On Sun, Nov 4, 2018 at 4:30 PM 'Paulo Matos' via Racket Users <
racket-users@googlegroups.com> wrote:

> Curious about what your places are doing.
> How come you started so many places without seeing this:
> https://groups.google.com/d/msg/racket-users/oE72JfIKDO4/zbFI6knhAQAJ
>
> Could you share what each of the places do?
>

Each process is:

0. Requesting an identifier (integer) from the queen process over a
(distributed) place channel,
1. Extracting multiple subsets of data from a database based on that id,
2. Performing a series of calculations on each subset of data, and
3. Communicating the results of those calculations back to the queen
process over the place channel.

The workers each hold one reference to the query-db. The results that come
back to the queen are thrown on a queue, which is processed and written to
a separate database.

All communications between distributed processes are happening as
serializable structures over the place channels.

I don't believe I'm seeing the behavior you're describing, but I would have
to investigate more closely. I believe, given the scale of data that I'm
processing, that I'm seeing a roughly linear speedup... but perhaps that
reaches a plateau.  I'll have to revisit the thread you referenced and
investigate more closely; I only gained access to the machine Thursday.

Cheers,
Matt

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to