Comments on: libkimap – the genesis http://ervin.ipsquad.net/2009/10/22/libkimap-the-genesis/ White noise from a gearhead Thu, 25 Feb 2010 15:42:09 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: Jan Kundráthttp://ervin.ipsquad.net/2009/10/22/libkimap-the-genesis/comment-page-1/#comment-253 Jan Kundrát Thu, 22 Oct 2009 09:16:34 +0000 http://ervin.ipsquad.net/?p=207#comment-253 Right, getting the UI part right is far from trivial (and my quick hack doesn't do that, it's just a low-level microbenchmark). It will be challenging to measure how long does it take for all items to pop up in the view, at least in my app. Right, getting the UI part right is far from trivial (and my quick hack doesn’t do that, it’s just a low-level microbenchmark). It will be challenging to measure how long does it take for all items to pop up in the view, at least in my app.

]]> By: ervinhttp://ervin.ipsquad.net/2009/10/22/libkimap-the-genesis/comment-page-1/#comment-252 ervin Thu, 22 Oct 2009 09:03:30 +0000 http://ervin.ipsquad.net/?p=207#comment-252 That's probably a point unclear from the post indeed. I didn't scientifically benchmark it <strong>yet</strong> (still need to find time to do it). That said, with a dumb UI slapped on top of the Akonadi resource which uses it, the difference of performance is stellar compared to KMail. Just to give an example, where it took several seconds for listing folders and getting metadata it's now instance.I know it's probably not a satisfying answer for you, it's definitely not a satisfying one for me, but such benchmarks take time to be done well.Since there's apparently interest in such benchmarking though I'll keep a mental note on publishing them here once I get around having them. That’s probably a point unclear from the post indeed. I didn’t scientifically benchmark it yet (still need to find time to do it). That said, with a dumb UI slapped on top of the Akonadi resource which uses it, the difference of performance is stellar compared to KMail. Just to give an example, where it took several seconds for listing folders and getting metadata it’s now instance.

I know it’s probably not a satisfying answer for you, it’s definitely not a satisfying one for me, but such benchmarks take time to be done well.

Since there’s apparently interest in such benchmarking though I’ll keep a mental note on publishing them here once I get around having them.

]]> By: Jan Kundráthttp://ervin.ipsquad.net/2009/10/22/libkimap-the-genesis/comment-page-1/#comment-251 Jan Kundrát Thu, 22 Oct 2009 08:53:13 +0000 http://ervin.ipsquad.net/?p=207#comment-251 Hi, how do you measure performance of the parser and what results have you achieved? I just tried to benchmark my parser [1] and according to QTest, I'm getting 0.33ms on average per loop iteration (it parsers three replies per one iteration, yielding roughly 110 microseconds per response); this is on a few years old Thinkpad T60 (2007-FRG).[1] http://gitorious.org/trojita/trojita/commit/e6afc47d3ec51051adaa748c73528bbd90ccdc2a Hi, how do you measure performance of the parser and what results have you achieved? I just tried to benchmark my parser [1] and according to QTest, I’m getting 0.33ms on average per loop iteration (it parsers three replies per one iteration, yielding roughly 110 microseconds per response); this is on a few years old Thinkpad T60 (2007-FRG).

[1] http://gitorious.org/trojita/trojita/commit/e6afc47d3ec51051adaa748c73528bbd90ccdc2a

]]>