Jump to content

Talk:Microprocessor chronology

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

To do list

[edit]

This is a promising article, but it needs some work to bring up the quality:

  • Specify clock frequency ranges: Instead of just specifying one speed, the range should be given. This is more informative and less biased. (Listing the max. for one chip and the min. for another is misleading and plain wrong!)
  • Check facts: I had a quick look and found obvious mistakes, though more checking is required. A quick look at the charts for each processor should give plenty of information that can be used for checking.
  • x86-centric: More architectures are needed to complete the list. The article is after all, 'Microprocessor Chronology', not 'x86 Chronology'
  • Transistor counts: If anyone has a better looking and more elegant way of presenting this information, please make the required changes.
  • Table appearance: If anyone can make the four tables uniform in column width, I think it will be better. Please do so if possible.

Rilak (talk) 03:48, 17 February 2008 (UTC)[reply]

  • Legend: Explanation to the column and their significance, especially the Moor's Law-column.
  • Architecture: A column noting their architecture, x86, IA64, PPC, Sparc, MIPS, ARM, etc

Henriok (talk) 09:50, 17 November 2008 (UTC)[reply]

Moore's Law

[edit]

I have concerns about presenting Moore's Law here. I think it gives (or has potential to) the impression that the microprocessors should have met the magical number of transistors in order to be with the times and that for a given year, the microprocessor that came closest to the meeting the magical number of transistors were more advanced than those further away, which cannot be further from the truth. Rilak (talk) 10:12, 17 November 2008 (UTC)[reply]

Agreed. For the earlier stuff expanding the process could interest. The fab for more recent stuff?
Architecture too like Mashey's RISC/CISC table? Something like hardwired/microcoded, accumulator/register/stack, fixed/variable length instructions, register width ....
Should microcontrollers go in? Early multiple chip processors (ie IBM RIOS/1 prior to RSC, earliest HPPA...) or keep to the simpler single chip processor criterion?
RDBrown (talk) 23:39, 26 November 2008 (UTC)[reply]
The inclusion of architectural details would be helpful, but I think only the basics should be covered here to prevent the table from becoming overwhelmingly large. Microcontrollers I think should not be included as there are far too many of them. As for multi chip processors, it depends how many chips there are and what reliable sources say about it. Rilak (talk) 05:09, 27 November 2008 (UTC)[reply]
I think the Moore's Law column should go. It's interessting and all but it doesn't fit in. -- Henriok (talk) 13:13, 27 November 2008 (UTC)[reply]
I think we should include microcontrollers in this article for now.
While there are an overwhelming number of individual microcontroller chips, I hope that listing only 1 row for each of the major architectures will be manageable.
Later, if it turns out that Rilak is right and "there are too many of them" (I think it is premature to make that judgement now), perhaps we should start a "microcontroller chronology" article. --DavidCary (talk) 17:34, 8 August 2013 (UTC)[reply]

speed bumpes and new fabrications

[edit]

I think we should concentrate of major architectural steps, like 386 -> 486 -> Pentium, and not smaller upgrades/versions like 386 -> 386SX -> 386SL. If we open up that can of worms there will be no end of every small version with minor differences to the archetypal processor. It's not all that interessting either. I will therefor prune the tree a bit. -- Henriok (talk) 13:31, 27 November 2008 (UTC)[reply]

The Transistor count article is quite similar to this. Merge? Collaborate? -- Henriok (talk) 17:04, 23 November 2009 (UTC)[reply]

Collaborate — GPUs & FPGA devices don't fit this article but make sense there.
Area and Process columns in their tables would be good and a Memory (SRAM and DRAM) table
would make sense in their article – more so if/when Area columns are populated.

RDBrown (talk) 02:25, 24 November 2009 (UTC)[reply]

A GPU chronology and FPGA chronology articles would be good. Then the Transistor count article could be eliminated ? Aero14 21:12, 28 November 2009 (UTC)[reply]

Great suggestion! I took the liberty to prep your comment as a launch point for such articles. -- Henriok (talk) 23:13, 28 November 2009 (UTC)[reply]

Expecting New Releases

[edit]

2013?

[edit]

Did nothing significant happen during 2013? -- Henriok (talk) 08:43, 3 May 2014 (UTC)[reply]

NEC microprocessors

[edit]

I've reverted several edits adding NEC's chips due to problems with the edits. The comment I left on the IP's talk page:

I've reverted your edits. Please read WP:RS on understanding what a reliable source is. As best I can tell, your Antique Chip Colletor's page is a single source. Second, please read WP:BRD. When you are reverted, please discuss the issue on the talk page before restoring your edits. Doing otherwise can get you blocked. And lastly, your edits don't even reflect your source - they claim NEC's uPD707 went in production in 1972, not 1971.

