>>>> "Ethan" == Ethan Dicks
<dickset(a)amanda.spole.gov> writes:
Ethan> On Thu, Apr 15, 2004 at 05:13:33PM -0400, Paul Koning wrote:
Tom> There is a (modified) version of BSD 2.9 that runs on the
Tom> Pro350, and is freely available
> Hm... interesting. A Pro-380 is close enough to
a -350 that it
> might also run there; if not then it shouldn't be hard to fix.
Ethan> Well, the Pro-350 uses an F-11 CPU (same as the 11/23 and
Ethan> 11/24), and the Pro-380 uses a J-11 CPU (same as the 11/73).
Ethan> The largest architectural difference is that the J-11 supports
Ethan> split I&D space. As a result, the Pro-380 should be able to
Ethan> load 2.11BSD which, I _think_, already supports the MSCP disk
Ethan> controller in the Pro series (that's the nature of the patches
Ethan> to 2.9BSD for the Pro-350... adding support for MSCP
Ethan> controllers).
The PRO disk controller isn't an MSCP controller, not within 100
miles. It's a one sector at a time programmed I/O controller,
somewhat on the same level of sophistication as a PC floppy
controller.
The most obvious difference between a Pro-380 and other J-11 based
systems is that the Pro clocks the CPU at 10 MHz and the others use
around 18 MHz. The reason is that the Pro system controller ASIC used
a synchronous design, so the CPU clock had to be an integer multiple
of the bus clock. The bus clock frequency was set at 10 MHz, and the
assumption was that Harris would deliver on its promise of a 20 MHz
J-11 chip. No such luck -- the best they ever managed was 18 MHz, and
that forced the Pro to the next lower integer multiple -- 10
MHz... :-(
Another difference is that the Pro-380 has a bunch of stuff on the
motherboard that's on option cards in the Pro-350 -- the graphics in
particular. And the memory isn't on the CTI bus, it's on a dedicated
card. Some vague memory says the graphics has more goodies, but I
don't remember. It's all in the manual on
bitsavers.org...
paul