I’m a computer science PhD student writing a paper with other master’s student about Optimization Algorithms for Large Nonlinear Equation Systems. For this work, we require to run a batch of algorithms and problems on high performance multi-core / multi-processor in order to analyze their behavior.
Our problem relies on the fact that we have no access to high-performance hardware and our fastest machine (6-core Ryzen CPU) would take over 9 months to compute all the solutions! As we would like to publish the article in a few weeks, would you be bothered to run our implementations on a machine with a high number of cores (32+)?
Our code was done in Julia Programming Language and all you would need to do was to execute a single file and send us the resulting data (2 .csv files and some .png graphs).
Hope to hear from you soon!
Thank you in advance for your time and consideration.
I believe that 2x64c Milan would be more than enough!
All its needed is to have the current stable release of Julia in the machine (probably some additional packages, but they can be easily installed in the Julia command-line REPL as the program asks them). I believe that you won’t need to spare more than 10 minutes to setup Julia + run the code.
How can I send you the source code files (about 50 KB uncompressed) to be executed?
You may use any disto that is the most convenient to you, as long its 64-bit.
The objective is to compare different algorithms among them self, so the only criteria is that they all get executed in the same conditions. For reference, we execute everything in Windows, but this is not a requirement.
I’m just polishing some data reporting on the code (so that it gatherers all necessary data without user intervention) and I will send it during the day (note that there will be some hours between us as I’m on the GMT time zone).
As I’m a relative new member in the forum, direct messages are not available, maybe if you send one to me, then I can reply?