Question

My question is a bit specific to the scenario I'm facing. I wasn't really sure what to search for to find and answer to it here.

I'm working on developing some CUDA-based programs as part of my research. I have access to a supercomputing cluster with GPU nodes ( http://www.rcac.purdue.edu/userinfo/resources/carter/ ). And my development machine, being a 13'' late 2011 MacBook Pro, doesn't have an nVidia GPU.

There are two ways I can run CUDA apps on the node. I can either use 'qsub' to schedule a job and wait for it to run and get the output files. Or I can start an interactive shell on the gpu-node itself using the 'qsub -I' command. This has some waiting time associated with starting it, but once it starts, I can pretty much run apps in realtime as if I'm on my local machine.

Now, I use something like TextMate or TextWrangler to write my code. Is there any way I can run a command in THAT terminal ( the qsub -I one ) directly from my text editor ( or any editor ). This way, I will be able to make changes and test them much faster than manually copying the file to the server, compiling and running it there.

-- Thomas

Était-ce utile?

La solution

Do you know that the newer version of Textmate allow you edit content on remote machine on your local textmate installation and instantly save it on the server. This trick might save you a lot of time. Of course this is possible with most editor command line (using a simple SSH Reverse tunnel trick). But being able to continue using Textmate make rmate an even better choice.

Check out the author blog about this here. http://blog.macromates.com/2011/mate-and-rmate/

And Ernie Miller's blog for an awesome step by step approach.

Licencié sous: CC-BY-SA avec attribution
Non affilié à StackOverflow
scroll top