5 ms·
Many interesting chips failed. My favorites: - National 32032 CPUs. Basically a VAX-on-a-chip, they were relatively slow and pretty buggy. (They were one of th
by kabdib 9y ago
Many interesting chips failed. My favorites:
- National 32032 CPUs. Basically a VAX-on-a-chip, they were relatively slow and pretty buggy. (They were one of the processors we considered using in the Atari ST, but the 68000 won out, thank goodness).
- INMOS Transputer. One of the early "just apply +5V and ground and happy computing!" systems with zero support chips needed, quite easy to put into a grid of compute elements. Very interesting nibble-based instruction set, worth studying. Unfortunately the high-level language INMOS was pushing (Occam) was really strange, and Transputer performance was never really great. There were also some cool microcode bugs (you could lock a Transputer up for minutes by executing a bit-shift instruction with a very large count).
- RCA 1802. Another early microprocessor. Not really a failure, but not a huge success, except for its radiation-hardened variants, which have flown on many space missions.
- Just about any floating-point coprocessor chip. Ugh. Thank goodness the early Pentiums stopped that madness [1].
[1] I heard that Intel wanted to charge for doing floating-point operations on Pentiums. The idea was that you'd pay to have an "enable" fuse blown, which would purchase (say) 100M operations before a disable fuse got blown. There'd be something like 64 fuses, and the last one would enable floating ops forever. The rationale was that the only people who really cared about floating point were folks running spreadsheets . . . and those people obviously had money and would pay for performance! People hated this idea; then 3D gaming started to be popular (with Quake, et al) and suddenly flops were in general use and it would have been a really awful marketing move by Intel.
- jhallenworld 9y agoThe Encore Multimax used the NS32032- late 80s students of WPI are familiar with with this machine because it replaced the Decsystem-20 as the student mainframe. Encore did make one product which had fairly wide use: their Annex terminal servers- sold to Xylogics.
- bpye 9y agoSome of the ideas from the Transputer can be seen in the more modern xMOS chips, though they use xC which is a C derivative adding many of the features Occam had. David May, the architect of the Transputer was also the founder of xMOS.
- xfer 9y agoi have a slicekit board, very fun to play around with hardware channels.
- nickpsecurity 9y agoTransputer lived on partly in Firewire.
- minipci1321 9y agoTransputer also lived a long and fulfilled second life in set-top box SoC chips by ST Micro, who acquired Inmos.
- kabdib 9y agoI knew about the Transputer links, but didn't know about their relationship to Firewire. That's pretty neat.
- crzwdjk 9y ago> Thank goodness the early Pentiums stopped that madness The idea of a separate co-processor that only did floating point ended with the 387. The 487SX (that was the coprocessor for the 486SX) was actually a 487DX (i.e. a full 486, with an FPU) that took over the system entirely and disabled the 486SX. But that whole scheme was still its own special kind of madness.