This is a reminder of the approaching deadline for submission of proposals.
If you wish to be considered for the allocation period starting on 1 April 2012,
then you need to submit your production project proposal no later than
Friday, 14 October 2011 at 18:00 (deadline).
Make sure that you have an active username and password so that you can
download an updated submission form from http://www.cscs.ch/nss/submission1/index.html
once we re-open the submission portal. At this very moment it is closed because
we are working on improvements to the submission process, trying to solve
the issues that some of our users had to face in the past. The portal should be
opened again in the beginning of October.
We would like to recall your attention on the following points:
1) We have introduced development projects (please visit our webpage for details)
2) In your production project proposal we expect to see a performance analysis of the
application(s) that you plan to use. We have a detailed ‘how to‘ on the webpage
Do not hesitate to contact us if you have need for further clarification.
3) If you have need for visualization as part of your production project, then please
describe in detail your requirements, both in terms of tools and support.
4) We accept now production projects which are targeted at data analysis instead of
high-parallel computation. See the list of special purpose machines below.
We further wish to recall your attention on our new systems:
Monte Rosa (Cray XE6) – Upgrade :
The new Rosa comes with AMD Interlagos 2 x 16-core 64-bit CPUs, 32 GB per compute node,
and high performance networking with Gemini 3D torus interconnect.
Production projects should target the Cray XE6, if they use highly scalable codes of excellent
performance, and make use of MPI or hybrid OpenMP/MPI parallelization schemes.
Tödi (Cray XK6) :
This is the first GPU/CPU hybrid supercomputing system with high scalability at CSCS,
designed to run data parallel and computationally intensive applications. It features
176 nodes, each with one 16-core AMD Opteron CPU, 32 GB DDR3 memory, and one NVIDIA Tesla X2090 GPU with 6 GB of GDDR5 memory.
Applications involving primarily local computation, heavy on floating point arithmetic but light on
communication, should benefit from this architecture. The system will initially be used chiefly for
development projects, in particular those which tune hybrid GPU/CPU and multicore applications.
Matterhorn (Cray XMT) :
The Cray XMT supercomputing system is a massively multi-threaded platform, with shared
memory architecture mainly for large-scale data analysis and data mining. It features
2 TB of globally shared memory and 64 Cray Threadstorm processors which allow to run
8’192 simultaneous hardware threads.
Rothorn (SGI Altix UV-1000):
This is a general-purpose system featuring large amounts of addressable memory (2TB of RAM in
total) and 32 Intel 8-core processors in 16 blades. OpenMP codes are most suitable, and memory-
intensive applications (e.g. in bioinformatics, data mining) will benefit in particular.
Detailed information on our submission process is available on our webpage