Jump to content

Does anyone here have experience renting compute time on a server?

Recommended Posts

I'm completing my master's thesis and of course two weeks before my deadline, the server from the university crashed and I quite possibly lost two weeks worth of work. I'm considering my options now to see if I could redo the work in a shorter time by paying for time on a better server. 


I'm running some artificial evolution software (Aevol) which is highly parallelized. It isn't too resource intensive as far as RAM goes, but it will use as many CPUs/threads as you can throw at it. I would need whatever cluster to let me run Ubuntu 18.04. 


The university server I was using was running 16 cores and it still was taking about two weeks for the experiments to complete, so any improvement over this is great.


I was getting so desperate, I was even considering buying a 3950x (16 cores, 32 threads) and building my own server since I was wanting to build a new gaming computer anyway, but it seems like pretty much all of the decent motherboards for that CPU are sold out. Probably it's better for my wallet this way. :p 


Does anyone here have any experience renting time on a cluster? If so, which company would you go with, and what do you think I could expect to pay for (ideally) 32 or more CPUs for about a week? Is an AWS EC2 instance likely all I would need, for example? 


This is all very new to me so I appreciate any information you all can give. Thank you! :hug:

Link to comment
Share on other sites

I have zero experience as well, but it looks like an Amazon service will fit the bill, particularly one of the Compute (c5) instances which they can offer up to 96  cores.


I'd say your other alternative is Microsoft, which can offer 120 AMD Epyc core:





These services are popular so I can imagine that both Microsoft and Amazon have made the entire process user friendly and simple, with plenty of support should you need it.

  • Thanks 1
Link to comment
Share on other sites

I ended up going with an Amazon Web Services C5 server just because I had already looked into that one and knew more or less what I was getting into there, but thank you @cusideabelincoln


I have my experiments running on an AWS server with 96 cores (had to contact customer support to get special permission to get that many :lol:) and the experiment that was taking just over 14 days on the university server (with 16 cores) is probably going to complete in 3.5 days on the AWS server. :cool: 


Thank you all! :hug:

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Recently Browsing   0 members

    • No registered users viewing this page.
  • Create New...