Jump to content

Talk:Apple silicon/Archive 1

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1Archive 2

Misc

Article is a rehash of Apple A4 and Apple A5 article. Straight copy of it. Don't know if currently adds any further value to the topic. --Visik (Chinwag Podium) 02:09, 7 July 2011 (UTC)

I think it's excellent. -- Henriok (talk) 08:44, 23 August 2011 (UTC)

The table

I, for one, think that the un-spanned table looks better and conveys information better. And that's the ultimate point. -- Henriok (talk) 12:18, 10 March 2012 (UTC)

Completing the table?

I'm aching to add previous SoCs, used in previous Apple devices. These are strictly designed and manufactured by Samsung though. But it would be really great to gather all the CPU's data from all devices from the original iPhone and forward. I'm talking about these devices: S5L8900 (iPhone), S5L8720 (iPod touch 2), S5L8920 (iPhone 3gs) and S5L8922 (iPod touch 3). -- Henriok (talk) 19:55, 19 March 2012 (UTC)

Needs citations

The section "List of Apple SoCs" is full of unsourced information. It would be better if people editing this article and the information in the tables used reliable sources to back up information. Jørgen88 (talk) 11:03, 13 September 2012 (UTC)

Rename article to List of Apple system on a chips?

I couldn't find the exact terminology "Apple Ax" on any technical documentation from Apple, instead these chips always referred to as system on a chips. For consistency and easier discoverability on search we should follow other wikipedia article titles on lists.

I propose we follow the lead on the other list of SoCs in wikipedia which has the title "List of Samsung system on a chips," and of multiple other Apple articles which start with "List of Apple...". Therefore this article would be named "List of Apple system on a chips." --112.203.62.55 (talk) 01:29, 13 September 2012 (UTC)

Good idea. I agree. --Biker Biker (talk) 06:28, 13 September 2012 (UTC)
Support - By the looks of it, this "article" appears to be an aggregated list with some detail on the SoCs and as most list do start with the title "List...", this one should be renamed accordingly. YuMaNuMa Contrib 06:32, 13 September 2012 (UTC)
Support -- Henriok (talk) 06:48, 13 September 2012 (UTC)
Support Is there a way to move the article to that title? It seems to be locked. --202.92.128.224 (talk) 03:40, 16 September 2012 (UTC)

Image placement

YuMaNuMa made a very good call by placing the chip images into the table. This fixed an ugly image stack on the right side of the page which meant that the images didn't clearly line up with the content to which they were associated. By putting the image in the table that relationship is firmly established. --Biker Biker (talk) 09:13, 16 September 2012 (UTC)

Twice as fast

Rating performance with phrases such as "twice as fast" is very weak and probably unjustifiable. Twice as fast as what? Who has made these claims? The A6 section is particularly badly worded. Such claims require to, at the very least, be attributed to a source and preferably to a reliable independent testing body. — Preceding unsigned comment added by 90.218.45.193 (talk) 12:21, 24 September 2012 (UTC)

The claim is sourced and what it compares to is stated in the article and the source. I don't see the problem here. Just check out the reference. -- Henriok (talk) 12:48, 24 September 2012 (UTC)

Requested move

The following discussion is an archived discussion of the proposal. Please do not modify it. Subsequent comments should be made in a new section on the talk page. No further edits should be made to this section.

The result of the proposal was moved. --BDD (talk) 18:41, 2 October 2012 (UTC) (non-admin closure)

Apple Ax → List of Apple system on a chips – Paraphrased from previous discussion: the term "Apple Ax" is not present in any official Apple documentation. The A4, A5, A5X and A6 chips are instead referred to as system on a chips. It was then proposed that we should follow the lead on the other list of SoCs in Wikipedia which was called "List of Samsung system on a chips," and of multiple other Apple articles which start with "List of Apple...". I agree with the previous anon and registered posters on those points and so I request the article be moved to "List of Apple system on a chips". 202.92.130.3 (talk) 04:50, 23 September 2012 (UTC)

Survey

Feel free to state your position on the renaming proposal by beginning a new line in this section with *'''Support''' or *'''Oppose''', then sign your comment with ~~~~. Since polling is not a substitute for discussion, please explain your reasons, taking into account Wikipedia's policy on article titles.

Discussion

Any additional comments:
I don't know what I'm talking about. The pluralization doesn't make sense to me, but using the Samsung article as an example, it seems cromulent. --BDD (talk) 18:41, 2 October 2012 (UTC)
The above discussion is preserved as an archive of the proposal. Please do not modify it. Subsequent comments should be made in a new section on this talk page. No further edits should be made to this section.

Naming convention and markings

Each of these devices use two names, depending on purpose and situation. One is the Samsung name and the other is the Apple name. We are using the Apple name in this article since it's the one that's printed on the package. In software though, these are named using the Samsung name.

