Stream: helpdesk (published)

Topic: Asynchronous programming in Julia

Davi Sales Barreira (May 20 2022 at 19:29):

I'm starting to learn how to do asynchronous programming in Julia, yet, the topic is brand new to me and Julia's Documentation is kind of terse in the subject. In the code below, I download a bunch of files using the @async macro, and then I do some processing. The thing is, I wish to only the the processing once the files have been downloaded...

function parsedocuments()
    for b in [1, 11, 21]
        downloadbatch(documents[b:b+9,:])
        Threads.@threads for i in b:(b+9)
            try
                @suppress parsedocument(documents[i, :])
            catch
                continue
            end
        end
    end
end

function downloadbatch(batch)
    output = datadir() * "/temp/"
    for row in eachrow(batch)
        @async run(`aws s3 cp $(row[:pasta]) $(output)`)
    end
end

I haven't been able to figure out how to do this... I mean, the code for parsedocument starts running before the download has finished.

Brenhin Keller (May 20 2022 at 19:33):

Two separate loops? You may also want a @sync in there somewhere..

Davi Sales Barreira (May 20 2022 at 20:31):

Brenhin Keller said:

Two separate loops? You may also want a @sync in there somewhere..

Tried adding a @sync, but it did not work.

Maarten (May 20 2022 at 21:10):

@sync for row in eachrow(batch)
        @async run(`aws s3 cp $(row[:pasta]) $(output)`)
    end

Kwaku Oskin (May 23 2022 at 13:38):

You should read tkf blog about async patterns, it is extremely useful.

Or just use his packages, like FLoops.jl and ThreadsX.jl

Kwaku Oskin (May 23 2022 at 13:39):

https://discourse.julialang.org/t/tutorial-concurrency-patterns-for-controlled-parallelisms/62651

Last updated: Oct 02 2023 at 04:34 UTC