mash@mips.com (John Mashey) (03/04/91)
Various people asked about the existence of 48-bit computers. I don't know of any, offhand that are being sold today. Ones sold in the past include: 1) Honeywell 800, 400, 1400-1800 2) CDC 1604, 1604A, 3[468]000 (the 3[1235]00 were 24-bitters) 3) English Electric KDF9 4) Burrounghs B5000, B5500, D825, B8500 One person asked about current Unisys (Burroughs) A-series. Correct me if I'm wrong, but I think these are derived from the Burroughs B6500/B7500 mainframes, which were the successors to the B5500 & friends, and they used 51 bits. 51 bits? Yes. The 48-bit words of the B5000 used up to the top 3 bits as flag bits, i.e., bit0 == 0 said rest of 47 bits were numeric operand. bit1 == 1 said rest of word was something else, separated by bit2. In the B6500 and later, they added 3 more bits for tags, which seemed to me to be cleaner, as it left the fairly clean number 48. The 48--bit machines look to me to be a late-50s, 1960s mainframe & mid-range phenomenon that has since died off. REFERENCE: Bell & Newell, COMPUTER STRUCTURES: READINGS AND EXAMPLES, McGraw-Hill, New York, 1971. -- -john mashey DISCLAIMER: <generic disclaimer, I speak for me only, etc> UUCP: mash@mips.com OR {ames,decwrl,prls,pyramid}!mips!mash DDD: 408-524-7015, 524-8253 or (main number) 408-720-1700 USPS: MIPS Computer Systems MS 1/05, 930 E. Arques, Sunnyvale, CA 94086
karger@osf.org (Paul A. Karger) (03/09/91)
Add to your list of 48-bit computers the Philco 2000 series. Perhaps their most famous application was as the first computers in the Cheyenne Mountain NORAD command center. They were replaced by Honeywell 6000 line machines in the early 70s as part of the WMMCCS computer buy.
haynes@felix.ucsc.edu (99700000) (03/09/91)
In article <666@spim.mips.COM> mash@mips.com (John Mashey) writes: >In the B6500 and later, they added 3 more bits for tags, which seemed to >me to be cleaner, as it left the fairly clean number 48. Well, yes and no. With the 48-bit B5500 you could write words to tape as 8 six-bit characters. With the 51-bit B6500 and later you could write words to tape as 8 six-bit characters by omitting the tags (or 6 eight-bit characters with 9-track tape) but now you've got the tag bits to do something with. So you need two different tape formats, one for pure data without the tags, and one that preserves the tags. Not that the problem is new or unique to the B6500. It goes back at least to the IBM 1401, where you could write six-bit characters on tape but if you wanted to preserve word marks you had 7-bit characters to deal with. And the G.E. (later Honeywell) machines have 36 bit words and 9-bit bytes because 9 is a divisor of 36 and 8 isn't. So again you can write tapes that hold all the bits of a word, or that hold 8-bit ASCII data with the ninth bit of each byte omitted.
avg@hq.demos.su (Vadim Antonov) (03/09/91)
In <666@spim.mips.COM> mash@mips.com (John Mashey) writes: >Various people asked about the existence of 48-bit computers. Soviet BESM-6 (mass production was started at '67 and finished about '84). Vadim Antonov DEMOS, Moscow, USSR
hrubin@pop.stat.purdue.edu (Herman Rubin) (03/10/91)
At the time of all these 48-bit computers, a byte was normally 6 bits. -- Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399 Phone: (317)494-6054 hrubin@l.cc.purdue.edu (Internet, bitnet) {purdue,pur-ee}!l.cc!hrubin(UUCP)
haynes@felix.ucsc.edu (99700000) (03/11/91)
In article <7572@mentor.cc.purdue.edu> hrubin@pop.stat.purdue.edu (Herman Rubin) writes: > >At the time of all these 48-bit computers, a byte was normally 6 bits. Also 48 is a nice number for packing all kinds of bytes and nibbles into, since it has so many divisors. I spoze that's less important now that the price of memory has gone down so much compared to 1964.
firth@sei.cmu.edu (Robert Firth) (03/11/91)
In article <13272@darkstar.ucsc.edu> haynes@felix.ucsc.edu (99700000) writes: >Also 48 is a nice number for packing all kinds of bytes and nibbles into, >since it has so many divisors. I spoze that's less important now that >the price of memory has gone down so much compared to 1964. Indeed it was. Here is one list, from the KDF9 programming manual, p 24: THE KDF9 WORD HAS 48 BITS ... IT MAY BE USED AS... Eight 6-Bit Alpha-Numeric Characters One 48-Bit Fixed-Point Number Two 24-Bit (Half length) Fixed-Point Numbers Half of a 96-Bit (Double Length) Fixed-Point Number One 48-Bit Floating-Point Number Two 24-Bit (Half length) Floating-Point Numbers Half of a 96-Bit (Double length) Floating-Point Number Three 16-Bit (Fixed point) Integers Six 8-Bit Instruction Syllables An instruction was 1, 2 or 3 syllables; an address was 15 bits. O, memory! We shall not see its like again.
