff at six campuses
regional partner
What is Advanced Research Computing
ANSWER: Any computing that takes too long, or is too complex for a desktop
regional partner
Computing Resources Available
A wide array of High Performance Computing and storage systemsBig Data and Data Analytics tools and environmentsLeading edge GPU computing systemsHighspeed, secure, fire-and-forget portal to transfer TBs of data (Globus Portal)Extensive software libraryGenAP computing platform to access and use genomic datasetsCompute Canada’s Cloud computing environmentData storage and back-up systems that provide stability and security options over your desktopDesktop and mobile videoconferencingVideoconferencing collaboration rooms at most ACENET partner institutions.
regional partner
ACENET Expertise AvailableDetermining computing resources neededDesigning, optimizing and troubleshooting computer codeCustomizing toolsIn-depth collaboration where needed
Installing, managing and maintaining advanced research computing equipment
Accessible and responsive support staffSpecialists:
– Big Data & Cloud– Molecular Dynamics– Research Data
Management– Visualization
regional partner
ACENET Training AvailableGroup and individual training from novice to advancedNovice to advancedStandard, discipline-specific, customizedSoftware Carpentry and Data Carpentry workshopsData management
regional partner
What is High Performance Computing
Many CPUs working on one problem• Many “serial” jobs running at once; or• Many CPUs working in close coordination.• Communication between CPUs is key!
Advantages• Faster results• Larger problems or more detailed simulations
Principal tool: Computer cluster
regional partner
ACENET Computers
Four clusters7000 CPU cores20 TB RAM500 TB disk storagePlus tapePlus softwareRemember: Shared resources!About 870 users
aggregatednumbers
regional partner
What is a Cluster?
storage
storage
storage
cl004
cl003
...
cl002
cl999
cl001
"head node"Internet
Dynamic Resource Manager
scheduler"Grid Engine"
regional partner
How Do I Get Connected?
Logging in:ssh Secure Shell
Moving Data Around:sftp Secure File Transfer ProtocolScp Secure Copy
Both provided with Mac OS X or Linuxssh -X [email protected] [email protected]
Windows? Try MobaXterm or PuTTY and WinSCP. Just google for them
regional partner
Command-Line Interface
regional partner
Applicationshttp://www.acceleratediscovery.ca/wiki/Software
Software requests to: [email protected]
MolecularGaussianGromacsQ-Chem
VASPNAMD
WebMOVMD
...
BioAbySSBowtie
rapsearch2MrBayesMigrate-n
PhyMLPLINK
...
Earth& OceanCDONCOGMTMM5
...
EngineeringANSYSFluentCFX
OpenFOAM…
Math&CSSageGAP
DiVinE...
AstroStarLab
IRAF...
regional partner
Software Developmenthttp://www.acceleratediscovery.ca/wiki/Software
ToolsTotalViewdbx, gdbValgrind
makecvs, svn
git...
LanguagesC/C++Fortran
PerlPython
RJava
...
ParallelAPIs
Open MPIOpenMP
...
LibrariesLAPACK
FFTWGSL
ACMLnetCDF
HDFBoost
...
CompilerSuitesGCCIntelPGI
SunStudio
Software requests to: [email protected]
regional partner
Data Analysishttp://www.acceleratediscovery.ca/wiki/Software
PythonNumpySciPySage
pyMPIpylab
matplotlib
RRmpi
Octave
Matlab**Distributed Computing Server
Software requests to: [email protected]
regional partner
Available Compilers At ACENET
C, C++, and FortranPortland Group (PGI) version 8
• pgcc, pgCC, pgf77, pgf90, pgf95GNU suite versions 3 and 4
• gcc, g++, gcc4, g++4, gfortran, g77Intel Compiler Suite XE 2011
• icc, icpc, ifortSunStudio version 12
• cc, CC, f77, f90, f95See www.acceleratediscovery.ca/wiki/Programming_Tools
regional partner
Environment Modules$ module listCurrently Loaded Modulefiles:1) pgi/8.0-6 3) totalview/8.8.0-22) openmpi/pgi/1.2.9
$ gcc –-versiongcc (GCC) 3.4.6 ...$ module avail gccgcc/4.4.3(default) gcc/4.6.4 gcc/4.8.0$ module load gcc$ gcc --versiongcc (GCC) 4.4.3 ... ← changed!$ module unload gcc
regional partner
MPI
OpenMPI is default implementationUse modules to vary the underlying compiler, e.g.
$ module unload openmpi$ module load gcc openmpi/gcc
See www.acceleratediscovery.ca/wiki/Open_MPI
regional partner
Parallel Job Processing
Notice that process count and host list is passed automatically from Grid Engine to Open MPI.
$ cat jobscript#$ -cwd#$ -j yes#$ -l h_rt=48:00:00#$ -l h_vmem=2G#$ -pe ompi* 32module purgemodule load gcc openmpi/gccmpirun ./application
regional partner
Where To Go For Help
Online:• gcc –help• man gcc
ACEnet wiki• http://acceleratediscovery.ca/wiki/ACEnet
Email support:• [email protected]• Please supply as much info as possible
www.acceleratediscovery.ca/wiki/Ask_Support
regional partner
What Cluster Should I Use?It doesn't matter,
ONE ACCOUNT IS GOOD EVERYWHERE
Multiple serial jobs? Shared memory parallel? MPI? Terabytes of storage? Some software has only local license
Fit in anywhereFundy best; Glooscap goodMahone, Placentia best; Fundy goodProbably Glooscape.g. Matlab DCS & Gaussian only at Placentia
See www.acceleratediscovery.ca/wiki/Software for availability, or contact your Computational Research Consultant