Others have argued for make; no need to say more. make is good.

Another option that's used commonly in the supercomputing centers is
NQS, a job queuing system that has a lot of the same characteristics as
the model used in JES.  NQS allows defining multiple queues with
priorities, and can limit the number of jobs in progress at any given
time.

A common model is to make the compilers and other CPU-intensive
applications owned by and runnable only by the NQS batch userids, so the
queuing system must be used.  It's kind of awkward if you're used to the
standard Unix model, but it does do an effective job of controlling
resources.

This is also a good model for using the resource management components
of the grid computing stuff.  If you have a common file system that can
be mounted between multiple Linux machines in a compile farm, using GRAM
to submit a job that picks a unused machine that is capped using the VM
SET SHARE command and does the compile there is a really elegant
solution. It also works cross-platform.

-- db

David Boyes
Sine Nomine Associates


> -----Original Message-----
> From: Linux on 390 Port [mailto:[EMAIL PROTECTED] Behalf Of
> Ferguson, Neale
> Sent: Thursday, March 20, 2003 9:02 AM
> To: [EMAIL PROTECTED]
> Subject: Re: "batch" compiles
>
>
> Using a make file to script the compiles that are required will help
> serialize a user's compilations. You can use a flag to
> increase levels of
> parallelism if required. You can do this to put them into the
> background:
>
> make -f <makefile name> >& make.log &
>
> To get a bitmore fancy, use the nice command to adjust the
> priority of the
> job:
>
> nice -n <-nn> make >& make.log &
>
> -----Original Message-----
> OK, so I have a corrupted mindset, coming from MVS <grin>.
> But suppose that
> I want to compile a LOT of programs. In MVS, I code up some
> JCL and submit
> it to run later. When it completes, I get a notify to my TSO
> id and look at
> the output in SDSF. I repeat this for however many compiles
> that I want to
> do. Perhaps doing the submissions over a period of time. How
> do I do that in
> Linux (or any UNIX)? In VM/CMS, I remember a CMSBATCH virtual
> machine which
> worked a bit like the MVS initiator.
>

Reply via email to