Hi !
Im using Opengl on a Mac with an Ati Nexus 128 (32 mb) , when I try to
setup a GL context with accumulation buffer (no matter what size), the renderer switches to software rendering .
Is this normal on any GL implementation that lacks of harware accumulation buffer?
I belived that in a good GL implentation only those features that harware don’t support are going to be emulated by software.
Is it true ?
The memory requirements are too high for performing accumulation buffering in consumer and mid-range graphics hardware. Even in the high end space, this relatively unused feature hasn’t been worth the cost overhead to implement.
For further information, search for accumulation on this board. There have been previous discussions with several valid arguments both for and against.
Regards,
Glossifah
yep, good implementations do emulate stuff they cant do in hardware… but just if its worth it… and the accumulationbuffers dont worth it, cause you have to do such a huge amount of memcpy and all and no one really uses it… means it would not be very fast anyways if partiallyhardwareaccelerated… and setting up the driver to get always the best out is difficult… but for much stuff wich is used often times it has hardandsoftware combined in good drivers…
on the other side, who knows if your driver is a nice done one 
cheers…