bug-bash
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Parallelism a la make -j <n> / GNU parallel


From: John Kearney
Subject: Re: Parallelism a la make -j <n> / GNU parallel
Date: Thu, 03 May 2012 22:12:17 +0200
User-agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:12.0) Gecko/20120428 Thunderbird/12.0.1

I tend to do something more like this


    function runJobParrell {
        local mjobCnt=${1} && shift
        jcnt=0
        function WrapJob {
            "address@hidden"
            kill -s USR2 $$
        }
        function JobFinised {
            jcnt=$((${jcnt}-1))
        }
        trap JobFinised USR2
        while [ $# -gt 0 ] ; do
            while [ ${jcnt} -lt ${mjobCnt} ]; do
                jcnt=$((${jcnt}+1))
                echo WrapJob "${1}" "${2}"
                WrapJob "${1}" "${2}" &
                shift 2
        done
        sleep 1
        done
    }
    function testProcess {
        echo "${*}"
        sleep 1
    }
    runJobParrell 2  testProcess "jiji#" testProcess "jiji#" testProcess
"jiji#"

    tends to work well enough.
it gets a bit more complex if you want to recover output but not too much.

Am 03.05.2012 21:21, schrieb Elliott Forney:
> Here is a construct that I use sometimes... although you might wind up
> waiting for the slowest job in each iteration of the loop:
>
>
> maxiter=100
> ncore=8
>
> for iter in $(seq 1 $maxiter)
> do
>   startjob $iter &
>
>   if (( (iter % $ncore) == 0 ))
>   then
>     wait
>   fi
> done
>
>
> On Thu, May 3, 2012 at 12:49 PM, Colin McEwan <address@hidden> wrote:
>> Hi there,
>>
>> I don't know if this is anything that has ever been discussed or
>> considered, but would be interested in any thoughts.
>>
>> I frequently find myself these days writing shell scripts, to run on
>> multi-core machines, which could easily exploit lots of parallelism (eg. a
>> batch of a hundred independent simulations).
>>
>> The basic parallelism construct of '&' for async execution is highly
>> expressive, but it's not useful for this sort of use-case: starting up 100
>> jobs at once will leave them competing, and lead to excessive context
>> switching and paging.
>>
>> So for practical purposes, I find myself reaching for 'make -j<n>' or GNU
>> parallel, both of which destroy the expressiveness of the shell script as I
>> have to redirect commands and parameters to Makefiles or stdout, and
>> wrestle with appropriate levels of quoting.
>>
>> What I would really *like* would be an extension to the shell which
>> implements the same sort of parallelism-limiting / 'process pooling' found
>> in make or 'parallel' via an operator in the shell language, similar to '&'
>> which has semantics of *possibly* continuing asynchronously (like '&') if
>> system resources allow, or waiting for the process to complete (';').
>>
>> Any thoughts, anyone?
>>
>> Thanks!
>>
>> --
>> C.
>>
>> https://plus.google.com/109211294311109803299
>> https://www.facebook.com/mcewanca




reply via email to

[Prev in Thread] Current Thread [Next in Thread]