Talk:Central processing unit/Archive 3

Archive 1

Archive 2

Archive 3

Article concept unclear/ambiguous?

"Central Processing Unit (CPU)" is a legacy phrase which naturally does not cope well with parallel processing or the rise of multi-core processors. This leads to the question "Does CPU mean a single core or the collection of all the cores within the physical or logical boundaries of some specified information processing device?" The explanations of "how the CPU works" seems to imply the article is referring to a single core.

Instead, how about providing words to:

(1) describe the (new) general concept of "Information Processing Device" (2) explain in a different article how an information processing device "Core Processor" works individually (3) explain how multiple similar/different core processors can work together to provide the total information processing "capacity" for the "information processing device" (4) recognize that other Information Processing Architectures may also be used

Finally, can we maybe deal with earlier architectures under "History"? — Preceding unsigned comment added by Jjalexand (talk • contribs) 16:04, 5 June 2012 (UTC)

"CPU" means a single core. A microprocessor can contain multiple CPUs.

Wikipedia shouldn't be introducing new general concepts, so it shouldn't introduce a new concept of an "Information Processing Device". There's already a page that discusses the type of device you're talking about. Guy Harris (talk) 20:34, 3 March 2013 (UTC)

Specifications

Cant find them, add them. Useless article if I can't find specifications easily. —Preceding unsigned comment added by 99.227.154.176 (talk) 03:07, 27 January 2009 (UTC)

How is it useless? It's supposed to be an article about CPUs, their common features, history, etc. A list of specifications? For which CPUs? Every one that's ever been made? That would be one hell of a list. However, if you feel the material warrants inclusion, you could start off by looking the information up yourself and editing the article to provide it. 24.68.148.215 (talk) 06:30, 18 December 2009 (UTC)

And Wikipedia isn't a collection of specifications in any case. Guy Harris (talk) 20:38, 3 March 2013 (UTC)

Automate archiving?

Does anyone object to me setting up automatic archiving for this page using MizaBot? Unless otherwise agreed, I would set it to archive threads that have been inactive for 60 days.--13:18, 21 December 2009 (UTC)

 Done--Oneiros (talk) 22:54, 29 December 2009 (UTC)

Changed "process counter" to "instruction pointer"

IP is an actual CPU register, which determines where in memory the CPU is fetching/executing instructions. I have never heard the term "process counter" used to define the register IP/EIP, who exactly wrote this article? —Preceding unsigned comment added by 24.225.22.126 (talk) 18:40, 8 April 2010 (UTC)

The register in question is given different names in different instruction set architectures. It was called the "program counter" on the PDP-10 and on the PDP-11 and is called the "program counter" in SPARC. The Power Architecture doesn't seem to explicitly list a register; its documentation speaks of the "current instruction address" and the "next instruction address". x86 and Itanium call it the "instruction pointer".

The article currently calls it the program counter, and links to that page for the benefit of those who haven't heard that term. Guy Harris (talk) 21:04, 3 March 2013 (UTC)

WP:3RR - justifiable cause

A Facebook user and I were in discussion concerning CPUs when it became obvious that by definition, he was confusing CPUs with microprocessors, being that CPUs are also known as processors. He also referred to this article to support his argument. I have edited the article accordingly to clarify this ambiguity. His repeated attempts to delete this edit is the reason for the WP:3RR incident. Do review the edit as seen below and if inadequate or in error, suggest a more suitable solution:

"The central processing unit (CPU) or the processor (not to be confused with the microprocessor, a component of the CPU itself) is..." Riggerus (talk) 18:46, 22 September 2010 (UTC)

I suggest: "The central processing unit (CPU) or the "processor" (not to be confused with "microprocessor", just one component of a CPU) is..." or similar. -- FG/^T|C 20:19, 22 September 2010 (UTC)

But not all CPUs use microprocessors. It's incredibly patronizing of an encyclopedia article to admonish the reader with allegations of confusion. Leave out the parenthetical phrase. The difference between a microprocessor CPU and a CPU in general will become apparent on reading further into the article. (Editing the Wikipedia article to support your arguement has many drawbacks.) --Wtshymanski (talk) 20:23, 22 September 2010 (UTC)

Agreed Wtshymanski. The microprocessors section explains clearly that CPUs were dramatically changed by the introduction of microprocessors so, there is little "confusion". My suggestion was that at least we should not refer to A CPU as THE CPU etc. -- FG/^T|C 13:14, 23 September 2010 (UTC)