If anyone has better sources for those NEC chips, let's add them. Tarl N. (discuss) 14:46, 15 April 2017 (UTC)[reply]

What is wrong with the AntiqueTech source? Either way, the dispute seems to be over the uPD707/uPD708 chips, not the other chips, which are reliably sourced, so I'll go ahead and restore the other ones while leaving out the uPD707/uPD708 for now. 86.177.160.213 (talk) 18:51, 15 April 2017 (UTC)[reply]
As best I can tell, AntiqueTecb source is a blog (single author, no editorial board), not a publication. Tarl N. (discuss) 19:07, 15 April 2017 (UTC)[reply]
The first line on the front page of AntiqueTech is "My artwork got a mention in the Wall Street Journal today". That tells me this is a blog, not a publication. Tarl N. (discuss) 19:13, 15 April 2017 (UTC)[reply]
Either way, resolve the issues here, do not simply restore your edits. Please read WP:3RR and WP:BRD. Tarl N. (discuss) 19:09, 15 April 2017 (UTC)[reply]
I only restored my edits which have nothing to do with the issues you've raised. Why did you revert all of my edits? You have not given any explanation for why you reverted all of my other edits which have nothing to do with the uPD707/uPD708. 86.177.160.213 (talk) 20:46, 15 April 2017 (UTC)[reply]
Because your edits didn't agree with your sources, and your sources did not qualify as WP:RS. You can't use Wikipedia as a source, you can't use blogs as a source. Enough mistakes make it problematic trying to sort out what might not be in error. Tarl N. (discuss) 04:33, 16 April 2017 (UTC)[reply]
You clearly didn't look at my sources for the μPD700 and TLCS-12, which are not blogs at all... (1970年代 マイコンの開発と発展 ~集積回路, Semiconductor History Museum of Japan) (Jeffrey A. Hart & Sangbae Kim (2001), The Defense of Intellectual Property Rights in the Global Information Order, International Studies Association, Chicago). One is a museum, while the other is an academic paper. These are reliable sources. 86.177.160.213 (talk) 05:44, 16 April 2017 (UTC)[reply]
Since it's been more than 24 hours, and you still haven't given an adequate explanation for reverting my edits on the μPD700 and TLCS-12, I'll go ahead and restore them. However, since you have given an explanation regarding the uPD707/uPD708, I'll avoid restoring those. Next time you decide to start edit-warring, at least pay attention to what you're actually reverting, instead of just reverting anything and everything. The issues you've raised only concern the uPD707/uPD708, not the μPD700 and TLCS-12, which are reliably sourced. 86.177.160.213 (talk) 18:29, 16 April 2017 (UTC)[reply]

Again, antiquetech.com is not a reliable source. Please read WP:RS. At this point, if you think there are pieces of your edits which stand alone, please discuss them here before restoring them. Just edit warring is not productive, and claiming I should be picking through your edits to figure out what I should revert and what I should restore isn't productive either. Tarl N. (discuss) 10:46, 17 April 2017 (UTC)[reply]

Fujitsu microprocessor

[edit]

A new addition was added for Fujitsu, but the sources are not as reliable as might be desired. Below is the comment I left on the editor's talk page:

cpu-museum.com does not seem to qualify as WP:RS. It seems to be registered to a single individual (Juri Prot) at a residential address. It provides no sources for its claims - indeed, it's probably just copied from the Fujitsu website. The Fujitsu website would be considered reliable if it had more detail than just saying "Developed the world's first 16-bit microcomputer on a single chip". I can't tell whether the date is when someone wrote a paper on it, built the chip, released it for public sales, or is simply self-puffery by a Fujitsu employee. Making claims about "world's first" usually needs more proof than a one-liner on the bottom of an obscure web page.

I don't necessarily disbelieve the claim, but I think we need better sources than are given (and going to the Wikipedia page on the chip, one of the sources now redirects to some kind of japanese game website). Tarl N. (discuss) 00:05, 28 April 2017 (UTC)[reply]

I have found a more reliable source:
Indeed, that's more reliable. However, it only says "developed in 1975", does not say it was released in that year (the common standard we hold the other chips to). It also says "around the time of the 1st 16-bit microprocessor" rather than "the 1st 16-bit microprocessor". Since the kit was released in 1977, we can suspect that they did not have production versions of the chip sitting around for two years. Please find a reliable source about the chip itself, rather than passing mentions in some other product. Tarl N. (discuss) 01:22, 28 April 2017 (UTC)[reply]
[edit]

Hello fellow Wikipedians,

I have just modified one external link on Microprocessor chronology. 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: 5 June 2024).

  • 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) 18:18, 10 June 2017 (UTC)[reply]

[edit]

Hello fellow Wikipedians,

I have just modified 3 external links on Microprocessor chronology. 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: 5 June 2024).

