Talk:Arithmetic logic unit/Archive 1

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1

History

Few issues... First, von Neumann didn't design ENIAC, though he certainly knew of it and was influenced by its design. Mauchly, Eckert, Brainerd, and Pender are the big names usually associated with ENIAC's design and manufacture. Second, it's pretty debatable whether the ALU was originally suggested for EDVAC or not. It would be good to try to find the first usage of the term and if that cannot be found, you can probably at least mention earlier devices that were more or less ALUs with I/O devices (like the IBM series of tube multipliers, adders, etc with punch card I/O). -- uberpenguin 15:48, 3 January 2006 (UTC)

Difference between ALU and core

There is an emergence of multicore processors. So what's the difference between ALU and core?

An ALU is an execution unit, directed to its chores by the control unit; both those units are part of the CPU core. Today, a core may often contain several ALUs as well as other types of execution units. --Wernher 03:18, 30 October 2005 (UTC)


the ALU is a basic building block of a CPU or CORE.
its the same relationship as the AND gate to the Half Adder as ALU is is to CPU/CORE
the ALU is a very complex and important block that needs more info on it on this page. it is a small part of a cpu core but its prob the most complex part and critical part to understanding how a cpu works. Eadthem (talk) 03:55, 28 December 2007 (UTC)

Division/Multiplication

Is it not true that division/multiplication can be done by bitwise shifting and addition/subtraction?—Preceding unsigned comment added by 60.240.107.52 (talk • contribs)

  • I think this is the answer to your question --ZeWrestler Talk 17:34, 11 November 2006 (UTC)

hmm, can you give me a history about the arithmetic Logic Unit —Preceding unsigned comment added by Wert zieta (talk • contribs) 02:56, 20 August 2008 (UTC)

schematic typo

I think F in the schematic should be called operator not operand? A and B are the operands. See http://en.wikipedia.org/wiki/Operand SvenPB (talk) 09:56, 8 February 2009 (UTC)

Is the 74AS888 an ALU?

PeterCamilleri (talk) 18:09, 28 July 2009 (UTC)OK; I have dug up more details and this is an impressive part. Too bad it came out too late to have a real impact. That however is opinion. What is not opinion is that this is NOT just an ALU. The TI data sheet calls it an 8 Bit Processor Slice. If you compare features, this part is actually a superset of the AM2901/2903 bit slices.

--PeterCamilleri (talk) 17:22, 11 May 2009 (UTC)

Computing without an ALU

I know of two computers that did not include ALU's; instead, they used a table lookup to do adds. Theses were the IBM 1620 (Can't Add; Doesn't Even Try) and RCA 301. I don't know whether anyone ever did it for production programs, but you could do octal arithentic by changing the add tables. Shmuel (Seymour J.) Metz Username:Chatul (talk) 15:02, 22 September 2010 (UTC)

Decimal representations

I know of several decimal computers, and none of them used ten tubes per digit. If there is such a beast, then a citation should be provided and the clause even true decimal systems, with ten tubes per digit split into two clauses, e.g., decimal computers using binary representations of the digits and even true decimal systems, with ten tubes per digit. Shmuel (Seymour J.) Metz Username:Chatul (talk) 15:34, 23 November 2010 (UTC)

Practical overview

This entire section needs to be re-written. Who is 'the customer' and it is ambiguous what the costs refer to, are we talking energy, monetary costs, complexity costs? Also is it really relevant in understanding what an ALU is to emphasize that better ALUs will cost more? Isn't that obvious enough to be omitted? — Preceding unsigned comment added by 58.107.82.54 (talk) 03:55, 26 September 2013 (UTC)

Reciprocal approximate?

The ALU on the Cray-1 did not have divide circuitry, but instead implemented Reciprocal approximation and Reciprocal iteration, as explained in Cray-1 Computer System Hardware Reference Manual (PDF). Cray Research. p. 3–28. 2240004. {{cite book}}: Cite has empty unknown parameters: |author-name-separator=, |doi_brokendate=, |deadurl=, |laydate=, |subscription=, |nopp=, |trans_title=, |trans_chapter=, |laysource=, |laysummary=, |author-separator=, |lastauthoramp=, |registration=, and |chapterurl= (help); Unknown parameter |separator= ignored (help)CS1 maint: postscript (link) Does this belong in the article, or is it TMI? Shmuel (Seymour J.) Metz Username:Chatul (talk) 17:33, 3 December 2013 (UTC)

ALU not for division?

The statement that ALUs don't usually do division is silly. Historically, the vast majority of ALUs produced commercially as primary processing units have had support for integer division and a modulus operation. Check out the PDP-8 instruction set, as an example, or ANY of the references, external links, "see also"s....

