Silly title, but it's kinda of topical - we just passed the 30th anniversary of some watershed events in Computer Engineering without much fanfare. Lots of things happened in 1985 that are important to the information revolution, but I'll focus on a few things that I think are crucial (at several different layers, mostly in computer system design, some meta-):
1) The first major successes with the Mead & Conway VLSI design methodology. Lots of students who had exposure to Mead & Conway had already worked on real projects in industry for several years. The first fabless semi companies had been founded, and products started to appear. The first Chipset was introduced by Chips & Technologies, which cloned the IBM Enhanced Graphics Adapter (EGA) card in Sep. 1985.
2) The ascendancy of CMOS - up to that time, it was an article of faith that bipolar was superior to MOS in high-performance digital computing. 1985 saw the first challenges to that notion. The first mainstream processor from Intel that was implemented in CMOS was introduced quietly in Oct. 1985, and it was the 80386 (i386 for short).
3) Hennessy & Patterson's ideas were being picked up by the industry, and the move away from CISC and towards RISC and single-cycle execution had begun. The first commercial RISC CPUs - MIPS R2000 was introduced, work on Sun Sparc began, IBM RT/PC and HP PA-RISC were introduced within a few months in 1986, Acorn RISC Machines (ARM) processors were prototyped in April 1985, etc.
4) The first commercial graphics accelerators, from SGI in the form of the IRIS 2000 and the later 3000-series machines. The importance of bit-mapped frame-buffer and graphics performance was becoming better understood.
5) The first commercial Floating Point Accelerator boards were introduced - mainly add-on boards for the PC/AT that ran as Attached Proccessors and helped accelerate applications that were specifically recompiled for the boards. Vendors included Opus, Definicon, Weitek, etc.
6) The first commercial supercomputer using CMOS ASICs was introduced within a few months - this was the ETA10, developed by a spin-off from Control Data Corporation (CDC). There had been CMOS mini-supercomputers that had been introduced over the previous few years, from vendors like Convex, Alliant, SCS, etc., but ETA was the first full-fledged big-iron supercomputer with performance comparable or better than flagship vector supercomputers like the Cray. The flagship ETA10 with 8-way SMP and 143 MHz processors had a peak performance of 9.1 GFLOPS (just a tad below the 10 GFLOPS that had been the original design goal).
7) Steve Jobs founded NeXT in October 1985, with the intention of developing a high-performance desktop workstation. It would use CMOS processors and ASICs, and a Microkernel OS (Mach from CMU). Eventually, Tim Berners-Lee would use a NeXT workstation at CERN in 1990 for the development and release of the first web-server and browser. NeXT would be acquired by Apple in 1996-97, and its OS would eventually become the basis for Mac OS X and iOS.
(There are several more computer-related events in 1985, but the ones listed above are what IMHO are the crucial combination on events, ideas, technologies, etc. that made most of the PC/Internet revolution possible. The things that have survived and dominated are 2) CMOS/x86, 3) RISC/ARM and 7) NeXT (in the form of Apple iOS).)
1) The first major successes with the Mead & Conway VLSI design methodology. Lots of students who had exposure to Mead & Conway had already worked on real projects in industry for several years. The first fabless semi companies had been founded, and products started to appear. The first Chipset was introduced by Chips & Technologies, which cloned the IBM Enhanced Graphics Adapter (EGA) card in Sep. 1985.
2) The ascendancy of CMOS - up to that time, it was an article of faith that bipolar was superior to MOS in high-performance digital computing. 1985 saw the first challenges to that notion. The first mainstream processor from Intel that was implemented in CMOS was introduced quietly in Oct. 1985, and it was the 80386 (i386 for short).
3) Hennessy & Patterson's ideas were being picked up by the industry, and the move away from CISC and towards RISC and single-cycle execution had begun. The first commercial RISC CPUs - MIPS R2000 was introduced, work on Sun Sparc began, IBM RT/PC and HP PA-RISC were introduced within a few months in 1986, Acorn RISC Machines (ARM) processors were prototyped in April 1985, etc.
4) The first commercial graphics accelerators, from SGI in the form of the IRIS 2000 and the later 3000-series machines. The importance of bit-mapped frame-buffer and graphics performance was becoming better understood.
5) The first commercial Floating Point Accelerator boards were introduced - mainly add-on boards for the PC/AT that ran as Attached Proccessors and helped accelerate applications that were specifically recompiled for the boards. Vendors included Opus, Definicon, Weitek, etc.
6) The first commercial supercomputer using CMOS ASICs was introduced within a few months - this was the ETA10, developed by a spin-off from Control Data Corporation (CDC). There had been CMOS mini-supercomputers that had been introduced over the previous few years, from vendors like Convex, Alliant, SCS, etc., but ETA was the first full-fledged big-iron supercomputer with performance comparable or better than flagship vector supercomputers like the Cray. The flagship ETA10 with 8-way SMP and 143 MHz processors had a peak performance of 9.1 GFLOPS (just a tad below the 10 GFLOPS that had been the original design goal).
7) Steve Jobs founded NeXT in October 1985, with the intention of developing a high-performance desktop workstation. It would use CMOS processors and ASICs, and a Microkernel OS (Mach from CMU). Eventually, Tim Berners-Lee would use a NeXT workstation at CERN in 1990 for the development and release of the first web-server and browser. NeXT would be acquired by Apple in 1996-97, and its OS would eventually become the basis for Mac OS X and iOS.
(There are several more computer-related events in 1985, but the ones listed above are what IMHO are the crucial combination on events, ideas, technologies, etc. that made most of the PC/Internet revolution possible. The things that have survived and dominated are 2) CMOS/x86, 3) RISC/ARM and 7) NeXT (in the form of Apple iOS).)
Comment