launchpad-reviewers team mailing list archive

[Robert Collins <robertc@xxxxxxxxxxxxxxxxx>]

On Sun, Sep 12, 2010 at 4:45 PM, Jeroen T. Vermeulen <jtv@xxxxxxxxxxxxx> wrote:
> I don't know what you mean with the part about the comment field.  Perhaps you're thinking of my Librarian branch that deals with discrepancies between the choices of default or slave stores?

No. View the source of a page - at the bottom there is a comment
listing some stats, including the time to render.

> Anyway, the reason I'm not testing locally is memory.  If I create the thousands of packages and templates I need for a realistic test, the differences in caching characteristics between my local system and a production system would probably invalidate the results.

The theory is that everything is fast except for rendering: adding
thousands of templates on on package should be fine : even at 1K per
row, that would be MB's, not 100's of MB's of increased:
 - pgsql buffer
 - storm buffer
 - objects in memory.

As long as you don't drive yourself into swap you'll be able to tell,
with a much lower turnaround time than reproducing on the actual
dataset.

If you aren't in swap, and this branch is a faster - say 15%
consistently faster - then its a pretty good sign.

-Rob
-- 
https://code.launchpad.net/~jtv/launchpad/templates-listing/+merge/35185
Your team Launchpad code reviewers is requested to review the proposed merge of lp:~jtv/launchpad/templates-listing into lp:launchpad/devel.