[Beowulf] Julia Language

John Hearns hearnsj at googlemail.com
Tue Sep 19 14:04:01 PDT 2017


I was going to say something deep about Python here, but again I am not
qualified. I program in Python every day - today I was writing a PBSPro
hook in Python, and it is supremely adapted for tasks like that. I can also
see why the bioinformatics community loves Python.
But would I start coding a CFD simulation from scratch in Python? Or
LS-DYNA Errr.....

Julians take pride in it being "Julia all the way down" - referred to as
the 'two languages problem' - there is not need  to bolt on C code to get
the performance.

Look over on the Julia discourse board. There are people there arguing the
merits of setting up the correct types for Angles such that scientific
calculations are performed naturally. The intention being that in the
future there is a rich set of Angle types available in the units an
engineer or scientist wants to use. They really do care about gettint this
right.

Also look at the Julia package system. Packages live on GitHub, and you
just clone them from there. It is stupidly easy.

One of the weaker parts in Julia in my opinion HAS been the concept of
having your personal Julia installation, in your home directory, and being
to add packages. This goes against the grain of our normal Soviet-style (*)
centralist HPC cluster software installs.
This is being dealt  with by the new Pkg3 system
https://github.com/JuliaLang/Juleps/blob/master/Pkg3.md

The real point I am making here is that you have to move beyond the model
of having a  Julia install on your MacBook to usefully using Julia on a
national HPC facility, seamlessly.


(*) Yes, I know I am teasing the assembled audience here. Of course you can
set PATHs and have packages installed in a shared location.






On 19 September 2017 at 22:17, John Hearns <hearnsj at googlemail.com> wrote:

> In reply to Jeff, I am no Julia expert.
> However I am going to stick my neck out here. Yes, learn Julia.
> Firstly there is a definite 'buzz' in the Julia community and you
> encounter science and engineering experts at every turn.
> Secondly, where I come from. My background is in high energy physics, from
> the era of old style FORTRAN.
> I gather that HEP codes are currently written in C++, probably by
> physicists who are not experts.
> I would really like to see clean, modern language used in physics
> simulations.
> I would like to use the term "strongly typed" - but in fact Julia is
> anything but.
> I gather the correct term is "duck typed" - which I rather like.
>
>
>
> On 19 September 2017 at 22:10, Jeffrey Layton <laytonjb at gmail.com> wrote:
>
>> John,
>>
>> Have you done much Julia coding? Can you talk about your experience?
>>
>> I have threatened to learn it for a while but your post has prompted me
>> to finally start learning Julia :)
>>
>> Thanks!
>>
>> Jeff
>>
>>
>> On Wed, Sep 13, 2017 at 7:43 AM, John Hearns via Beowulf <
>> beowulf at beowulf.org> wrote:
>>
>>> I see HPCwire has an article on Julia.  I am a big fan of Julia, so
>>> though it worth pointing out.
>>> https://www.hpcwire.com/off-the-wire/julia-joins-petaflop-club/
>>> Though the source of this seems old news - it is a presentation from
>>> this year's JuliaCon
>>>
>>> JuliaCon 2018 will be talking place at UCL in London so mark your
>>> diaries. Yours truly should be there.
>>>
>>>
>>> _______________________________________________
>>> Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin Computing
>>> To change your subscription (digest mode or unsubscribe) visit
>>> http://www.beowulf.org/mailman/listinfo/beowulf
>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.beowulf.org/pipermail/beowulf/attachments/20170919/8e3122bc/attachment.html>


More information about the Beowulf mailing list