On Tue, 1 Feb 2000, Erik Thiele wrote:
> a) use palemu target and get a 8 bit visual, and use ggiPutBox with it
> b) use ggiCrossBlit to convert 8bit memory-target to 16 bit visual.
> which one is faster?

ggiCrossBlit can be optimized for each kind of back-end target to use
features in that target to speed things up.  Palemu probably does its
best to optimize the transfer, but it is (slightly) limited to working 
through GGI primitives or directbuffer.  I suppose it can be optimized in
some ways (by using extension libs) for the back-end target, but it is 
less likely to be since that introduces target specific code into what's 
intended as a generic utility.  

So, your mileage may vary, depending on how thoroughly the target for 
your display is implemented and optimized -- either could be slower on 
a given target -- but I would expect that ggiCrossBlit would have better 
potential to be faster and that when it is slower, that's just a matter 
of unoptimized code that can/should be written better.


Reply via email to