[net.micro] S-100 bus faster than Multibus?

rconn@brl@sri-unix.UUCP (08/26/83)

From:      Rick Conn <rconn@brl>

The IEEE-696 S-100 bus standard does allow for 16-bit data paths and 24-bit
addresses (16M).

While I don't have much experience with the Multibus, I have heard just
the reverse -- that the Multibus is faster than the S-100, but I'm not sure
why.  From looking at S-100 bus lines on a scope, I've seen a lot of
cross talk occurring, and some of the resultant spikes dip very close
to the active region (3.5V).  Some S-100 motherboards are shielded better
than others, providing less cross talk and (I suppose) greater speed.
I am running S-100 at 4MHz, and the best that I've heard of is 6MHz
(or was it 8?).

	Rick

ken@turtleva.UUCP (Ken Turkowski) (08/30/83)

I've heard that the S-100 bus is faster than the Multibus.  Can anyone
substantiate this claim?  I've also heard that extensions to the S-100
will handle 16-bit data and 16Mbyte address space.

			Ken Turkowski
		    CADLINC, Palo Alto
		{decwrl,amd70}!turtlevax!ken

phil@amd70.UUCP (08/31/83)

The Multibus has not and will not be extended to support 32 bit data
or addresses. There will be a new bus, Multibus II, which supports
32 bits. Several companies, including AMD, have come out in support
of it. For further details, you should wait until Multibus II is
released to the public, sometime late this year. (November?)

I too have heard S-100 is faster than Multibus. I don't know the details
of how S-100 operates, but the Multibus has a theoretical limit of
10 megabytes per second and a practical limit of much less. You can sure
buy a lot of boards for the Multibus, though.

phil@amd70.UUCP (Phil Ngai) (09/02/83)

What is all this stuff about the clock rate of the Multibus? The basic
data transfers are done asynchronously, with a minimum of 200 nS per 16
bit word. If the slave needs more time, the master waits. As soon as
transfer acknowledge (XACK) is returned, the master terminates the cycle.
Bus exchange is done synchronously using a 10 MHz clock.

	Phil Ngai
	Advanced Micro Devices

fair@dual.UUCP (Erik Fair) (09/07/83)

	The Dual System 83/20, which is an S-100 based UNIX system, runs the
bus at 4 MHZ for our 8 MHZ 68000 CPU, and at 5 MHZ for the 10 MHZ 68000 CPU.
Multi-Bus companies? SUN? Altos? How fast are you going?

	Erik E. Fair	{ucbvax,amd70,zehntel}!dual!fair
			Dual Systems Corporation