30 Jun
2010
30 Jun
'10
6:10 p.m.
On Mon, 2010-06-21 at 20:46 +0100, Lawrence Wilkinson wrote:
On Mon, 2010-06-21 at 14:25 -0400, Dave McGuire
wrote:
Ok, diagnostics complete ok and I'm on to getting it to run a program.
Basic execution and single-instruction-stepping works. Just fixed a
VHDL problem related to RX opcode handling, but from now on the majority
of problems are likely to be microcode-related.
The main memory problem turned out to be the main clock too slow by a
factor of 4. The memory runs asynchronously relative to the CPU, and it
had the correct timing, so it was effectively responding in a quarter of
the normal access time which messed things up. Happy Happy Joy Joy.
If anyone has a Digilent Spartan3 board and would like to play with it,
I can try to come up with a configuration that will work without
external switches. You won't be able to do much!
--
Lawrence Wilkinson lawrence at ljw.me.uk
The IBM 360/30 page http://www.ljw.me.uk/ibm360
On 6/21/10 7:03 AM, Roger Pugh wrote:
Ah, yes. VCF was its debut. I haven't got a page up yet, but you can
see Andrew's photo at:
http://www.flickr.com/photos/carrierdetect/4718559358/
Basically, it's a 1000K gate Spartan 3 which contains the microcode
store (2k x 55) and 8k x 9 of main memory. Front panel switches are
physical rotaries and pushbuttons; blinkenlights are via a VGA display
for now.
Right now it fails the microcode startup diagnostics with a main memory
parity error, probably due to a race condition somewhere (the
memory /used/ to work).
I quite fancy an IBM 360 / 30 FPGA board now i
know a little about big
iron!
Urr? Is someone working on that?