Re: DataQueue performance jt400Native 10x slower then RPG

[Colin Williams <colinwilliams007@xxxxxxxxx>]

Have you eliminated jvm startup time from that?

Before you start writing the messages to the dataqueue output a log message
and after you finish output a log message to confirm you are eliminating
the overhead of jvm startup.



On Fri, 10 Apr 2020, 16:35 , <Franz.Rauscher@xxxxxxxxxxx> wrote:

> Hi!
>
> I have a performance problem with my IBM-i scenario.
>
> I have a RPG Client which writes messages to a dataquene. the messages are
> about 9000bytes in size. i am writing 1024 messages.
>
> I have two server implementaions which run on the same IBMi machine one in
> Java and one in RPG
>
> The java version takes about 150ms to read 1024 messges. The RPG version
> only about 15ms. so the RPG version is 10 times faster.
>
> I verified that i user jt400Native and the AS400ImplNative and
> BaseDataQueueImplNative is used so that the native API calls from JVM/Pase
> are used.
>
> Both server implementations are running in al loop so that i can exlude
> any startup perfomance impacts. The timings do not vary over time. RPG is
> alwas 10x faster.
>
>
> Why is the Java/Pase version so much slower??
>
> Greetings,
> Franz
>
> My Environment is: Power S812 64GB RAM (1Core no VIOS). IBM-i 7.4 TR1.
> There is no load on Server
>
>
> --
> This is the Java Programming on and around the IBM i (JAVA400-L) mailing
> list
> To post a message email: JAVA400-L@xxxxxxxxxxxxxxxxxx
> To subscribe, unsubscribe, or change list options,
> visit: https://lists.midrange.com/mailman/listinfo/java400-l
> or email: JAVA400-L-request@xxxxxxxxxxxxxxxxxx
> Before posting, please take a moment to review the archives
> at https://archive.midrange.com/java400-l.
>
> Help support midrange.com by shopping at amazon.com with our affiliate
> link: https://amazon.midrange.com