The present edit is so much better, takes contention completely out of the picture. Thanks Wtshymanski, FG --Riggerus (talk) 09:31, 24 September 2010 (UTC)

Parallelism

I would like to suggest that most of the discussion regarding Parallelism be split off to another article. While there are features of a give CPU design that better enable parallel processing, the discussion goes way beyond CPU attributes. DG12 (talk) 00:36, 22 May 2011 (UTC)

References

A few issues:

• Mixes WP:FOOTNOTE and the deprecated Footnote3
• Citations formatted with and without templates
• Explanatory notes mixed with footnotes

---— Gadget850 (Ed) ^talk 20:56, 21 August 2011 (UTC)

Why "Central"?

I mention this as a suggestion to the article's regular editors. The term "central" isn't explained. Its has its origin in a distinction between the "central" and the "peripheral" parts of a computer. for a source see Computer structures (Daniel P. Siewiorek, C. Gordon Bell, Allen Newell) google books. In 2011 we don't refer to a SATA controller as a "peripheral processing unit" but we still have "central processing unit" as a legacy term. patsw (talk) 18:04, 1 December 2011 (UTC)

It's not only a legacy term, as it is still distinguished from other processors such as a Graphics processing unit. -R. S. Shaw (talk) 06:29, 2 December 2011 (UTC)

Agreed. It was a legacy term, until graphics chips got more complex and expensive, and morphed into processing units. patsw (talk) 17:46, 2 December 2011 (UTC)

Proposed merge with Order code processor to History of general purpose CPUs

same topic, alternative name Widefox; talk 14:16, 30 April 2013 (UTC)

• oppose Whilst there's an argument that both apply to the same topic, the contexts are very different. One is the enormously widespread term for this common item, the other an obsolete term used by one manufacturer, a long time ago.

Obviously CPU is the prime article. Order code processor also meets WP:N and can stay (please be aware of WP:RECENTISM). Merging OCP into CPU would fail WP:UNDUE – although a historically notable term, it has zero relevance today and doesn't warrant even a footnote in the main article on a space-valuable high-profile computing vital article. Andy Dingley (talk) 15:21, 30 April 2013 (UTC)

They appear to be synonyms (please correct me if I'm wrong) so it's to prevent a wp:povfork / wp:dictdef, with the CPU article covers this period (60s) already, so I'm struggling to see that argument. The RECENTISM would be to dismiss old stuff like this. Anyhow, it is about this dictdef / stub with little chance of advancement. If opposition is based purely on size of CPU, instead of CPU, we could retarget merge to History of general purpose CPUs. Widefox; talk 12:30, 1 May 2013 (UTC)

I don't know if they're synonyms or not. As far as I know, I assume they are. If ICL had some distinctly different implementation of a CPU under this name (and there were such things, back in these early days), then it might have sufficient historical note to be worth adding. As a simple linguistic synonym though, it doesn't. Especially not as one that seems to have had no ongoing influence in terminology. Andy Dingley (talk) 13:45, 1 May 2013 (UTC)

I take it you have no objection to the retarget? Widefox; talk 14:28, 1 May 2013 (UTC)

As I see no content to OCP other than it being an obscure ICL term, I see no reason to mention it in any other article, including History of general purpose CPUs. The ICL concept doesn't appear to have had any influence in the history of CPUs, beyond a name that wasn't used outside that company (Sources to the contrary welcome). Per RECENTISM we should keep the article, but that's about it. Andy Dingley (talk) 14:39, 1 May 2013 (UTC)

Exactly my point - it can't be both WP:N and a dictdef - it has no future - I disagree with you there - wp:not. We agree about lack of content, exactly the reason to merge per WP:MERGE. Your main point about weight I agree with, so happy to sideline into the History of general purpose CPUs as 60s-80s mainframes aren't too obscure even if only one vendors term. I will retarget, (but keep the discussion here). Widefox; talk 16:23, 1 May 2013 (UTC)

CPU not appropriate for multi-core

"A computer can have more than one CPU; [..] those ICs are called multi-core processors." I agree. But I often see (including on WP) them just called "CPU" or CPUs (not really better, that could be separate chips). Has the terminology shifted and is CPU allowed (and should be mentioned here?)? Or should I change elsewhere? comp.arch (talk) 12:36, 30 August 2013 (UTC)