The Samsung name is this: S5L8920. The S5L was Samsung's designation for ARM based application processors family. The 8920 is the name of the specific model belonging to a series of processors.

Apple started to use 8700 series in iPods, and the 8900 series of processors were a more powerful series of application processors from Samsung that Apple used for the original iPhone. As far as I know, these were stock Samsung application processors that are by other vendors. The ones used by Apple is stamped with Apple's logo, but there are pure Samsung parts out there without it. Since Samsung is a technology manufacturer the OEM's (Apple in this case) are allowed to brand and rename their parts how they like. What may differ between OEM's are amount of RAM and since the SoC and RAM can be mounted in the same package, its reasonable to make some differentiation of the finished product.

The Apple name is like this: APL0298. The APL stands for "Apple" of course and the 0298 is the Samsung numeric designation for the unbranded chip backwards. It's the same part. At least the SoC part is the same. On Apple's packaging there's also markings describing the included RAM chips that may or not be manufactured by Samsung.

There are either two lines of text on the packaging or four. The chips with two lines describe the part model number or the SoC and the manufacturing date (YYMM), and serial numbers of different sorts, perhaps manufacturing process, factory line, and such. This follows Samsung's naming convention. The ones with four lines of text also describe the RAM that's been packaged with the SoC. What type, how much, their manufacturing date, and serial numbers. These markings follow whatever naming convention the RAM manufacturer is using. It's therefor easy to tell if a package is including RAM or not, and what manufacturer of RAM is being used. The A5X and A6X is not, but all other parts (at this date) are. On the parts with four lines, the upper and lower describe the SoC part, the left and the right ones describe the RAM.

I can't find any references to any other vendor that uses the S5L8920 besides Apple, and I guess that Apple was by far the largest customer and might have paid Samsung a lot of money to keep it exclusive to them. And from that part on, Apple seems to have pretty much owned this family of processors. It'll be interesting to see what the designation of the processors become once they are completely out of Samsung's grasp. Even the A6X is still manufactured by Samsung, but that's as far as Samsung's commitment go. The A7 might use a completely different naming convention. We'll se.

With the introduction of the S5L8920/APL0298 in this list, I ust add the S5L8922/APL2298 that was used in the iPod touch 3G. -- Henriok (talk) 14:30, 10 November 2012 (UTC)

Actually, I believe the APL designation is a Samsung designation for "application" not "Apple" as they classify their ARM cores as targeting an "application" profile. This is likely a reference to how the Cortex cores are also divided between Cortex-A "application" profile, Cortex-R "real-time" profile and Cortex-M "microcontroller" profile. 50.53.15.59 (talk) 18:23, 27 January 2013 (UTC)
Good point. Are there any other Samsung SoCs with the APLXXXX designation that's not used in Apple's products? -- Henriok (talk) 19:55, 27 January 2013 (UTC)

iPod touch 3rd Gen 8 GB

It seems at least some people want to deny the existence of an iPod touch 3rd Gen 8 GB and keep removing it from this article. Although the iPod touch 3rd Gen 8 GB is extremely similar to the iPod touch 2nd Gen 8 GB they are indeed separate products and even have at least one significant difference (besides identifying themselves differently as iPod3,1 vs. iPod2,1). This difference is in the their wireless chip. The iPod touch 3rd Gen 8 GB has a Broadcom BCM4329 whereas the iPod touch 2nd Gen 8 GB has a Broadcom BCM4325. This is the reason for the difference in Apple's quoted Bluetooth capability of 2.1 vs. 2.0. The Broadcom BCM4329 also supports 802.11n but evidently Apple never enabled this with their firmware.

In light of this, please do not delete references to "iPod Touch (3rd gen.; only 8GB)" regardless how similar it might be to iPod touch 2nd Gen 8 GB. They are different products. 50.53.15.59 (talk) 17:29, 27 January 2013 (UTC)

People (myself included) are denying the existence of an 8 GB iPod touch 3G because it's not a 3rd generation iPod touch; it's a minor revision of the iPod touch 2G. The 8 GB model still identifies itself as an "iPod2,1," which is the identifier for all iPod touch 2G models. It uses the same IPSW as the other iPod touch 2G models (consequently, its last firmware was iOS 4.2.1, not 5.1.1 like the iPod touch 3G), and has the same exact specifications as one. (Since it uses the S5L8720, it doesn't have Voice Control or support multitasking out of the box in iOS 4.0, whereas the iPod touch 3G does since it uses the S5L8922.) Its bootrom version is also more in-line with an iPod touch 2G's than a 3G's. The iPod touch 2G has a bootrom version of "iBoot-240.4" and the iPod touch 3G has a bootrom version of "iBoot-359.5;" the 8 GB model's bootrom version is "iBoot-240.5.1." As for the 8 GB model using the BCM4329 wireless chipset, I would like to see a teardown that proves this. It is to my understanding that the 8 GB model still used a BCM4325 chipset, and that wouldn't be so farfetched; the iPhone 3GS used the BCM4325 for its life. I'm actually not sure if Apple stated the supported Bluetooth version for the iPod touch 2G (I think they said 2.1, but I cannot find an archived version of their site that proves this), but Broadcom does state that the BCM4325 does indeed support Bluetooth 2.1.[1] --Dialexio 19:13, 27 January 2013 (UTC)
Well this source supports my assertion but I have seen some that appear to support your position as well:
50.53.15.59 (talk) 02:24, 28 January 2013 (UTC)

