> Can anyone else run a couple tests too and confirm some of
> the finding?
The latest checkins lowered my 100 iteration test from 536 ms to 443 ms.
Significant, but nothing like the first one. Still, since it's not posted
yet anyway, I'd advocate running "ant dist" again. ;-)
-- Scott
> Last patch wasn't soo spectacular, but the first second to last was
> pretty good again. Got a build on my laptop from 14.11 to 11.41 :-)
> I was working on the laptop, so results aren't too reliable though.
>
Thank you.
>> That is my last easy patch, I will try to go inside
>> Canonicalizer20
Here it is my last easy patch,
It caches the getProviderIsInClass it seems to be called a lot, it
seems that the first cache of translateURI..JCE is not enougth. All in
All in my test cases it seems to speed the things a little, from 1.4sec
to 10ms, but i think is not a great thing in a bigger
Here it is my last easy patch,
It caches the getProviderIsInClass it seems to be called a lot, it
seems that the first cache of translateURI..JCE is not enougth. All in
All in my test cases it seems to speed the things a little, from 1.4sec
to 10ms, but i think is not a great thing in a bigge
[EMAIL PROTECTED] wrote:
Thank you for including the code, it was a great honour.
I'm currently working in other optimization, again in Transforms class, it
seems that this class is invoked a lot(I´m optimizing a Liberty
implementation, not based in opensaml, for Ericsson). The method is:public T
> Thank you for including the code, it was a great honour.
No, thank you. That rocked.
-- Scott
Erwin van der Koogh wrote:
But these results are quite a bit better so I think Berin won't mind
including them in 1.1 ;-)
. Nope.
There is a new "1.1" in xml.apache.org/security/1.1
I have moved the first one into the "superceded" directory.
Some stuff came up tonight that meant I couldn't up
Thank you for including the code, it was a great honour.
I'm currently working in other optimization, again in Transforms class, it
seems that this class is invoked a lot(I´m optimizing a Liberty
implementation, not based in opensaml, for Ericsson). The method ispublic Transform
item(int i) is ca
Thank you for including the code, it was a great honour.
I'm currently working in other optimization, again in Transforms class, it
seems that this class is invoked a lot(I´m optimizing a Liberty
implementation, not based in opensaml, for Ericsson). The method ispublic Transform
item(int i) is ca
I'm sure I don't have a vote, but I'd like to chime in and express my
hope that these changes make it into 1.1. I ran some tests similar to
Scott's and got good results:
Anyone who contribute code or other stuff has a vote ;-)
But these results are quite a bit better so I think Berin won't mind
I'm sure I don't have a vote, but I'd like to chime in and express my
hope that these changes make it into 1.1. I ran some tests similar to
Scott's and got good results:
Avg Time (ms): 372 - an old build w/Sun provider
Avg Time (ms): 395 - 1.1 pre fix w/Sun provider
Avg Time (ms): 225 - 1.1 pos
+1 from me.
--- Berin Lautenbach <[EMAIL PROTECTED]> wrote:
> > "2.0" fake release: 840 ms avg
> > 1.1 RC Berin put out: 915 ms avg
> > Build of CVS Head: 540 ms avg
> >
> > Seriously. Maybe I'm just a special case and the patches target my use
> > case, but, wow.
>
>
> Wow!
>
>
> "2.0" fake release: 840 ms avg
> 1.1 RC Berin put out: 915 ms avg
> Build of CVS Head:540 ms avg
>
> Seriously. Maybe I'm just a special case and the patches target my use
> case, but, wow.
Wow!
It's interesting in other respects too. The library makes extensive use
of XPath calls all
I looped over a Junit test I put into OpenSAML that reads in a SAML
assertion, signs it (RSA/SHA1), dumps it, reads it back in, verifies it, and
takes a time reading. Since the whole thing was CPU bound, the numbers are
practically identical across runs.
I used my laptop, a 1.8 GHz with 512MB, XP,
> If the increase is anything over "I think it runs faster" I'd say we
> give it a try.. how about you Berin? I spoke to Axl tonight and he's
> out in the middle of the bush in deep (South) Africa (no kidding) and
> probably won't be able to respond for another few days.
. Tell you what - if
> Scott, could you try and see if you can get some feel for the speed
> increases? They don't have to be perfect, but rough estimates
> would really help to make a decision.
I've got a copy of the 1.1 jar in our cvs now. I can try to get some timings
on SAML signing with that and then do a build
Heh, is there time to get this into 1.1 since it's not officially out yet?
I don't know. I don't particulary like to make changes to it, but if this
turns out to be a decent improvement I think we should give it a try.
I didn't have time to properly test 1.1 until tonight and I did most of the
t
Heh, is there time to get this into 1.1 since it's not officially out yet?
We're getting ready to ship a new Shibboleth release. ;-)
-- Scott
18 matches
Mail list logo
