[Xcb] [PATCH] Making XCB less of a memory hog

[Jamey Sharp]

On Sun, May 13, 2007 at 11:42:26AM +0200, Tilman Sauerbeck wrote:
> Hi Jamey,
> I suck. This will teach me to come up with hard evidence before
> submitting invasive patches :|
> ...
> Interesting. Is there a benchmark available that I can use to compare
> the runtime performance?

What, you expect me to have hard evidence for *my* invasive API changes?

After I wrote that e-mail I realized that I'm not sure I actually
*measured* a performance improvement; I may have just theorized there
would be one. Sorry. :-) I'm sure I measured a code-size improvement,
which is what I cared about at the time.

That said, recently I've been working with the noop-xcb benchmark, which
you can find here:
	http://minilop.net/svn/jamey/trunk/test/xlib
Since the NoOperation request is trivial for the server to process,
XCB's request-issuing performance shows more clearly in this benchmark
than it would when testing with other requests. It seems to be the worst
possible case for XCB performance, in fact.

More on this in my next mail, whenever I get it written.

--Jamey
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://lists.freedesktop.org/archives/xcb/attachments/20070516/a23a1342/attachment.pgp