24-bit computing: Difference between revisions

Content deleted Content added
See also: It's not a *widely-used* term for a 24-bit word - it's the term Bull chose (and it's arguably just an alternative way of saying "word"); I don't think it's used elsewhere.
Usage: Added the Four-Phase IV/70, one of the few true 24-bit computers on this list.
 
(34 intermediate revisions by 22 users not shown)
Line 1:
{{Short description|Computer architectures using a 24-bit word}}
{{otheruses}}
{{redirect|24-bit}}
{{Refimprove|date=October 2009}}
{{N-bit|24|(3 octets)}}
 
Notable '''24-bit''' machines include the [[CDC 924]] – a 24-bit version of the [[CDC 1604]], [[CDC 3000|CDC lower 3000 series]], [[SDS 930]] and [[SDS 940]], the [[ICT 1900 series]], the [[Elliott Brothers (computer company)|Elliott]] 4100 series, and the Datacraft minicomputers/[[Harris CorporationComputer Systems|Harris]] H series.<ref>{{cite web | url = http://www.quadibloc.com/comp/cp0303.htm | title = Real Machines with 24-bit and 48-bit words | accessdateaccess-date = 2011-02-11 | last = Savard | first = John| archiveurlarchive-url= https://web.archive.org/web/20110107183451/http://quadibloc.com/comp/cp0303.htm| archivedatearchive-date= 7 January 2011 <!--DASHBot-->| deadurlurl-status= nolive}}</ref>
 
The term SWORD is sometimes used to describe a 24-bit data type with the S prefix referring to [[wikt:sesqui|sesqui]].{{cn|date=October 2018}}
The IBM [[System/360]], announced in 1964, was a popular computer system with 24-bit addressing and [[32-bit]] general registers and arithmetic. The early 1980s saw the first popular personal computers, including the IBM [[PC/AT]] with an Intel [[80286]] processor using 24-bit addressing and [[16-bit]] general registers and arithmetic, and the [[Apple Inc.|Apple]] [[Macintosh 128K]] with a Motorola [[68000]] processor featuring 24-bit addressing and 32-bit registers.
 
The range of unsigned integers that can be represented in 24&nbsp;bits is 0 to 16,777,215 ({{n-ary|FFFFFF|16}} in [[hexadecimal]]). The range of signed integers that can be represented in 24&nbsp;bits is −8,388,608 to 8,388,607.
The [[eZ80]] is a microprocessor and microcontroller family, with 24-bit registers and therefore 24-bit linear addressing, that is [[binary compatible]] with the [[8-bit|8]]/16-bit [[Z80]]. {{cn|date=September 2015}}
 
== Usage ==
The [[65816]] is a microprocessor and microcontroller family with 16-bit registers and 24-bit [[bank switching|bank switched]] addressing. It is binary compatible with the [[8-bit]] [[6502]].<ref>Brett Tabke. [http://www.defence-force.org/computing/oric/coding/annexe_2/ "A 6502 Programmer's Introduction to the 65816"]. Commodore World magazine, Issue #16. 1996.</ref>
The IV/70, was introduced by [[Four-Phase Systems]] in 1971. The IV/70 has an integer word size of 24 bits. [[Byte]] addressing is not supported directly but bytes are handled by instructions that pack three bytes per word. The IV/70 CPU is a 9-chip [[Integrated_circuit#LSI|LSI]] [[microprocessor]] based on three [[Four-Phase Systems AL1|AL4]] 8-bit slice register/ALUs.<ref>[http://archive.computerhistory.org/resources/text/Four-Phase_Systems/Four-Phase.SystemIV_70.1970.102646129.pdf Brochure, System IV/70, Four Phase Systems]. From the Computer History Museum. Accessed online June 11, 2010.</ref>
 
The IBM [[System/360]], announced in 1964, was a popular computer system with 24-bit addressing and [[32-bit]] general registers and arithmetic. The early 1980s saw the first popular personal computers, including the [[IBM Personal Computer AT|IBM PC/AT]] with an [[Intel 80286]] processor using 24-bit addressing and [[16-bit computing|16-bit]] general registers and arithmetic, and the [[Apple Inc.|Apple]] [[Macintosh 128K]] with a [[Motorola 68000]] processor featuring 24-bit addressing and 32-bit registers. Some late-1980s Apple computers such as the [[Macintosh SE/30]] and [[Macintosh IIx]] retained some 24-bit code in their [[ROM image|ROMs]] despite being advertised as 32-bit computers. As a result, these computers require the installation of the [[MODE32]] memory manager to address more than 8Mb of RAM.<ref>{{Cite web |last=Staff |first=L. E. M. |date=1989-01-19 |title=Mac SE/30 |url=https://lowendmac.com/1989/mac-se30/ |access-date=2025-06-18 |website=Low End Mac |language=en-US}}</ref><ref>{{Cite web |last=Staff |first=L. E. M. |date=1988-09-19 |title=Mac IIx |url=https://lowendmac.com/1988/mac-iix/ |access-date=2025-06-18 |website=Low End Mac |language=en-US}}</ref>
The range of unsigned integers that can be represented in 24&nbsp;bits is 0 to 16,777,215 ({{n-ary|FFFFFF|16}} in [[hexadecimal]]). The range of signed integers that can be represented in 24&nbsp;bits is −8,388,608 to 8,388,607.
 
