Running a local application that uses JuliaHub only for heavy compute.

[ii8]

Is it possible to run an application on my local machine to gather data and then upload it to JuliaHub to perform some computation on it? Automatically and repeatedly I mean.

I had imagined one could use julias distributed constructs like addprocs and @spawnat to interact with the JuliaHub cloud and in that way execute certain computationally intensive parts of an application remotely.

[pfitzseb]

You can't use the native Julia machinery for this, no. Currently you'll need to use JuliaHubClient.jl (a package in the JuliaComputingRegistry, which you can access on juliahub.com) to programmatically submit jobs.

[layne-sadler]

@ii8 it sounded like you were also looking to programmatically upload data. Would you be manually triggering these batches or would they run on a set schedule?

[ii8]

@layne-sadler manually I think, not 100% sure what you mean. My local application would gather the data, send it to JuliaHub for making calculations, get a result back, and then continue gathering data and repeat the process. My problem is just that I have to gather the data locally but my machine cannot handle the compute part. Certainly I don't want JuliaHub to schedule or start anything by itself, only for it to respond to my local machine when it needs help computing.

[StefanKarpinski]

You can submit batch jobs from your local machine to JuliaHub via JuliaHubClient.jl, which might be what you want?