As Pionier partners are also members of PRACE RI, we also helped running the PRACE booth promoting the ideas of collaborative European HPC. Apart from that we presented results of our research on using hybrid CPU/GPU chips for HPC computations.

The Birds of a Feather (BoF) session we took part in covered also other topics such as building supercomputer out of ARM cores, using DSP processors or the benefits of cooling the servers with hot water.

Supercomputing Conference is a yearly opportunity not only to present ourown so lutions and projects. It turns out that not everything can be explained and discussed only on digital way. This time we were able to host at ur booth a group of specialists whom we know only virtually, and meet with users of similar systems installed all over the world.

Thank you for visiting our booth!

Nov/12

15

How NOVI works

If you have not got the opportunity to have your hands on NOVI project there is a video that was recorded at FIA 2011 (24-28 October) in Poznan. Professor Vassilis Maglaris provided a foreword, while PSNC partner showed the concept of NOVI Innovation Cloud and the quick prototype of NOVI virtual switch (NSwitch).

The video is available at:

http://www.week.fi-poznan.eu/online/articles/1663/NOVI_demonstration_VIDEO.html

Nov/12

15

Join NOVI

NOVI provides tools and methods to compose virtualized heterogeneous e-Infrastructures. Resources belonging to various levels ie. networking, storage or computational resources are managed by separate yet inter-working providers each running NOVI instance. Provided that you have resources that you would like to share with NOVI we describe below what you need to do.

NOVI Innovation Cloud adopts Service Oriented Architecture where functionality of the system is modeled as a set of services that communicate with each other. Infrastructure provider must implement a set of NOVI services specific to the given testbed. These include embedding, mapping, monitoring as well as federating. For the sake of readability further details regarding services will be provided only to those that will contact NOVI project.

When these core services are implemented they must pass repeatability and scalability tests. The underlying substrate may not be repeatable, but service layer must provide consistent results and perform its computational tasks in acceptable time limits. One bottleneck may impact onto the whole federated environment.

If you are a platform provider and you would like to join NOVI do not hesitate to contact NOVI!

CoolEmAll is pleased to announce that representatives from the project will be part of a special extended panel on the European Commission’s Smart Data Centre projects at DataCenterDynamics London on 14th and 15th November.

Full details can be found at “DataCenterDynamics London”

(http://www.datacenterdynamics.com/conferences/2012/london-2012?l=programme)

Virtualization improves resources utilization, but it comes with the price of virtualization overhead. It not only affects reduced memory or CPU power, but also impacts on network measurements.

The lack of the time deterministic processing in the virtual environment and issues with time synchronization in platforms limit applicability of delay measurements. Random and systematic measurement errors start to appear due to virtualization.

Monitoring gets tricky in federated environments. In homogeneous platforms tools are usually of the same type, while in heterogeneous environments tools are usually different. The tools on two different ends must cooperate with each other. This problem is important only when measurements are made on the physical substrate, because in a slice the experimenter have full control over resources and probably will deploy homogeneous tools.
If you are interested in other research topics that are evaluated in NOVI visit our website and don’t hesitate to contact us. The website is available at:

http://www.fp7-novi.eu/

Online tools are a really useful way to let people know what we’re up to at Project CoolEmAll. As well as http://www.coolemall.eu website, we use LinkedIn, Twitter and other online tools to communicate with the outside world. But sometimes the old fashioned ways work well too.

That’s why we have decided to put together a printable pamphlet which gives a great overview of the project. You can download the pamphlet here:

 

http://coolemall.eu/documents/10157/31079cb2-11a9-4ae0-a0ce-b76ae161ca24

 

We’ll also be handing out hard copies at the numerous events we plan to attend, and speak at, over the course of the project.

Have you seen our small competition on SC’12 booth? Not yet? It is the last chance to win a gadget there – come to our booth tommorrow and try to find the winning code between the others.

It is not so hard – stop for a couple of minutes and have fun!

 

Nov/12

13

We’ve started

Many thanks for the first visitors at our stand! This year, it is not difficult to find us, so the first day we were able to welcome you (as you see on pictures below). Grand launch of the exhibition, new topics, premieres, but also an opportunity to face meetings. That is why we are waiting for you! We invite you to booth # 1901.

 

Nov/12

13

So, here we go!

Welcome to the joint booth of Poznan Supercomputing and Networking Center and the PIONIER Consortium at the next Supercomputing Conference! This time we host in the capital of the state of Utah, in the snow-covered Salt Lake City.

 

We warmly invite you to visit our booth #1901. It’s easy to find us, we are just at the entrance. See then what we are presenting this year … and try to win a gadget in our small competition – QRcontest!

 

So, here we go – today at 6:00PM – the grand opening of the exhibition. Enjoy!