Mail Archives: djgpp/1998/09/19/14:04:27
Reply-To: | <arfa AT clara DOT net>
|
From: | "Arthur" <arfa AT clara DOT net>
|
To: | "DJGPP Mailing List" <djgpp AT delorie DOT com>
|
Subject: | RE: Floating/fixed point
|
Date: | Sat, 19 Sep 1998 19:02:48 +0100
|
Message-ID: | <000001bde3f7$b5dc8fc0$9f4e08c3@arthur>
|
MIME-Version: | 1.0
|
In-Reply-To: | <Pine.GSO.3.96.980918214250.10258D-100000@sutf17.reading.ac.uk>
|
Importance: | Normal
|
> > I'll have to see your benchmark code to be convinced. Float-integer
> > conversion is very slow, as are if() statements. So no, people
> wouldn't use
> > fmul instead of mul.
> >
>
> But,(unless I'm mistaken) according to the intel pentium data sheet, a
> 'mul' instruction uses the FPU pipeline to execute an integer multiply;
> and then converts it back to an integer again... So why is a 'mul'
> implicity faster...
Question: why use an FPU to do integer math? I could understand using a FPU
to do divides (where there is a remainder), but why multiplications?
It's all very well having the conversion float->integer and integer->float
on a hardware level, but on a software level it can be much slower. Don't
ask me why, but it happens.
James Arthur
jaa AT arfa DOT clara DOT net
ICQ#15054819
- Raw text -