Hosted Mode 1.5 M2 RPC performance

[gregor]

Just been doing some tests on 1.5. M2 RPC in hosted mode (basic return
string on button click). Results (rough by eye):

1.4.61 1.5 M2
===== =====
First run: 0.75s 2.25s
Refresh: 0.25s 0.5s

Anyone else noticed this or is it just me (or Intellij)?

regards
gregor

[Bruce Johnson]

Hi Gregor,

Thanks for the numbers. Being quantitative about it makes it a *lot* easier for us to investigate the issue with a concrete idea of the effect it's having on you.

We have heard similar reports from other people, too, so it's something we are beginning to investigate now.

--
Join us at Google's biggest developer event of the year
May 28-29, San Francisco
http://code.google.com/events/io

[gregor]

Hi all,

I've done some further tests and can report the following results
which are very reassuring. I did ten runs for each of the eight
situations (1.4.61 Hosted first run & refresh, 1.4.61 deployed first
run & refresh, 1.5.M2 Hosted first run & refresh, 1.5.M2 deployed
first run & refresh) and report the average timings in milliseconds.
The test is a simple RPC method to return a String called from a
button click. Times were taken at three points:

-Just before prepare RPC async proxy (prepare)
-Just before RPC method call (call)
-On string retrieved in onSuccess(..) (complete)

The figures are therefore: prepare = call - prepare, and excecute =
complete - call. The deployed tests involved restarting JBoss 4.2.1GA
each iteration.

1st run refresh
prepare execute prepare execute
1.4.61(H) 587.5 187.4 453.1 26.5
1.4.61(D) 0 132.9 0 31.6
1.5M2(H) 1735.5 553.2 1020.5 204.6
1.5M2(D) 0 107.8 0 28.3

Clearly 1.5M2 is not behaving itself in hosted mode, but once deployed
it seems if anything faster than 1.4.61. It would be nice to get the
1.5 hosted mode RPC sorted out, but what really matters to me is it's
deployed performance, and this makes me very happy ;)

regards
gregor

On Apr 12, 3:03 pm, "Bruce Johnson" <br...@google.com> wrote:
> Hi Gregor,
> Thanks for the numbers. Being quantitative about it makes it a *lot* easier
> for us to investigate the issue with a concrete idea of the effect it's
> having on you.
>
> We have heard similar reports from other people, too, so it's something we
> are beginning to investigate now.
>

> On Fri, Apr 11, 2008 at 9:52 PM, gregor <greg.power...@googlemail.com>