renglish@hplabsz.HP.COM (Bob English) (03/12/91)
In article <666@spim.mips.COM> mash@mips.com (John Mashey) writes: >Various people asked about the existence of 48-bit computers. >I don't know of any, offhand that are being sold today. The PA-RISC architecture allows for 32-bit, 48-bit, and 64-bit implementations. I don't know of any that have actually been shipped with more than 48-bits of supported address space. Under HP-UX, processes are limited to between 30 and 32 bits of local address space (depending on how you count). Under MPE-XL, the entire address space is made available to tasks, and is used (among other things) to map file system pages. --bob-- renglish@hplabs Standard Disclaimer
fred@sma2.uucp (Fred Brooks) (03/14/91)
In article <6684@hplabsz.HP.COM> renglish@hplabsz.HP.COM (Bob English) writes: >In article <666@spim.mips.COM> mash@mips.com (John Mashey) writes: >>Various people asked about the existence of 48-bit computers. >>I don't know of any, offhand that are being sold today. > >The PA-RISC architecture allows for 32-bit, 48-bit, and 64-bit I think that the Harris H series is a 24/48-bit computer. The AN/UYK-62 is a Harris H-300. -- Defend your 2nd amendment rights. What is the New World Order? Fred Brooks Portland Oregon
ram@unislc.uucp (Ram Madduluri) (03/19/91)
fred@sma2.uucp (Fred Brooks) writes: >In article <6684@hplabsz.HP.COM> renglish@hplabsz.HP.COM (Bob English) writes: >>In article <666@spim.mips.COM> mash@mips.com (John Mashey) writes: >>>Various people asked about the existence of 48-bit computers. >>>I don't know of any, offhand that are being sold today. >> >>The PA-RISC architecture allows for 32-bit, 48-bit, and 64-bit > I think that the Harris H series is a 24/48-bit computer. The AN/UYK-62 >is a Harris H-300. Unisys A-Series machines are 48bit machines and are currently available in the market.... Ram Madduluri -- _ Ram Madduluri Unisys UNIX Networks
mash@mips.com (John Mashey) (03/19/91)
In article <1991Mar18.220827.27542@unislc.uucp> ram@unislc.uucp (Ram Madduluri) writes: >>In article <6684@hplabsz.HP.COM> renglish@hplabsz.HP.COM (Bob English) writes: >>>In article <666@spim.mips.COM> mash@mips.com (John Mashey) writes: >>>>Various people asked about the existence of 48-bit computers. >>>>I don't know of any, offhand that are being sold today. >Unisys A-Series machines are 48bit machines and are currently available >in the market.... Is that true? I thought the A-series was 51 bits or more, upward-compatible from B6500, etc. Did these go back to the 48 used by B5500? -- -john mashey DISCLAIMER: <generic disclaimer, I speak for me only, etc> UUCP: mash@mips.com OR {ames,decwrl,prls,pyramid}!mips!mash DDD: 408-524-7015, 524-8253 or (main number) 408-720-1700 USPS: MIPS Computer Systems MS 1/05, 930 E. Arques, Sunnyvale, CA 94086
cfj@iSC.intel.com (Charlie Johnson) (03/20/91)
In article <1991Mar18.220827.27542@unislc.uucp>, ram@unislc.uucp (Ram Madduluri) writes: |> fred@sma2.uucp (Fred Brooks) writes: |> |> >In article <6684@hplabsz.HP.COM> renglish@hplabsz.HP.COM (Bob English) writes: |> >>In article <666@spim.mips.COM> mash@mips.com (John Mashey) writes: |> >>>Various people asked about the existence of 48-bit computers. |> >>>I don't know of any, offhand that are being sold today. |> >> |> >>The PA-RISC architecture allows for 32-bit, 48-bit, and 64-bit |> |> > I think that the Harris H series is a 24/48-bit computer. The AN/UYK-62 |> >is a Harris H-300. |> |> Unisys A-Series machines are 48bit machines and are currently available |> in the market.... |> |> Ram Madduluri |> -- |> _ |> Ram Madduluri |> |> Unisys UNIX Networks Does the Harris H series still supported ?? I worked on both the 24 bit and 48 bit Harris machines for about a year and half. What about the Vulcan OS on the Harris machines. Now that was an experience!! -- Charles Johnson Intel Corporation, Supercomputer Systems Division 15201 NW Greenbrier Pkwy Beaverton, OR 97006 phone: (503)629-7605 email: cfj@isc.intel.com
dricejb@drilex.UUCP (Craig Jackson drilex1) (03/21/91)
In article <1126@spim.mips.COM> mash@mips.com (John Mashey) writes: >In article <1991Mar18.220827.27542@unislc.uucp> ram@unislc.uucp (Ram Madduluri) writes: > >>Unisys A-Series machines are 48bit machines and are currently available >>in the market.... > >Is that true? I thought the A-series was 51 bits or more, upward-compatible >from B6500, etc. Did these go back to the 48 used by B5500? The A-Series, like the B6500 before it, has 48-bit words visible in the applications architecture. E-mode level Beta and Gamma machines also have 4 tag bits with each word; the B6500 had 3. All of the machines had various other bits associated with each word for parity, SECDEC, etc. If one wishes, one can consider the A-Series a 52-bit architecture from an operating systems viewpoint. However, the tag bits cannot take on arbitrary values--the meanings are defined by the hardware. From the view of applications other than compilers and linkers, it is a 48-bit machine. The tag bits are not addressable. (If they were, the machine would not be secure--the A-Series (and the B6500) have no hardware distinction between kernel and user mode.) -- Craig Jackson dricejb@drilex.dri.mgh.com {bbn,axiom,redsox,atexnet,ka3ovk}!drilex!{dricej,dricejb}
koll@NECAM.tdd.sj.nec.com (Michael Goldman) (03/22/91)
BTW, the Intel 386/486 has a seldom mentioned 14-bit segment register to handle multiple 32-bit spaces (for multiple large disks for virtual memory?) so one could think of it as a 46-bit machine. The reason they chose 14 bits instead of 16 is that ..(militarily sensitive material deleted).. which clearly proves to anyone that we'll all have to scrap our base 16 calculators and switch to base 14 (easily divisible by 7, 2, & 3.5) ;) ("Segments!? Did someone say SEGMENTS!!?? SLOOWWLY I TURNED, STEP BY STEP...")