[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: --enable-final again
From:       Mario Weilguni <mweilguni () sime ! com>
Date:       1999-05-28 15:23:38
[Download RAW message or body]

Am Fri, 28 May 1999 schrieb Alex Zepeda:
>On Thu, 27 May 1999, Mario Weilguni wrote:
>
>> I did some testing with and without --enable-final, and I must say on Intel
>> machines it's definitly not worth the efforts:
>
>> Considering all the troubles you get from --enable-final (huge memory
>> usage, C++ vs C hassle, include file dependencies e.g. those stupid
>> X11 headers...) it's not worth 14 kb. I'm sure there are better ways
>> to reduce the memory footprint.
>
>Memory (RAM) useage is not in any way directly related to the size of the
>binary on disk.  Smaller binaries might just have poorer code.  The true
>test is when you run the binary.  Does it use less RAM, does it run
>faster? etc etc etc
>

The code is tighter packed (aligned) due to some compiler flags, making
execution slower on 486er systems (who really uses a 486er and KDE?), and some
other optimizations (-fno-exceptions). When I said the library is smaller I of
course meant code, no symbols or such, because I stripped the libraries to
check this.

And why does everbody emphasize speed? Speed is mostly a null-issue for
GUI-stuff, and most of the speed of kde-programs is limited by disk access and
of course the user - you can't type faster because code is optimized for a P-II.

Mario

--
Waste not, get your budget cut next year.

[prev in list] [next in list] [prev in thread] [next in thread]