The [[ARM1]], supported 24-bit memory address, as it can access 16MiB memory.
 
The [[eZ80]] is a microprocessor and microcontroller family, with 24-bit registers and 24-bit linear addressing. It is [[binary-code compatibility|binary compatible]] with the [[8-bit|8]]/16-bit [[Z80]]. Although eZ80 supports 24-bit adds, subtracts, and moves, most ALU operations are limited to 8-bit.<ref>{{cite book |title=eZ80 CPU User Manual |date=July 15, 2009 |publisher=Zilog |edition=15, April 2015 |url=http://www.zilog.com/docs/um0077.pdf |access-date=16 June 2024}}</ref>
 
The [[65816]] is a microprocessor and microcontroller family with 16-bit registers and 24-bit [[bank switching|bank switched]] addressing. It is binary compatible with the [[8-bit]] [[6502]].<ref>{{cite magazine |author=Brett Tabke. [|url=http://www.defence-force.org/computing/oric/coding/annexe_2/ "|title=A 6502 Programmer's Introduction to the 65816"]. |magazine=Commodore World magazine, Issue #|issue=16. |year=1996.}}</ref>
 
Several fixed-point [[digital signal processor]]s have a 24-bit data bus, selected as the basic word length because it gave the system a reasonable precision for the processing audio (sound). In particular, the [[Motorola 56000]] series has three parallel 24-bit data [[computer bus|buses]], one connected to each [[Memory space (computational resource)|memory space]]: program memory, data memory X, and data memory Y.<ref>[{{cite web |url=http://cache.freescale.com/files/dsp/doc/inactive/DSP56000UM.pdf |title=24-BIT. DIGITAL SIGNAL PROCESSOR. FAMILY]}}</ref>
 
[[Engineering Research Associates]] (later merged into [[UNIVAC]]) designed a series of 24-bit [[drum memory]] machines including the Atlas, its commercial version the [[UNIVAC 1101]], the [[ATHENA computer]], the [[UNIVAC 1824]] guidance computer, etc. Those designers selected a 24-bit word length because the Earth is roughly 40 million feet in diameter, and an [[intercontinental ballistic missile]] guidance computer needs to do the [[Earth-centered inertial]] navigation calculations to an accuracy of a few feet.<ref>[{{cite web |url=http://vipclubmn.org/CP24bit.html "|title=UNIVAC 24-bit computer genealogy"]}}</ref>{{failed verification|date=October 2017}}
Several fixed-point [[digital signal processor]]s have a 24-bit data bus, selected as the basic word length because it gave the system a reasonable precision for the processing audio (sound). In particular, the [[Motorola 56000]] series has three parallel 24-bit data [[computer bus|buses]], one connected to each [[Memory space (computational resource)|memory space]]: program memory, data memory X, and data memory Y.<ref>[http://cache.freescale.com/files/dsp/doc/inactive/DSP56000UM.pdf 24-BIT. DIGITAL SIGNAL PROCESSOR. FAMILY]</ref>
 
[[OpenCL]] has a built-in intrinsic for multiplication (<code>mul24()</code>) with two 24-bit integers, returning a 32-bit result. It is typically much faster than a 32-bit multiplication.<ref>{{cite web |title=integerFunctions(3) Manual Page |url=https://www.khronos.org/registry/OpenCL/sdk/2.2/docs/man/html/ctz.html |website=www.khronos.org}}</ref>
[[Engineering Research Associates]] (later merged into [[UNIVAC]]) designed a series of 24-bit [[drum memory]] machines including the Atlas, its commercial version the [[UNIVAC 1101]], the [[ATHENA computer]], the [[UNIVAC 1824]] guidance computer, etc. Those designers selected a 24-bit word length because the Earth is roughly 40 million feet in diameter, and an [[intercontinental ballistic missile]] guidance computer needs to do the [[Earth-centered inertial]] navigation calculations to an accuracy of a few feet.<ref>[http://vipclubmn.org/CP24bit.html "UNIVAC 24-bit computer genealogy"]</ref>
 
== See also ==
* [[Catena (computing)|Catena]], a term used for a 24-bit unit of data on the Bull Gamma 60 computer
 
== References ==
{{Reflist|30em}}
 
{{CPU technologies}}