  • 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) 09:49, 29 January 2018 (UTC)[reply]

HP FOCUS, 1982

[edit]

The entry for the HP FOCUS in 1982 seems off. Does anyone have a more reliable date? The Wikipedia article on the chip says "Launched in 1982", which usually means when a project is started, not when production chips were available. From the technology, clock rate and number of transistors, this chip looks more like 1985-1985 than 1982. Tarl N. (discuss) 04:56, 27 November 2018 (UTC)[reply]

Answering my own question, HP Journal from Aug 1983, available under http://www.hpmuseum.net/exhibit.php?catdoc=12 (it's got a funny captcha, so I can't reference it directly) states 18 MHz clock, 450,000 transistors and 2.5 µm pitch. It describes this as 1.5 µm lines spaced 1.0 µm apart - I don't know if that's called 1.5µm technology or 2.5 µm. But it specifies in late 1983, "New HP 9000" computer, which makes me wonder if 1982 is off by a year. Either way, it was about 4-5 years ahead of the rest of the market. Tarl N. (discuss) 05:13, 27 November 2018 (UTC)[reply]
If you have full access look at: Beyers, J.W.; Dohse, L.J.; Fucetola, J.P.; Kochis, R.L.; Lob, C.G.; Taylor, G.L.; Zeller, E.R. (October 1981). "A 32-bit VLSI CPU chip". IEEE Journal of Solid-State Circuits. 16 (5): 537–542. doi:10.1109/JSSC.1981.1051634. which I suspect is the Conference proceedings referenced in the Aug 83 HP Journal. RDBrown (talk) 08:37, 27 November 2018 (UTC)[reply]

RISC-V

[edit]

On Oct, 21 2022 I added a reference to RISC-V architecture in section 2010s, but was reverted: https://en.wikipedia.org/w/index.php?title=Microprocessor_chronology&oldid=prev&diff=1156493555 by: @Tarl N.

I think that a reference to RISC-V architecture cannot be missing in this article regarding the history of microprocessors.

Please explain your idea about what can be in the article and what cannot Efa (talk) 11:50, 10 July 2023 (UTC)[reply]

Your entry had no specifics. As such, it had no useful information. This is not a list of general architectures, it is a list of specific microprocessor chips. Tarl N. (discuss) 20:50, 10 July 2023 (UTC)[reply]
@Efa: By the way, if you wish to notify a user, use the {{ping}} template in the same paragraph as your signature. The notification occurs during the process of expanding the tildes to an explicit signature, and does not apply to anything done outside that last paragraph. Regards, Tarl N. (discuss) 21:00, 10 July 2023 (UTC)[reply]
I.e., individual microprocessors listed in RISC-V § Implementations might belong in the tables in this article, but RISC-V doesn't belong as an entry by itself, as it's an instruction set, not a microprocessor, just as individual ARM, PA-RISC, SPARC, x86, ESA/390 and z/Architecture, MIPS, etc. microprocessors might belong in those tables, but those instruction sets don't belong as entries by themselves.
For example, the very first RISC-V implementation would probably belong, as would the first 64-bit implementation, and other notable implementations. Listing every RISC-V implementation would be too much, just as listing every ARM core would be too much. Guy Harris (talk) 21:53, 10 July 2023 (UTC)[reply]

I couldn't find data on older (non-prototype) implementations than SiFive's E31, a 32-bit single core with the RV32IMAC set and privileged instructions. I couldn't find the E31's transistor number. unsigned entry from Efa (talkcontribs) 09:45, 11 July 2023‎ (UTC)[reply]

The E-31 entry looks good. It's not unusual for the number of transistors to not be available. Tarl N. (discuss) 15:44, 11 July 2023 (UTC)[reply]

Added U54-MC, is one of the first 64 bit RISC-V core--Efa (talk) 12:25, 17 July 2023 (UTC)[reply]

Added U74-MC, is the first high performance 64 bit RISC-V core with MMU/Linux capable. It is the core used in the well known SoC JH7110, mounted on VisionFive2. I'm not sure if add S76 (no MMU) and E76 (32 bit microcontroller)--Efa (talk) 16:39, 31 August 2023 (UTC)[reply]

tables fields

[edit]

In my opinion we should add other columns to all tables such as:

  • the reference ISA
  • Word size (bits). Recent tables miss this field
  • cache size and type
  • performances MIPS/MFLOPS
  • efficiency MHz/W or MIPS/W

Efa (talk) 12:38, 12 July 2023 (UTC)[reply]

Dubious pre-4004 mps

[edit]

I see that two cpus were added before the 4004. The TMS with no details and no reference, and the AL1 with the wrong date (engineering systems didn’t exist until 1970, let alone when they shipped), with a blog for a citation. I figure I’d allow some discussion here before removing them. Tarl N. (discuss) 03:51, 19 January 2024 (UTC)[reply]