On 5/25/05, Viktor Pracht <[EMAIL PROTECTED]> wrote:
> Hi
> 
> We were discussing the VGA nanocontroller on IRC, and figured out its
> usefulness depends on the available RAM bandwidth. Extreme examples:
> * An 8 bit path shared for instructions and code results in about 2 FPS.
> * Independent 32 bit paths for instructions and data result in 32 FPS.
> 
> So, the question is (mainly to Timothy): how could the raw VRAM
> interface look like, how could the cached RAM interface look like, and
> what bandwidths can we expect from them?
> 
> And, considering that the 4 RAM chips prefer long data streams, which is
> the exact opposite of heavily self-modifying nanocontroller code, what
> are the latencies for random access?


I've been thinking about it, and while I really like the idea of
instructions being lookup tables in RAM, it may not give us the
performance we need.  Things will already be slow.  SO, I suggest we
develop a simple processor and use an FPGA RAM block to store both
nearly 500 instructions and the register file.

_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)

Reply via email to