Have you ever wanted to visit CERN, or maybe even work there? Well guess what — one of the prizes for the Intel Modern Code Developer Challenge 2015 is a trip to CERN — and another one for a 9 week internship there!
The CERN and Intel sponsored competition is looking for a bright young developer (you must be a student) to improve the performance of the code used to simulate brains — specifically to simulate the growth of cells in the cerebral cortex. It’s called the Human Green Brain Project.
In this Challenge, you’ll be working with this code to improve its runtime performance, so researchers can make life-changing scientific breakthroughs faster. Download the code, optimize it, and submit it to the Challenge. The students who submit the fastest optimized code will win the prizes and help accelerate science – that could be you!
Improve it, and you’re literally accelerating science and research discovery. Oh — and you’ll get a chance to visit or work at the CERN OpenLab. What are you waiting for? Go enter!
[Thanks for the tip David!]
Too bad the competition is limited to students. I wonder if being enrolled in the School of Hard Knocks is sufficient?
Why just students?
Dirty cheap labour without questioning!
Lots of talent (of non-student interested people) goes to waste
What no clock building wtf
cern had enough publicity with their doomsday black holes, the last thing they need is for people to find out they’re building “bombs”
The registration seems horribly unsafe. The registration confirmation link gets you to a page showing your password in cleartext. Meaning they don’t hash it!
Let’s hope security on CERN’s other stuff is better…* ;-)
* Which it most probably is
M3gaFr3ak, please clarify for me so I can understand the concern. You receive an email that includes your cluster login credentials (login/pw). This is separate and not connected to the registration process or site login.
I want to make sure I understand your concern so I can address this properly.
after clicking the confirmation link I was sent to a page that showed me my account password (NOT THE CLUSTER PW).
hi, this is resolved.
Been there in Geneva at the ATLAS detector. It was a bit of a letdown. You take an elevator down to the tubes. There you can stand on a small platform and make photos, but to you won’t see the inside of the detector if they haven’t opened it for inspection and you can’t walk inside the ring.
moderncodechallenge.intel.com uses an invalid security certificate.
The certificate is not trusted because no issuer chain was provided
Just remember to not try too hard while at cern. Installing expensive software (licensed to cern, but not to interns!) to work on your assignment during the weekend can cost you $30K
https://cds.cern.ch/journal/CERNBulletin/2015/28/News%20Articles/2030047?ln=en
Hey, quick question is the subversion cx3dp-core empty?
hxxps://svn.ini.uzh.ch/pub/cx3dp-core/
Just installed TortoiseSVN (missing some VC+ libs) so I can’t run the repo browser. Regular web browser not displaying any contents.
*sigh* Anyway the cx3d.zip is available. ANNNNDDDD. It’s in Sun Java. I mean Oracle java.
Has anyone registered all info at the Intel Site? Is it a way to get a free edition of the Intel compiler or something? (i.e. I might not win the internship but does that mean I get a license key for the compiler?)
I haven’t dove in to the source of the cx3d aside from cursory examination. So I don’t know if there is any fork/join or java.util.concurrency calls in the current codebase.
So, I’m just going to postulate if the cx2dp isn’t there what one will have to do take the existing java and hit it over the head with a j2c to get some c code. Or lexer/parser the java source (or attempt to compile the java into javabyte code THEN get that into LLVM IR compatible code to get to C and hit it with gcc or conversely the icc…) into work a version of C that be compiled, once you got that then you’ll have to take the C code and make it parallelized or built a nice front end to use AWS/Azure or Apache Spark.
Bearing in mind it won’t be for x86 I’m considering as the focus would be on VLIW. Focus would probably to leverage AVX2 but primarily AVX-512 on the XEON PHI’s. Mostly to have a framework and say “Yes, beastly the PCI card co-processors CAN DO!”
http://www.amazon.com/Intel-Xeon-3120A-Heptapentaconta-core-Coprocessor/dp/B00D6020KO (I studied the specs no it will not run the newest version Crysis across 6 screens, it can only do it virtually.)
Or you could take the .java and get it to work on ARM via Davlik JVM. *free shrugs?* And then shove and mold the code with a HMP framework.
Oh, here is what I could find
https://code.google.com/archive/p/cx3dp/source/default/source
/sarcasm Here is the outline in video format of what really needs to be done:
https://www.youtube.com/watch?v=SWMGd_rzRdY
shot out to [breitwieser] for exporting a clean version from the google code graveyard.
https://github.com/breitwieser/cx3d-cpp
FYI – Project seems like it is designed more to mimic removing the ladder from your sims swimming pool.
http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0071615
I don’t think it’s designed to actually help up develop NZT.