× The internal search function is temporarily non-functional. The current search engine is no longer viable and we are researching alternatives.
As a stop gap measure, we are using Google's custom search engine service.
If you know of an easy to use, open source, search engine ... please contact support@midrange.com.



Thank you

On Tue, Jun 6, 2023 at 3:57 AM Dsternb <dieter.bender@xxxxxxxxxxxx> wrote:

... it depends:

- if we are talking about typical interactive workload coding is not
important. The system will distribute the workload transparently over the
available cores. Database design might be important (e.g. full tablescans
etc. => parallel database feature mighjt help).

- talking about single stream batchjobs (typical "night-jobs") one cpu is
heacy on work, all others are sleeping - easy to find out by watching cpu
usage. 18 years ago (2005) I was asked if a high-end AS/400 (I don't
remember how it was named in those days) would manage one million complex
transactions (reading some tables, calculating some metrics and writing
some
tables) within 8 hours and up to what limt this would scale well.
Making some benchmarks, I found out that you would need 2 jobs per cpu to
load the system up to its limits. (going into details: I/O is handled
independantly on AS/400 by I/O processors).
As a proof of concept we designed a process, splitting the one million
transactions into a configurable number of jobstreams running in parallel.
(going into details: we generated diffrent cursors and the worker-process
used dynamic SQL)
As the single transactions might come into lock conflicts, we journaled
all
data and used commit for all programms and implemented an automatic
process
to recover the rolled back transactions. All RPG loving guys told us, that
would not perform well. We've cut the overall transaction time to 2
hours;
the answer to scalability was: if we could double the number of CPUs
performance would be double two.

Parallel database feature could help a little bit, but not half as
splitting
one single stream to two and neglectable for the optimal degree of
parallelisation on a system of today.

You might find some discussions on this in java400-l about batch
performance
of java compared to rpg (the rpg loving community disliked my posting on
this - some up to now)

D*B

<Darryl>
I have had conflicting opinions.

Do we have to code specifically for multiple cores or does the OS handle
using multiple cores transparently?

The question is for a single LPAR.

TIA, Darryl Freinkel.
</Darryl>

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxxxxxxxx
Before posting, please take a moment to review the archives
at https://archive.midrange.com/midrange-l.

Please contact support@xxxxxxxxxxxxxxxxxxxx for any subscription related
questions.



As an Amazon Associate we earn from qualifying purchases.

This thread ...

Replies:

Follow On AppleNews
Return to Archive home page | Return to MIDRANGE.COM home page

This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].

Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.