[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [openrisc] Re: PC as GPR?



> The purpose of a CPU is not to execute instructions as fast as possible,
> it is to run software.  It shouldn't be fast but efficient.  There is a
> balance to find between simplicity (hardware view) and usability
> (software view).  Currently the hardware view is overrepresented.  Maybe
> if I keep annoying you we will end up with something better :)

Sure, MHz is not important, overall performance is important (or power/size
if you have other requirements).

But it is hard to come from an idea to verified working HW. Not to mention
all the software. If I'd know how exactly windowed register file is
implemented in Sparc GCC, we would have windowed register file. ;-)

So bottom line is to actually have working systems, even though they might
not be the best. Of course we will correct all the mistakes in next
generation.

If you think there is something that can be added easily (to HW, SW etc),
start working on it. ;-)

For example right now, the interesting topic is FP. How to add it to the
architecture (ok, there are floting point registers and instructions, but
this is not implemented yet in HW, so it is subject to change - once it will
be implemented, it will be harder to change ;-). Of course somebody also
needs to do a real fast pipelined FPU.

regards,
Damjan


--
To unsubscribe from openrisc mailing list please visit http://www.opencores.org/mailinglists.shtml