Very encouraging indeed! Great work everyone :o)
On 8/7/07, Craig Andera <[EMAIL PROTECTED]> wrote:
>
> OK, so I got the results of the test, and the results are very, very
> good. With the new caching code in place and John's page optimizations, we
> are, on average, **faster** than FlexWiki 1.8 by about 30% (average time
> to last byte of 636ms versus 939ms).
>
>
>
> There are a whole boatload of caveats here, including:
>
>
>
> · Some pages are still slower in 2.0 than in 1.8, so the average
> is not super meaningful.
>
> · This is against the www.flexwiki.com corpus.
>
> · I ran this in a virtual machine on my laptop.
>
> · A small percentage of pages failed completely in either 2.0, 1.8,
> or both.
>
> · 2.0 is generally slower before a page is cached, and generally
> faster after a page is cached.
>
> · So far, only default.aspx has been tested – other things like
> RecentChanges.aspx, etc. haven't been measured.
>
>
>
> In other words, your wiki running 2.0 may be slower or faster than
> 1.8depending on your data, hardware, etc. etc. But I still feel pretty good
> about these numbers.
>
>
>
> I'll probably try one or two more optimizations. I'll also want to test
> some pages besides default.aspx, and I want to rerun the tests a few times
> to verify the results. Of course, I need to strike a balance between adding
> the absolute minimum of caching code I can (we had an absolute nightmare of
> a time stabilizing the caching code in 1.8), and performance. So if I
> don't see a significant speedup with the next round of cache, I'm going to
> stop where I am now.
>
>
>
> Very encouraging!
>
>
>
> *From:* [EMAIL PROTECTED] [mailto:
> [EMAIL PROTECTED] *On Behalf Of *Craig Andera
> *Sent:* Monday, August 06, 2007 8:57 PM
> *To:* 'FlexWiki Users Mailing List'
> *Subject:* Re: [Flexwiki-users] FlexWiki 2.0 Progress Report
>
>
>
> Thanks – it is indeed very gratifying!
>
>
>
> But credit has to go to John Davidson, who did some very important
> page-level optimizations on top of the core engine stuff.
>
>
>
> Also, it's still slower than FlexWiki 1.8. J
>
>
>
> *From:* [EMAIL PROTECTED] [mailto:
> [EMAIL PROTECTED] *On Behalf Of *Jimmy Sieben
> *Sent:* Monday, August 06, 2007 5:18 PM
> *To:* FlexWiki Users Mailing List
> *Subject:* Re: [Flexwiki-users] FlexWiki 2.0 Progress Report
>
>
>
> That is very encouraging news, that a feature like caching could be added
> as an orthogonal change.
>
>
>
> Nice work.
>
>
>
> *From:* [EMAIL PROTECTED] [mailto:
> [EMAIL PROTECTED] *On Behalf Of *Craig Andera
> *Sent:* Monday, August 06, 2007 4:03 PM
> *To:* 'FlexWiki Users Mailing List'
> *Subject:* [Flexwiki-users] FlexWiki 2.0 Progress Report
>
>
>
> Oh wow – I'm rerunning the test now, and the numbers are coming in
> somewhere in the 500-1000ms range. Still slower than FW 1.8, but much,
> much faster than it was without caching. The test will take a few hours to
> wind down – I'll post results tomorrow. The best part is, the caching
> provider is actually only about 80 lines of code right now.
>
>
>
> It'll be interesting to see where the next bottleneck is.
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc.
> Still grepping through log files to find problems? Stop.
> Now Search log events and configuration files using AJAX and a browser.
> Download your FREE copy of Splunk now >> http://get.splunk.com/
> _______________________________________________
> Flexwiki-users mailing list
> Flexwiki-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/flexwiki-users
>
>
-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Flexwiki-users mailing list
Flexwiki-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/flexwiki-users
