[Qt-interest] why does QGrahicsView framework uses floats for coordiantes ?

[Jason H]

FPU multiplication is fast. (Scale, transform, any matrix multiplication etc)
FPU addition is slow (translate)

GPUs can (in theory) be 
used for acceleration (CUDA?).

The "lighter" comment is because 
GVIs are elemtary shapes. They don't have QObject overhead. But they 
don't do a lot that widgets do. QWs derive from QObject and all that 
jazz.

And you failed to define "faster"


----- Original Message ----
From: Pritam Ghanghas <pritam_ghanghas at infosys.com>
To: "qt-interest at trolltech.com" <qt-interest at trolltech.com>
Sent: Wed, June 30, 2010 8:17:42 AM
Subject: [Qt-interest] why does QGrahicsView framework uses floats for coordiantes ?

Hi

why does QGrahicsView framework uses floats for coordinates ?
is it just to have larger maximum value possible or some other reason?
and
is there a performance penalty because of that if my hardware doesn't
have a FPU? 
I found my Qwidgets to be signifacntly faster than QGraphicsItems. which
is unexpected considering docs say that QGaphicsitems are lighter than 
Qwidgets.


-- 
Thanks,
Pritam



_______________________________________________
Qt-interest mailing list
Qt-interest at trolltech.com
http://lists.trolltech.com/mailman/listinfo/qt-interest