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

List:       kde-core-devel
Subject:    Re: 0 or 0L for empty pointers?
From:       Thiago Macieira <thiago () kde ! org>
Date:       2006-06-24 23:30:48
Message-ID: 200606250130.48585.thiago () kde ! org
[Download RAW message or body]


Lubos Lunak wrote:
>On Friday 23 June 2006 18:17, Allan Sandfeld Jensen wrote:
>> On Friday 23 June 2006 17:20, André Wöbbeking wrote:
>> > Any numbers on this especially with recent compilers?
>>
>> Apple claimed to gain 20% smaller binary and 5-10% speed-up by
>> disabling it in KHTML. This was half a year ago, and they are using
>> their Apple's branch of gcc.
>
> That seems very unlikely. Are you sure you're not confusing it with
>exceptions? RTTI should involve next to zero runtime overhead when not
> used and as it's just little data per class 20% size seems way over the
> top.

Technically, disabling RTTI disables exceptions, since exceptions require 
RTTI to catch properly. So, you get the benefits of -fno-exceptions here.

A 20% decrease in binary size is possible for the removal of exceptions 
and RTTI. Speed improvements? Only if you were using gcc 2.95, which we 
don't support anymore.

-- 
Thiago Macieira  -  thiago (AT) macieira.info - thiago (AT) kde.org
  thiago.macieira (AT) trolltech.com     Trolltech ASA
    GPG: 0x6EF45358                   |  Sandakerveien 116,
    E067 918B B660 DBD1 105C          |  NO-0402
    966C 33F5 F005 6EF4 5358          |  Oslo, Norway

[Attachment #3 (application/pgp-signature)]

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