Undo name change

This is a page about Apple's SoCs, it's not about anything named "Apple" so you should probably not name the article in this manner: "Apple (system on chip)" -- Henriok (talk) 08:10, 1 July 2013 (UTC)

I am not sure what the name was previously (it seems to have had quite a few names in the past), however, I agree. The article is about Apple SoCs so it is incorrect to put "system on chip" in parentheses as if it was about something else Apple but could be disambiguated by such. Apple SoCs are a subcategory and/or part of some products produced by Apple so the article logically should be about SoCs with Apple as an adjective. 50.53.15.59 (talk) 23:52, 13 July 2013 (UTC)

SoC table list - two tables one for "Apple-designed" and one for "rebranded" SoC

An editor recently made an edit separating the SoC table into two tables one for "Apple-designed" and one for "rebranded" SoCs. I think making that division is a bad idea for two reasons. For one, I'm not sure if the chips in the "List of Apple rebranded SoCs" could legitimately be referred to as "rebranded", and the edit did not provide any references for how the decision was made. Second, it's not clear how this improves the article; or even what was the intended goal of the edit. I've reverted the edit, but wanted to open this up to general discussion and see what others thought (in case I'm off-base). Personally, I think the table should be left as a single table, but I would not be averse to adding a new section at the top, before the A4 section, clarifying the role and design origins of the earlier chips. —RP88 (talk) 13:07, 19 September 2013 (UTC)

Also, when doing research in this topic, several (most?) of the pre-A4 devices didn't show up in devices from any other vendor than Apple, and isn't listed for purchase from Samsung. They can therefor be Apple designs too, even if Apple didn't communicate that in their marketing. "Design" can be interpreted as a wish list or work order from Apple to Samsung, so that they got a bespoke part that only they could use. Samsung probably did most of the low level circuitry design, but Apple could certainly have made the high level design.. We just don't know. We do know that they are all branded with Apple's logo, but that doesn't say much since OEM parts often have an area on the packaging for their customer's logo even if the suff underneath is generic. Apple's IC department wasn't completely without talent before they bought Intrinsity and P.A. Semi, so I'd be very surprised if they didn't influence the earlier designs more than just a s casual buyer. It'd rather keep the table as one, but we could perhaps indicate which models are confirmed Apple designs. And even then.. The A4 is less of an Apple design than the A6… -- Henriok (talk) 13:41, 19 September 2013 (UTC)

Reading old talk page entries and looking at old page revisions, it's apparent that this article was originally just about the Apple Ax SoCs before they each had their own article. So this article retains some awkwardness from the earlier transition to separate articles and the addition of the earlier generations of SoCs that were designed by Samsung. So I can sort of see the source of the confusion that lead to the division of the table. However, i think a better solution would be improving the layout and clarity of this overview article. I'll give it some thought. —RP88 (talk) 13:27, 19 September 2013 (UTC)

I started it, and I just began with the A4 since the information was more readily available. It was my intention to include older devices too, and I still have devices to add. Why now include the processors that went into the AirPort stations, AppleTV gen1 and Newtons? -- Henriok (talk) 13:41, 19 September 2013 (UTC) 13:41, 19 September 2013 (UTC)
Thanks for a bit of history about the article. With regards to the other devices, the first-generation AppleTV didn't use a SoC, it used an off-the-shelf Intel Pentium M. The early AirPort base stations used off-the-shelf PowerPC and MIPS CPUs, while the later ones use off-the-shelf Broadcom router SoCs. The Newtons used ARM CPUs in combination with Apple-designed system ASICs, so they didn't use true SoCs. So, believe it or not, I think the article already mentions all the chips that can legitimately be called Apple SoCs. —RP88 (talk) 14:28, 19 September 2013 (UTC)

