I'm now running MAXI030 at 32MHz! It seems this is the sweet spot. I really want to get to the bottom of why the VHDL isn't able to run faster, despite Quartus reporting a fMax of around 80MHz. I suspect there's something subtle I don't understand going on. My address decoder being clocked is not helping I'm sure, but implementing it in combinational logic is a nightmare.
Linux bootup is around 2m 30s, which is pretty respectable I think. This includes firing up OpenSSH, inetd, Apache2 and a few other things.
The next thing I will look at is probably improving the DRAM performance by investigating how the /STERM '030 input can be used to load 4 longword chunks into the CPU cache. I suspect it will be a big job, but it should be fun.
Discussions
Become a Hackaday.io Member
Create an account to leave a comment. Already have an account? Log In.