<div class="gmail_extra"><br><br><div class="gmail_quote">On Fri, Nov 9, 2012 at 12:55 PM, Tres Seaver <span dir="ltr"><<a href="mailto:tseaver@palladion.com" target="_blank">tseaver@palladion.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">On 11/09/2012 09:33 AM, Leonardo Santagada wrote:<br>
> Great! Do you already have a benchmark suite to test its performance<br>
> on pypy? I think the next step would be to run ZODB on pypy and<br>
> measure performance, or is there more code that needs porting?<br>
<br>
</div>I'd like to finish cleanups, including 100% coverage, and think about<br>
Py3k porting for BTrees. We may need to refactor the pure-Python bits<br>
somewhat to allow better JITting by PyPy, too -- likely taht would need<br>
some help from them?<br>
<div class="im"></div></blockquote></div><br>The pypy guys would probably like to help zodb, one thing that they keep repeating is that to help any project they need a self running benchmark, something you can just run and then analise the results. If you guys have it I'm betting the pypy team can help to make it fast. My tip would be to call it the zope/plone database benchmark so that it would carry more the weight of the killer apps using it (that have a good market share in the python world).<br clear="all">
<div><br></div>-- <br><br>Leonardo Santagada<br>
</div>