I was kind of expecting a revert, I think A4 is confimed (first?) "Apple-designed", I had actually changed in some other article from Samsung to Apple (that uses another product code than in the pictures here) and then reverted myself, since I can't back up that they are Apple designed. I just have no source. I guess it's Samsung - still only know that they are the manufacturer. Don't know that they are the designer, what does that mean? It's non-custom ARM and non-custom (I guess) PowerVR. Who else designs? Maybe it's not non-trivial to put togeter.. Newton used however just an ARM 610 CPU , not "SoC"(?). comp.arch (talk) 13:58, 19 September 2013 (UTC)

Yes, the A4 is first SoC where Apple explicitly claims that it is "designed by Apple". The chips prior to the A4 are at least "Apple specified", but probably to what degree they are "Apple designed" is unclear. They've not off-the-shelf parts; they've never shown up in other mobile devices (although certainly some of the Samsung functional units are in contemporary off-the-shelf Samsung SoCs). —RP88 (talk) 14:28, 19 September 2013 (UTC)

OK, I've made an attempt to adjust the structure of the article to accommodate sections for the SoCs prior to the "A" series. I did some minor rewriting and relocation of intros to accommodate the new sections. Actual content for the early SoCs is still mostly just stubs that are thinly sourced. I'll try to fill them out later. However, if you fell this is ugly, worse than the old layout, or worse than the split table, please feel free to revert. —RP88 (talk) 16:59, 19 September 2013 (UTC)

GFLOPS of A6X and A7 (APL0698)

I've noticed that there is an editor who, for some reason, disagrees with values for the GFLOPS of the A6X and A7 (APL0698) (see [2], [3], [4], [5]). Editor, so far you haven't explained your objection to these values. You've made your change and have been reverted several times by a number of different editors. Rather than just continuing this pattern, I was hoping you'd come here an explain why you feel the value from the references are wrong. I think you will find that offering a persuasive case here on the talk page will be a more successful strategy. Hopefully we can resolve the issue. —RP88 (talk) 07:58, 15 November 2013 (UTC)

Looking at the edits and going back to the associated references, I think I may know what is confusing this editor. AnandTech often compares different generations of GPU to each other by using a reference clock (usually 300 MHz) in order to show the generation over generation performance-per-clock improvement. Not all of these "Mobile SoC GPU Comparison" tables make this clear. Sometime AnandTech makes this distinction explicit (see, for example, the GPU chart in the iPad Air review which lists GFLOPS at both the 300 GHz reference and also an "as shipping" value), but sometimes this is not made obvious (such as in the iPhone 5s review). —RP88 (talk) 08:28, 15 November 2013 (UTC)

Rename article

Exynos links here with "A by Apple". Apple A7 and related pages use "Apple system on chips". Above I actually see that it was moved to "List of Apple system on a chips". SoC is "System on (a) chip". The system is not on many chips, however what is the plural of SoC? SoCs? Should this page be renamed to "Apple SOCs"? Or thinking outside the box, just "Apple chips"? Should it include "List of" in front? Is there any way to have a plural title without implying that that chip in SoC is not singular? comp.arch (talk) 17:14, 30 October 2013 (UTC)

