28 Jul
2012
28 Jul
'12
1:55 p.m.
I wrote:
I've been considering doing such a thing anyhow
for the UDA50, KDA50,
and HSC microcode.
IIRC, they use two 2910 sequencers for two
independent threads of execution.
jimpdavis wrote:
Really? It interleaved fetch / execute with two
sequencers. Quite
sophisticated,
but it does make sense if they both ran the same code and the
operations took some time, or you had multiple drives to service.
As I type this, it is apparent to me, that this is the only way to
go.to reduce part count and eliminate duplicate firmware.
Although the two microsequencers executed code from the same PROMs, they
did not execute the same code or perform the same function. One is used
for the bus interface, and one is used for the drive interface.