Announcements, 26th May 2015

Shaheen II Cray XC40 Workshop Announcement

Date: 7th June to 11th June 2015

Where: KAUST: Auditorium Al-Haytham (down the steps between Bldg2 and Bldg3)

KAUST Supercomputing Lab and Cray are offering a series of three courses:

*Sunday 7th June to Tuesday 9th, 2015 Introduction to the new Shaheen II Cray XC40

*Wednesday 10th June 2015 Efficient Parallel I/O

*Thursday 11th June 2015 Port and optimize your own code on the Cray XC40

Please book the appropriate days in the registration form! Link to registration: http://tiny.cc/ShaheenII_Workshop_June15 Deadline for registration: 6th June 2015 All courses will run from 9:00 AM to 5:00 PM.

Abstract

In May 2015, KAUST acquired and installed a new Cray XC40 supercomputer: Shaheen II. This machine delivers 25 times the sustained computing capability of KAUST’s current system (Shaheen I, IBM BG/P).

Shaheen II, with a theoretical peak performance of 7.2 PFLOP/s, is composed of 6,174 dual sockets compute nodes based on 16 cores Intel Haswell processors running at 2.3GHz representing a total of 197,568 cores. The processors are tightly integrated with a richly layered memory hierarchy and dragonfly interconnection network with a total storage space in the parallel file system of 17 PB.

The homogenous design of Shaheen II Cray XC40 with only Haswell makes it suitable for most, if not all, of KAUST applications.

This workshop is targeted to any potential Shaheen II users looking forward to port and run their codes in the new system. Temporary accounts on Shaheen II will be provided to accepted registrants. Minimum Linux/Unix knowledge is required.

Sunday 7th June to Tuesday 9th, 2015

Introduction to the new Shaheen II Cray XC40

This course will provide an introduction to the new Shaheen II Cray XC40 supercomputer, covering the programming environment, porting applications, profiling, and debugging, along with live demos. Furthermore, a series of hands-on sessions will give attendees the opportunity to practice with their own applications.

Wednesday 10th June 2015

Efficient Parallel IO

This fourth day is targeting users requiring parallel I/O at scale. The topics covered are I/O hardware aspects and Lustre parallel filesystem, along with I/O strategies with an overview of MPI-IO, NetCDF, HDF5 and ADIOS. Most of the topics discussed can be applied not only on a Cray XC40 system but also on any computer platform.

Thursday 11th June 2015

Port and optimize your own code on the Cray XC40

This fifth day is dedicated to efficiently porting and optimizing your code on the XC40. Registrants are encouraged to bring their own code, while Cray and KSL computational scientists will help them use the new HPC system efficiently.

For further information on this workshop and registration, you may contact:

Samar Aseeri : samar.aseeri@kaust.edu.sa

Bilel Hadri : bilel.hadri@kaust.edu.sa

Tip of the Week: Check Your Allocation:

Allocations for the computational resources on KSL systems are being managed using a program called "clusterbank". Users have access to this system via the "cbank" command. For example, the command "cbank list allocations" will display a list of all of the allocations currently active and available for your use:

user@fen3:~> cbank list allocations -p k999`<br>
#    End        Resource Project            Jobs       Charged     Available
---- ---------- -------- --------------- ------- ------------- -------------
4    2010-11-01 neseracc k999                  0           0.0      100000.0
                                         ------- ------------- -------------
                                               0           0.0      100000.0
Units are in core-hours.

Additional similar useful commands are as follows:

~> cbank list allocations -u username (lists all user allocations)
~> cbank list charges -p k## (lists account charges)
~> groupies k## (lists project members)
~> cb k## (gives complete data about allocations)

Follow us on Twitter

Follow all the latest news on HPC within the Supercomputing Lab and at KAUST, on Twitter @KAUST_HPC.

Shaheen II Cray XC40 Workshop Announcement

Date: 7th June to 11th June 2015

Where: KAUST: Auditorium Al-Haytham (down the steps between Bldg2 and Bldg3)

KAUST Supercomputing Lab and Cray are offering a series of three courses:

*Sunday 7th June to Tuesday 9th, 2015 Introduction to the new Shaheen II Cray XC40

*Wednesday 10th June 2015 Efficient Parallel I/O

*Thursday 11th June 2015 Port and optimize your own code on the Cray XC40

Please book the appropriate days in the registration form! Link to registration: http://tiny.cc/ShaheenII_Workshop_June15 Deadline for registration: 6th June 2015 All courses will run from 9:00 AM to 5:00 PM.

Abstract

In May 2015, KAUST acquired and installed a new Cray XC40 supercomputer: Shaheen II. This machine delivers 25 times the sustained computing capability of KAUST’s current system (Shaheen I, IBM BG/P).

Shaheen II, with a theoretical peak performance of 7.2 PFLOP/s, is composed of 6,174 dual sockets compute nodes based on 16 cores Intel Haswell processors running at 2.3GHz representing a total of 197,568 cores. The processors are tightly integrated with a richly layered memory hierarchy and dragonfly interconnection network with a total storage space in the parallel file system of 17 PB.

The homogenous design of Shaheen II Cray XC40 with only Haswell makes it suitable for most, if not all, of KAUST applications.

This workshop is targeted to any potential Shaheen II users looking forward to port and run their codes in the new system. Temporary accounts on Shaheen II will be provided to accepted registrants. Minimum Linux/Unix knowledge is required.

Sunday 7th June to Tuesday 9th, 2015

Introduction to the new Shaheen II Cray XC40

This course will provide an introduction to the new Shaheen II Cray XC40 supercomputer, covering the programming environment, porting applications, profiling, and debugging, along with live demos. Furthermore, a series of hands-on sessions will give attendees the opportunity to practice with their own applications.

Wednesday 10th June 2015

Efficient Parallel IO

This fourth day is targeting users requiring parallel I/O at scale. The topics covered are I/O hardware aspects and Lustre parallel filesystem, along with I/O strategies with an overview of MPI-IO, NetCDF, HDF5 and ADIOS. Most of the topics discussed can be applied not only on a Cray XC40 system but also on any computer platform.

Thursday 11th June 2015

Port and optimize your own code on the Cray XC40

This fifth day is dedicated to efficiently porting and optimizing your code on the XC40. Registrants are encouraged to bring their own code, while Cray and KSL computational scientists will help them use the new HPC system efficiently.

For further information on this workshop and registration, you may contact:

Samar Aseeri : samar.aseeri@kaust.edu.sa

Bilel Hadri : bilel.hadri@kaust.edu.sa

Tip of the Week: Check Your Allocation:

Allocations for the computational resources on KSL systems are being managed using a program called "clusterbank". Users have access to this system via the "cbank" command. For example, the command "cbank list allocations" will display a list of all of the allocations currently active and available for your use:

user@fen3:~> cbank list allocations -p k999
#    End        Resource Project            Jobs       Charged     Available
---- ---------- -------- --------------- ------- ------------- -------------
4    2010-11-01 neseracc k999                  0           0.0      100000.0
                                         ------- ------------- -------------
                                               0           0.0      100000.0
Units are in core-hours.

Additional similar useful commands are as follows:

~> cbank list allocations -u username (lists all user allocations)
~> cbank list charges -p k## (lists account charges)
~> groupies k## (lists project members)
~> cb k## (gives complete data about allocations)

Follow us on Twitter

Follow all the latest news on HPC within the Supercomputing Lab and at KAUST, on Twitter @KAUST_HPC.