9 of 10Image
DEC Rainbow 1001
If the VAX 11/780 was the high-water mark of DEC's dominance, the Rainbow 100 was a bad omen for the future. An idiosyncratic design that married a Z80 chip with an 8088, it managed to be compatible with very little and annoy users very much.
Taken as a stand-alone computer, the hardware was built to DEC's usual high standards; in the PC market, it was a bad fit. The user couldn't format floppy disks, which had to be bought pre-formatted from DEC. It wasn't PC compatible, so it couldn't run much software. It had an odd keyboard layout. It cost $3,000.
There are many signs that DEC itself wasn't too keen on the Rainbow, as little marketing was done and the sales teams got more money for less work by selling minicomputers. Plus, it was mostly built around technologies that DEC hadn't invented, which, in a fiercely proud engineering company, rarely endears a product.
The result, in the end, was that DEC missed the microcomputer revolution and was buried, much as most of the mainframe manufacturers had dismissed minicomputers 20 years before.
Photo credit: David Alcubierre/Wikimedia
DEC Alpha processor
In 1992, DEC had one final burst of genius — the DEC Alpha processor.
Although the company had had considerable success with the single-chip versions of its minicomputers, the architectural limitations had made them obsolete in any format. The Alpha chip was designed on RISC principles and was intended to support a thousandfold increase in performance over its projected 25-year lifespan.
There is evidence that the Alpha would have succeeded. The first versions, the 21064 line, produced unparalleled performance for the CMOS process on which they were built, foreshadowing Intel's later success in pushing that technology. Later versions introduced on-chip secondary caches, high-speed out-of-order execution and on-chip memory controllers. They would have supported simultaneous multi-threading, had the design not been cancelled after Compaq's 1998 purchase of DEC.
Although the Alpha continued as a server chip for some time, software and hardware support dropped off and it was finally killed when Intel persuaded its customers that the Itanium was going to be the muscle chip of choice.
Photo credit: Dyl/Wikimedia Commons
DEC StrongARM processor
One of DEC's last pieces of innovation was the StrongARM, a curious combination of ARM's low-power processor architecture and elements of the Alpha's high-performance design. The chimera was surprisingly successful, finding its way into PDAs and set-top boxes and helping establish the ARM architecture in a number of new — and subsequently very important — niches.
The design proved more important as a bargaining chip than a processor chip, as it was passed to Intel as part of a complex deal to settle an intellectual-property dispute. Intel developed the design into the XScale processor, which was in turn sold to Marvell Technology Group in 2006, prior to Intel's announcement of its Atom low-power x86 processor.
Photo credit: Dirk Oppelt/Wikimedia Commons
Get the latest technology news and analysis, blogs and reviews delivered directly to your inbox with ZDNet UK's newsletters.