South African students are supercomputing champions

[Rkootknir]

Well done!

Are there no technical articles on the event?

I've haven't been able to find a technical writeup yet... However, there is a FAQ on the HPC website detailing a couple of technical details here: http://www.hpcadvisorycouncil.com/events/2013/ISC13-Student-Cluster-Competition/attendee_reg.php

 

[b@nD]

Detail !

Can we know WHO these peeps are ( NAMES ) where they come from , and what institution they attend.
Maybe a bio or two ?

Thanks\

 

[ActivateD]

Congrats to the team they did South Africa proud. I wish they grow from this experience and inspire others.

 

[ponder]

Can we know WHO these peeps are ( NAMES ) where they come from , and what institution they attend.
Maybe a bio or two ?

[video=youtube;QZecH6u3t4U]http://www.youtube.com/watch?v=QZecH6u3t4U[/video]


http://studentclustercomp.com/?p=1416

South Africa Scores Massive Upset at ISC’13

June 20, 2013 · Dan Olds

In the greatest upset in cluster competition history, the team from South Africa’s CHPC won the overall championship at the ISC’13 Student Cluster Challenge. They were one of only two teams that successfully completed every task.

In the win, the CHPC team overcame several obstacles, including not having access to their competition configuration until a day after they arrived at the show. They were also the youngest team, composed exclusively of undergraduate students.

The team, sponsored by Dell, made good use of the PowerEdge R320/R720 servers provided by the company. Their cluster had eight nodes, each with dual Xeon E5-2660 processors, for a total of 128 CPU cores. The South Africa system also sported 512 GB of memory (64GB per node) and utilized a Mellanox FDR Infiniband interconnect.

Like every other team this year, the South African team used accelerators to give their cluster a turbocharged punch. While some teams used Intel Phi co-processors, Team South Africa went with eight NVIDIA K20 cards, which seemed to do the trick.

Congratulations, South Africa! What a way to put yourself onto the Student Cluster Competition map!

We’ll be following up with much more material from the ISC’13 Challenge including interviews with the teams, details about their configurations and strategies, and analysis of what we learned from this competition. Stay tuned, there’s much more ISC’13 Challenge material on the way…

http://insidehpc.com/2013/06/20/student-cluster-invictus-team-south-africa-wins-the-day-at-isc13/

Student Cluster Invictus – Team South Africa Wins the Day at ISC’13

We are pleased to announce that Team South Africa won the ISC’13 Student Cluster Challenge this week. A hearty congratulations goes out to these young folks, as this was their first international cluster competition.

I had a chance to talk to Happy Sithole from South Africa’s CHPC just now, and he said that all the Student teams this year had accelerators in their clusters that they used to get better Linpack results. But the key to winning the day for Team South Africa was that they were the only team able to accelerate two of their benchmark applications in addition to Linpack. Even more remarkable, they only had access to their Dell cluster kit for three days before the competition.

Looks like they competed against older students many of them busy with Ph.D studies etc and they setup in less time.

Hopefully in the next few days there will be more videos & articles online (plenty of stuff for ISC'12 on youtube etc from last year).

 

[ponder]

http://www.hpcadvisorycouncil.com/events/2013/ISC13-Student-Cluster-Competition/team_chpc.php

Muhammad Atif

Evelyn Khoboka

No Pic
Katlego Moukangwe

Kerren Ortlepp

Jan-Willem Steeb

Ryan Strange

 

[hsark]

well done team SA its great to see young guys i'm sure they'll get rosy jobs in the future due to the lack of skills in science and IT and with SKA on the way the demand will even be higher

 

[Swa]

From Schmidt's statement I take that this is data and not knowledge or information.
He refers to user data, and from what I see on youTube something like 80 to 90% is either rubbish, no good to anyone, and duplications.
Once we look at useful information and knowledge the picture must look quite different.

No, the statement is from Bernie Fanaroff. If Schmidt is right then not much data was created up to 2003. Since then the amount has been exponentially increasing. Fanaroff can't exclude this data when referring to our entire existence. He's overestimating the importance of his own project. The data that's going to be sent over the cable actually dwarfs in comparison to some of the cables in the world.

 

[krieg]

Well done!!!

 

[dabbler]

Great job well done