All of those issues have been fought over before. The name of this article has been rather contentious in the past, it's been renamed many times:
Date Change Reason given
4 October 2011 Apple Ax (System on Chip) → Apple Ax (link) unnecessary disambiguation
2 October 2012 Apple Ax → List of Apple system on a chips (link) RM; see talk page
10 October 2012 List of Apple system on a chips → Apple system on chips (link) System on a chips is not correct English
24 October 2012 Apple system on chips → List of Apple system on chips (link) Moved the article as per what appears to be a unanimous consensus and fixed the grammatical issue in title
25 October 2012 List of Apple system on chips → Apple system on chips (link) This article is not only a list.
4 December 2012 Apple system on chips → Apple System on a Chip (link) Apple System on a Chip is the technical term
2 January 2013 Apple System on a Chip → Apple System on Chips (link) This page is about a series of system on chips from Apple, not a specific, system on a chip.
1 July 2013 Apple System on Chips → Apple (system on chip) (link) naming standardization
So part of the reason you may not be getting a lot replies is that editors don't want to get into the naming issue again. I personally prefer "Apple system on a chip", singular and with no parenthesis (like the IBM Selectric typewriter article), including the "a" and no SOC caps (like the system on a chip article). However I doubt those reasons are a good enough reason to change the name, particularly given how many different names it's had in the past. —RP88 (talk) 13:04, 19 November 2013 (UTC)
The current title doesn't make sense. The article's subject is a processor line designed by Apple, not a processor line called "Apple".
Your suggested title, Apple system on a chip, seems better than any of the above. Another possibility is Apple A-series. (It appears that reliable sources usually hyphenate.) —David Levy 10:27, 3 December 2013 (UTC)
I think the name "Apple A-series" is, unfortunately, somewhat misleading as a name since Apple's early SOCs from 2007 to 2009 are not officially part of the A-series. —RP88 (talk) 23:25, 3 December 2013 (UTC)
That's a good point. Per WP:COMMONNAME, the relevant question is whether reliable sources typically refer to those processors via the "A-series" terminology (not whether Apple does so officially). I don't know the answer.
Either way, Apple system on a chip is better than any title that's been used so far, and all of the earlier titles were better than the current one, which doesn't make sense (a concern previously noted by two other users) and evidently was instituted without discussion.
If there are no objections, let's switch to Apple system on a chip (at least for the time being). We can consider alternatives later, but we should eliminate the current illogical/misleading title now. —David Levy 10:46, 4 December 2013 (UTC)
I approve this. -- Henriok (talk) 10:50, 4 December 2013 (UTC)
I say we should use Apple systems on chip or just Apple SoCs (would we say Apple/Intel CPUs or spell CPU out?). WP:PLURAL seems to apply here (note for systems, not for chip, chips would be wrong/misleading). If you want to add "List of" in front, I'm ok with that. "This article is not only a list." (25 October 2012) Why exactly not? Seems to be to me.. See [6] [7] [8] [9]. comp.arch (talk) 14:13, 5 December 2013 (UTC)
Interestingly while googling the now obvious plural form I found "A multiprocessor systems-on-chip (MPSoC) is a system-on-chip (SoC) that contains multiple instruction-set processors (CPUs)." [10]. The ARM (the CPU) doesn't have multiple instuction sets. I didn't look into this further, that is if a GPU counts as a separate instuction set (is it general enough, in this case the PowerVR) and MPSoC applies for each chip. Try to come up with the plural spelt out.. :) comp.arch (talk) 14:32, 5 December 2013 (UTC)
How about Apple SoC line? That seems to reflect common usage while retaining accuracy and avoiding any grammatical awkwardness.
Keep in mind that we can discuss alternatives later. Right now, our top priority should be ridding ourselves of the current title as quickly as possible. —David Levy 18:26, 7 December 2013 (UTC)
I don't like "Apple SoC line" as a name at all. A quick Google search seems to show that it doesn't appear to be used by any reliable source, a strong indication that it does not reflect common usage. —RP88 (talk) 02:39, 10 December 2013 (UTC)
It's a collective description (analogous to our "List of..." titles), not a product designation in and of itself. Come to think of it, simply Apple SoC might work (and could even be interpreted as plural).
But really, let's just eliminate the current title and settle on a permanent one later. Apple system on a chip is still my favorite option, and it appears to have the most support at the moment. —David Levy 04:13, 10 December 2013 (UTC)

Why exactly isn't this a list article? And should it then start with "List of" like "List of Samsung System on a Chip"? In any case, I think we should use List of Apple systems on chip or just List of Apple SoCs in plural similar to List of colors, not List of Color. comp.arch (talk) 15:00, 16 December 2013 (UTC)

Unlike the Samsung article, this one contains a great deal of information beyond a list. —David Levy 03:12, 20 December 2013 (UTC)

SoCs in iPod Nano, Shuffle, and Classic

Someone recently added some iPods (Nano, Shuffle, and Classic) and the Apple TV (1st gen) to the "Utilizing devices" column of the SoC table. They didn't include any references with their additions, and some seem unlikely, such as the claim that the iPod Shuffle (4th gen.) uses an A4 or that the iPod Nano (7th gen.) uses an A5. The addition of the 1st gen Apple TV was certainly wrong (it used a Pentium M). I've reverted these changes. For reference, here is what iFixit teardowns show for these chips:

iPod Nano (3rd gen.): Apple 337S3473 (Samsung S5L8702)
iPod Nano (4th gen.): Apple 339S0049 (Samsung S5L8720)
iPod Nano (5th gen.): Apple 339S0081 (Samsung S5L8730)
iPod Nano (6th gen.): Apple 339S0104 (Samsung S5L8723) (i.e. Samsung APL3278A01)
iPod Nano (7th gen.): Apple 339S0193
iPod Shuffle (3rd gen.): Apple 339S0082
iPod Shuffle (4th gen.): Apple 339S0128

If someone would like to add these devices to the "Utilizing devices" column, please identify the source that makes the claim, as it seems to be contradicted by iFixit. —RP88 (talk) 23:48, 7 June 2014 (UTC)

Speculation about A1, A2, A3

Someone added the names "Apple A1", "Apple A2" and "Apple A3" to this article. They didn't include a rationale for adding these names, and I can't find a trustworthy source identifying these chips with these names. I reverted this change, but I am happy to discuss this if someone feels strongly about including these names. —RP88 (talk) 23:48, 7 June 2014 (UTC)

