On Mon, Aug 10, 2026 at 3:56 PM Mark Linimon via cctalk <[email protected]> wrote: > > I am being slammed by email today so I will have to be brief. > > I was hired in 1987 by Mizar, Incorporated, to bring up AT&T > SVR4 v2 Unix (sic) for them to sell as a VMEBus product. IIRC my > development machine at that time had a 321SB board in it. My > understanding is that the only difference between it and the 3B2 > machines was the bus interface. VMEBus was going to be "the future" > of expansion buses for industrial computers at that time. > > Even by the standards of the time this CPU was underpowered. IIRC > we made our own boards based on it. (VMEBus master boards made by > manufacturers who did not specialize in VMEBus, as we did, tended not > to be well-designed, and only implemented a minimum of features. > Their multiple bus master support was non-existant, for instance.) > > The product did not sell. After a year they switched me over > to the design of the next "future" processor, a sparc32 board. > That did not sell well, either. > > Then we tried to sell the idea of MIPS-based boards to interested > customers but not much happened. > > During that time Mizar was making most of its money off the 68k > product lines. The 68000-, 68010-, and 68020- VMEBus machines > were still selling well. We were selling these machines either > to run bare-metal, PDOS, or vxWorks. (The latter was what I was > assigned to the whole time.)
I think Motorola did the lion's share of VME through longevity, although some early UNIX workstations like Sun2/3 and SGI also used it as an expansion bus. I grew up down the street from Motorola Computer Group and have a small collection of oddities. The MVME boards are interesting UNIX machines (System V/68K and 88K) and still supported by NetBSD (68k). Veering slightly off topic, VME was somewhat displaced by CompactPCI (and to a lesser extent ATCA/uATCA), although it probably will outlive everything else (it shows up in weird places with long lifetimes like ASML and other semiconductor equipment, particle accelerators, M1 Abrams tanks). These days it seems like the nearest neighbor is a SOM. Maybe the cost/difficulty of creating custom boards has just dropped low enough. > > Older people will recall that Motorola "lost its mojo" around that > time and were having trouble scaling up the 68040 to faster clock > speeds (this is because much of their technical design team walked > out the door to start Ross Technology.) So we were trying to figure > out what we could offer to stay competitive. > > tl;dr: we didn't. I left at the end of 1989 after running my mouth > to management too much. > > I can go on a treasure-hunt to see what I still have from those days > but it won't be this week. I do have manuals and maybe even a board > or two. The manuals for VME boards tend to be gold, usually full details on the bring up and register space, I guess because the customers were sophisticated. > AFAIK no one else ever even got AT&T-based processors out the door > at all. I haven't heard of other uses of the WE32k. There were certainly many choices in that era. > mcl