Anyway, it seems like that section was cribbed directly from another source that didn't know what it was talking about. — Preceding unsigned comment added by 75.156.103.185 (talk) 16:33, 4 December 2013 (UTC)

ALU Diagram

The ALU diagram contains some letters, such as "F" which are never documented in the text. This is very confusing. — Preceding unsigned comment added by WirlWhind (talk • contribs) 19:06, 7 January 2014 (UTC)

Complex operations

The whole Complex operations section is problematic. Phrases like "Design an extraordinarily complex ALU" and "Tell the programmers" make it read like it's part of a manual or something. So I've tagged it with {Inappropriate tone} and added a few "HTML comments" here and there. - dcljr (talk) 19:33, 11 January 2011 (UTC)

I disagree. The section explains the issues in a manner that I find clear; or are you concerned with the use of the word "complex" & "simple" here? BTW, I am not a chip designer by any measure; I don't even know how to code or read assembler. -- llywrch (talk) 19:49, 24 March 2011 (iTC)
The issue is that it appears the article is giving instructions to the reader, as though it were telling you to do something. --Mman2112 (talk) 18:24, 7 March 2014 (UTC)

What is the size of ALU? F.e. in transistors or logic operations

It would be useful to specify. For example by specifiing ALU sizes for some processors and-or other schemes which use ALU.

Does anybody know how to find out this info? Pavel Senatorov, developer of programming language Ya (talk) 16:46, 13 August 2014 (UTC)

I agree that a gate count or transistor count or both would be a good addition to this article.
The "74181" article links to manufacturer's data sheets that show all the logic gates and how they are connected in that popular ALU. That article claims "75 logic gates" in one place and "63 logic gates" in another -- would you mind getting the correct number from those datasheets and updating both this ALU article and the 74181 article to the correct count?
My understanding is that other adders and ALUs often show a similar schematic diagram, or at least mention the gate count or transistor count, in their respective datasheets.
The articles "AMD Am2900", "list of 4000 series integrated circuits", "list of 7400 series integrated circuits", and "list of Intel microprocessors#3000 Family" articles mention several other adders and a few more ALUs, and link to their datasheets. --DavidCary (talk) 22:39, 28 June 2015 (UTC)
The natural sizing unit is the logic gate; relay, tube,diode and transistor counts are too dependent on the technology used. Shmuel (Seymour J.) Metz Username:Chatul (talk) 16:01, 8 July 2015 (UTC)
Unfortunately even functionally equivalent ALUs can, and often do have different circuitry. These differences can be influenced by a number of factors. There may be architectural differences for the purpose of performance (vs. functional) enhancement. Different technologies/processes have different maximum fan-outs and control gates, leading to more/fewer intermediate logic levels. And what if the ALU is implemented in an FPGA? If you want a lively debate, just ask a room full of engineers to reach consensus on the number of gates/transistors in an FPGA-based ALU. For these reasons I don't think it's practical to discuss ALU component counts here, except perhaps to explain how the component count grows (e.g., linearly, exponentially) with various ALU attributes and features. That said, it may be possible to state the component counts of historical, early ALUs that had a well-defined number of relays, tubes or discrete transistors. Lambtron (talk) 19:26, 13 July 2015 (UTC)

History of ALUs

I recently organized and expanded the history section. While doing that, I realized that it covers quite a lot of technical information that can only be understood after reading the technical sections below. It might be a bit unorthodox, but wouldn't it be logical to move the history section further down in the article, so as to make the history easier to understand? Lambtron (talk) 19:45, 17 July 2015 (UTC)

No one has objected, so I've moved History below the tech sections. Lambtron (talk) 17:31, 22 July 2015 (UTC)

74AS888 Information

{I am new to Wiki; Not sure where or how to ask so here goes.} I could not help but notice that the article includes a photo of the 74AS888 part from TI. I have been trying to find more information on this part with no success. Is anyone aware of where I might more info on this enigmatic part?

The 74AS888 chip (released circa 1982) is classified as "8-bit processor slices". It consists of an 8-bit ALU, 16x8 register file, support circuits for multi-cycle multiply/divide/normalize, data selectors, and a bit of glue logic. For more info, see TI's 1986 "LSI Logic Data Book". Lambtron (talk) 18:42, 21 July 2015 (UTC)
A copy of the 1986 "LSI Logic Data Book" that Lambtron mentioned is at http://www.datasheetarchive.com/74AS888-datasheet.html and http://bitsavers.informatik.uni-stuttgart.de/pdf/ti/_dataBooks/1986_LSI_Logic.pdf .
That includes the detailed technical data that a person would need to use that chip as part of a complete CPU, on page 2-293, as well as a "Bit-Slice Processor User's Guide" starting on p. 3-67, "Floating-Point System Design Using the 'AS888/'AS890" starting on p. 3-125, etc.
Which educational or commercial systems, if any, used one or more 74AS888 chips? --DavidCary (talk) 02:14, 21 November 2015 (UTC)

External links modified

Hello fellow Wikipedians,

I have just modified one external link on Arithmetic logic unit. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 18 January 2022).

  • If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
  • If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 02:16, 9 July 2017 (UTC)

Broken link as of 10/2019

Reference [1] has a broken link as of today 107.146.232.246 (talk) 04:36, 22 October 2019 (UTC)

Tagged as {{dead link}}.Lambtron talk 14:45, 22 October 2019 (UTC)

Why is it a unit ?