Rumors about A8

Someone added unsourced speculation about the Apple A8 to this article. I've revered this change, as we don't normally include unsourced rumors. If you disagree with my revert, I'd be happy to discuss it with you here on this discussion page. —RP88 (talk) 23:48, 7 June 2014 (UTC)

The Inheritance-chart

I think the chart is wrong since I think it should depict the technological family tree, not how they are used in products. The A6X is a sister to A6, not a child of A5X. There is much more kinship between A6 and A6X than between either and A5X and A7, so why their line is dotted is har to understand, and if their line is dotted, so should the A5/A5X. In the A6 article, it's stated that the A6X is a variant, but not the other way around, curiously. In the A6X article, it says that it is the successor to A5X, which I now believe is confusing on the border of wrong. We should probably just smack the A5/A5X articles together to one, and the A6/A6X into one too. -- Henriok (talk) 07:53, 11 September 2014 (UTC)

Microarchitecture vs Instruction Set

Please be aware the ARMv7 and ARMv8-A are microarchitecture cores and NOT an instruction set. Each ARM microarchitecture core implements one or more instruction set, such as ARM, Thumb, Thumb-2, ... and vary from core to core as ARM architecture and ARM cores have evolved over the years. Please be careful how you describe ARMv7, ARMv8-A, and other ARM core to prevent future confusion. • Sbmeirow • Talk • 22:59, 9 February 2015 (UTC)

No, they're not. They're architectures, which specify multiple instruction sets. Multiple different microarchitectures can and do implement the same architecture. Guy Harris (talk) 01:28, 10 February 2015 (UTC)

A9 Images

Henriok, on an unrelated note, you've done a bunch of drawings for the previous models of the Ax series. Do you plan on doing a pair of images for the A9 (APL0898 & APL1022)? I assume you already have appropriate textures and stencils. However, if you are too busy, let me know, and I'll attempt to create some appropriate images (but I don't want to step on your toes if you're already working on some images). —RP88 (talk) 16:04, 2 October 2015 (UTC)

Yes! They are coming this weekend (I've had a lot to do.. and a conference). There are good photos of the A9 (The Samsung version), and I'd put something nice and provisional for the A9x from Apples iPad Pro video. -- Henriok (talk) 16:16, 2 October 2015 (UTC)
Henriok, iFixit's photo of the Samsung A9 (APL0898) is quite good, but half the lettering on their photo of the TSMC A9 (APL1022) is illegible. The most readable version of the TSMC APL1022 that I know of is this photo from Techinsights. —RP88 (talk) 18:05, 2 October 2015 (UTC)
Excellent! Thanks! I didn't have the pic from Techinsights. -- Henriok (talk) 20:56, 2 October 2015 (UTC)

Article Name still relevant?

Now that the Apple Watch chip is included, should the article name be renamed or should a new article be created? Since the S1 is a SiP (System in Package) instead of a (System on Chip), changes should be made to reflect this. — AYTK (talk) 15:55, 25 April 2015 (UTC)

This is not a new discussion by any means, but recently I came up with this: As per the excellent renaming of the article Apple motion coprocessors, wouldn't "Apple application processors" be a acceptable name for this article? It should encompass the A and S families, as well as not sounding very strange (as it currently does). -- Henriok (talk) 14:39, 29 September 2015 (UTC)
That's actually a good idea. Keeps it consistent with the Apple motion coprocessors article as well as keeping it relevant for both SiP and SoCs. — AYTK talk. 18:02, 29 September 2015 (UTC)
Allegedly, the Apple A9 includes the M9 on the chip; that would arguably make it more than just an application processor. Guy Harris (talk) 18:31, 29 September 2015 (UTC)
I found this definition of an APU: "A mobile application processor is a system on a chip (SoC) designed to support applications running in a mobile operating system environment." That would certainly fit all devices on this list, even if the SoC in the future would integrate baseband, and other coprocessors. -- �Henriok (talk) 14:39, 30 September 2015 (UTC)
The suggested name as well as the current name have problems in varying degrees, so there doesn't appear to an obvious best choice for the name like there was for Apple motion coprocessors. The "System on a chip" is mostly accurate in varying degrees for all of the chips. All of the Ax/S1 chips contain a "system on a chip", but many of these are actually "package on package" (PoP) configurations which technically contain two chips (usually the SoC and a DRAM chip) while the S1 is actually multiple chips (including a SoC) in a single package. So I agree that the current name, while close, isn't perfect. "Application processor" is also an OK name, but has its own problems. For mobile devices the term "application processor" was originally used to distinguish the processor that ran a phone's UI from the "baseband processor" which ran the communication stack. Over time more and more of the system such as the screen controller, memory controller, graphics coprocessor, camera coprocessor and other functionality was integrated into the same chip as the application processor leading to a "system on a chip" (lots of other wireless functionality was integrated into the chip containing the baseband processor as well). So technically the term "application processor" really only refers to the main CPU inside the SoC, sort of like how "System on a chip" only refers to one of the chips inside a Ax PoP or S1 package. Changing the name to "Apple application processors" would expand the scope of this article considerably, since Apple has used application processors that were not a "System on a chip" in its Newton (see list at MessagePad#Newton_device_models) and early iPod devices. —RP88 (talk) 11:47, 2 October 2015 (UTC)
I'd be OK to expand the scope in that direction though.
One point that I didn't make is that the new name would solve the awkward problem with "prularization". This article is about more than one(1) SoC. And we had a debate where it should be "Apple systems on a chip", "Apple system on a chips", "Apple System on Chip", "Apple SoCs" and so forth.. each awkward and/or jargony so we settled on this name, which was the least bad. "Apple application processors" would solve that one easily. BUT.. Wikipedia doesn't have an article for Application processor, something that would have to be addressed. -- Henriok (talk) 14:43, 2 October 2015 (UTC)
Henriok, that is a good point about the pluralization problem. Apple application processors or Apple mobile application processors are indeed much less awkward in that regard, and I'd lean towards one of those names on that point alone (WP has a mobile processor article, but the article is really weak). —RP88 (talk) 15:53, 2 October 2015 (UTC)
If there aren't any objections, should we proceed with the article renaming to either suggestions by RP88? Personally, I'm leaning towards the latter name Apple mobile application processors since it makes it clear that the processors are targeting mobile devices. — AYTK talk. 01:28, 4 October 2015 (UTC)
Apple mobile application processors would be fine with me. —RP88 (talk) 05:01, 4 October 2015 (UTC)
Me too. -- Henriok (talk) 07:55, 4 October 2015 (UTC)

Apple A9 table

Could someone split the model numbers APL0898 and APL1022 into two lines in the section List of Apple processors? I don't see the reason for this as each model number deserves its own line. — Preceding unsigned comment added by 84.173.205.75 (talk • contribs) 16:18, 1 December 2015‎ ‎

That sounds like a good idea to me. I went ahead and split row for the A9 into two rows, one for APL0898 and one for APL1022. —RP88 (talk) 18:32, 1 December 2015 (UTC)
I merged the subrows in columns where the APL0898 and APL1022 are the same. Guy Harris (talk) 18:35, 1 December 2015 (UTC)

Hello fellow Wikipedians,

I have just modified 3 external links on Apple mobile application processors. 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, please set the checked parameter below to true or failed to let others know (documentation at {{Sourcecheck}}).

checkY An editor has reviewed this edit and fixed any errors that were found.

  • 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) 15:33, 16 October 2016 (UTC)

I found the first two pages elsewhere at Apple, and used the new URLs. The third works, but wasn't flagged as an actual dead link, which it is, so I fixed it. Guy Harris (talk) 18:39, 16 October 2016 (UTC)

GPU specs and Flops

I notice that the GPU specs and FLOPS have been changed recently. However, reading a few of the cited sources there are no GPU clock speeds given for many of these SoCs. And there certainly are not any FLOP ratings. I believe these should be removed unless a citable source can be found. (While this seems like a simple case, this will remove a lot of information from the page, so I am hesitant to do it without starting a conversation first.) Dbsseven (talk) 22:04, 8 November 2016 (UTC)

Yes. Please remove unsourced info. --Henriok (talk) 22:25, 8 November 2016 (UTC)

Apple Watch's S-Series being ARMv8 / 64bit is not substantiated

In the table it is claimed that the S1P and S2 are ARMv8 (64bit) but that is not substantiated by the references given.

The clock rates and other CPU attributes look suspect too. — Preceding unsigned comment added by Supersofts (talk • contribs) 19:23, 15 December 2016 (UTC)

Hello fellow Wikipedians,

I have just modified 5 external links on Apple mobile application processors. 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.

checkY An editor has reviewed this edit and fixed any errors that were found.

  • 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) 03:23, 8 July 2017 (UTC)

Apple Watch S3 processor

See https://www.apple.com/newsroom/2017/09/apple-watch-series-3-features-built-in-cellular-and-more/ "Powering Apple Watch Series 3 is the S3, Apple’s third-generation architecture featuring a faster dual-core processor ..." This info is not yet documented in this article. 84.173.216.75 (talk) 19:33, 15 September 2017 (UTC)

Done. Guy Harris (talk) 20:20, 15 September 2017 (UTC)

Geekbench comment sources

Based on recent edits by myself and @Dnywlsh: regarding the note at the top of the Geekbench benchmark section, it seems to be worth discussing the comment: "Note: These benchmark results are not comparable with other CPU/SoCs architecture due to the software codes optimization.[1][2]" While these appear to be in good faith, it appears to be worth discussing the sources and their reliability.