Why are such operations grouped as a single unit ?
Do they share circuits ?
Are they so simple it makes no sense to implement only parts ?
Is it just conceptual to ease representation ? Musaran (talk) 21:32, 8 April 2019 (UTC)

How would this affect the content of this article? --Mortense (talk) 18:02, 17 February 2020 (UTC)

Opposite to change

I dissagree with this change: ([[1]]). In my opinion your contribution is not right because the ALU is not necessarily an electronic circuit. The first units were carried out mechanically. Hence, it should be written that it is a digital circuit that can be implemented as mechanical, electromechanical, electronic etc. You also wrote that I deleted important information without explaining - what exactly do you mean? -79.191.112.146 (talk) 12:47, 4 September 2020 (UTC)

Ok, I ran through this edition again and added back information that might be considered unnecessarily removed. You were probably right about them. Do you think anything else should come back or be removed from this edition? -79.191.112.146 (talk) 13:01, 4 September 2020 (UTC)
It's interesting that mechanical devices can perform arithmetic functions, but that doesn't justify calling them "arithmetic logic units". To justify this, a reliable source must explicitly state that a particular mechanical device is commonly referred to as an "arithmetic logic unit". Without such a supporting citation, it's WP:OR to infer that a mechanical device is an ALU. Lambtron talk 14:28, 4 September 2020 (UTC)

Biological ALUs

I moved biological ALUs to History because it's a modern research area (vs. common practice) and to satisfy MOS:LEADREL. This could be worthy of mention in the lede if coverage is significantly expanded. Lambtron talk 15:08, 7 September 2020 (UTC)

Rationale for renaming the article to "arithmetic logic unit"

The most common terms for ALU in the computer science literature seems to be "arithmetic logic unit" and "arithmetic-logic unit"; hence the recent renaming/move from "arithmetic and logical unit". Double redirects have been removed. --Wernher 16:09, 2 Apr 2005 (UTC)

Lambtron, about the revert, I didn't see this previously at the talk page, yes, I do know often written with space, without ndash, while e.g. Britannica has https://www.britannica.com/technology/arithmetic-logic-unit with hyphen (9 out of 10 times). And techtarget with "What is an arithmetic-logic unit (ALU)? [..] In some processors, the ALU is divided into two units: an arithmetic unit (AU) and a logic unit (LU)." That's what it is two functions, arithmetic AND logic. For consistency with e.g. load–store unit, and ndash – (not - a hyphen) would be appropriate. At least if we choose to change the wording and rename the article, then do it right, not with - hyphen, because that's grammatically incorrect. comp.arch (talk) 14:38, 15 November 2021 (UTC)
Hi comp.arch, I agree there are valid reasons for calling it "arithmetic-logic", "arithmetic/logic", or "arithmetic and logic". Having said that, "arithmetic logic" is commonly used in both reliable sources and in engineering. Is there a compelling reason to change the title of the article (and all instances of the term in Wikipedia), or would it be sufficient to simply mention the alternative terms in the article? Lambtron talk 15:03, 15 November 2021 (UTC)
@Comp.arch: To my surprise I've found that the arithmetic-logic unit is more prevalent than arithmetic logic unit: [2]. I think it should be renamed per WP:COMMONNAME. Feel free to open an RFC. --AXONOV (talk) ⚑ 16:54, 1 December 2021 (UTC)
Here's a more informative graph, which shows usage from the invention of the ALU up to today. Lambtron talk 17:29, 1 December 2021 (UTC)

How do they make it so small?

I'm not the one to do this because I know nothing about it. But for people brought here by a sense of wonder, a subtopic section should be included that addresses the BIG question, how the heck to they manage to put the machinery that makes microprocessors work into such a small space? 2600:8801:B011:300:A0B9:F42B:4906:8956 (talk) 15:09, 19 February 2021 (UTC)Thanks, James.

The process of constructing such an ALU is the same process used to build all the other parts of a CPU.
A surprisingly large variety of different techniques have been used in practice to build mainframe CPUs and minicomputer CPUs.
By far the most common approach (and the approach that has produced a long series of ever-smaller ALUs, each one by far the smallest ALU ever made so far) is semiconductor device fabrication.
How can we make it clear to our readers that *this* article isn't going to go into those details, but they can still find out by going to *that* article?
--DavidCary (talk) 05:07, 27 May 2022 (UTC)
There should no be subsection on that. ALU is a mathetmical and schematic concept of either digital or analog implementation. A design which is usually a part of a bigger Integrated circuit design. The most modern implementation is laid out in the semiconductor device fabrication article. There are many ways to implement ALU. You can use even relay for that. It's matter of choice of manufacturing process. AXONOV (talk) ⚑ 08:00, 27 May 2022 (UTC)
Retrieved from "https://en.wikipedia.org/w/index.php?title=Talk:Arithmetic_logic_unit/Archive_1&oldid=1215676257"