I will attempt to summarize Dnywlsh's comments:

  • The sources are "opinion pieces"
  • Sources articles are not sourced
  • The content is out of date
  • OSNews a blog.

I believe these to be incorrect.

  • The PC World source is clearly a reputable source by a reputable and the author Gordon Mah Ung an expert in the field[3]. (In particular the expert author is central as it is directly addressed in WP policy WP:NEWSORG.)
  • OSNews is not a blog, but rather a news website as noted it's WP article. If there is evidence OSNews is a blog and therefore does not meet the criteria of reputability, evidence must be provided (and weighed against the already present sources in the OSNews page.)
  • The articles themselves are the sources. I believe our role as editors is to evaluate the source, not the source's source. Otherwise this would defeat the point of WP policy encouraging secondary sources (see WP:SECONDARY).
  • The argument that the sources/content are out of date is unsupported and unlikely as the sources are from 2015 and 2016, later than many of the products listed on the Geekbench table.
  • Dnywlsh correctly points out that the PC World article states "some have claimed" without citation. However this is not the content reference here and the purported claims are addressed by the author directly in that article.
  • There is no evidence these articles are opinion pieces. An expert author discussing a topic on which they are knowledgable does not inherently make it an opinion piece. This need evidence or it is original research.

An expert's statements on the relevance of these results is directly applicable to this section. As I read them they meet the standard of reliable sources and should be included.Dbsseven (talk) 23:00, 1 October 2017 (UTC)

- Unfortunately, that's not what I said. The OSNews piece is written like a blog entry, with the author, Thom Holwerda, simply expressing his personal opinion about the subject. In his "article" (written and presented like a personal blog entry), he disagrees with John Gruber. Neither of these journalists are "experts in the field", they're simply technology journalists expressing their opinions. Notice that the OSNews piece does not cite any sources. The only links in the "article" take you to related OSNews pieces or John Gruber blog entries. I see no evidence that the OSNews opinion piece meets Wikipedia's standards for an appropriate source, since it itself is not sourced and is pure opinion. Note that the author says "I never cared much for benchmarks", etc. The entire "article" is based on his personal anecdotal evidence and opinion.
As for the PCWorld article, it was written in November 2015, and picks apart GeekBench 3, a test that is no longer used or relevant to this discussion. The scores listed on this Wikipedia page are GeekBench 4 scores, a completely new test. It's not clear that Gordon Mah Ung's (a tech journalist, not an "expert in the field") opinion about GeekBench 3 also applies to GeekBench 4. It's also not clear that his opinions about GeekBench 3 are accurate, since the article is solely written by him, are his personal opinions, and haven't been corroborated by an actual expert in the field. I suppose we could also debate what makes one an "expert" in this area. I personally don't feel that tech bloggers and journalists fall into that category. The page that you linked to is simply a PR statement by PCWorld bragging about his tech journalism experience. It actually says nothing about his technology knowledge or expertise or qualifications. Does he have a degree that makes him an expert in this field? Did he work at ARM or Apple for many years? I don't know. They just mention his 15 years at Maximum PC. Hardly relevant unless we're discussing his qualifications as a journalist. If I purposely wrote a blog entry with complete lies, would it also be deemed an appropriate source for Wikipedia, simply because you classify it as a secondary source? This is a slippery slope when we start deeming tech journalists' personal opinions as valid and correct sources without question.
If you finished the paragraph, he bases his entire argument against GeekBench 3 based on what "some have claimed": "Some have claimed the problem is how Geek Bench 3 weights its results. SHA2 encryption, for example, is overly represented for CPU performance. Given the hardware acceleration in the A9X, it’s showing Apple’s chip to be far faster than that it actually is." We don't have a source that says this is accurate, or that GeekBench 4 suffers from this same apparent shortcoming. He doesn't even seem to know whether that statement is accurate. He's simply saying that is what "some people" are claiming is the reason.
I'd like to see current, accurate sources that support the statement in question: "These benchmark results are not comparable with other CPU/SoCs architecture due to the software codes optimization." What in the "software code's optimization" makes them not able to be compared? Neither of the sources adequately answer that question. Dnywlsh (talk) 01:30, 2 October 2017 (UTC)
Hi @Dnywlsh:. I am sorry you feel I mischaracterized you arguments. I have quoted them below where appropriated and address them individually.
  • "Written like a blog entry" - This is a personal opinion and therefore original research. Should be discussed here if to be used as a standard for editorial decisions.
  • "Gordon Mah Ung's... opinion" and "I personally don't feel" - Unambiguously a personal opinion of the editor and therefore original research.
    • Opinion pieces are explicitly valid for inclusion per WP:NEWSORG "Editorial commentary, analysis and opinion pieces... are reliable primary sources".
    • In fact, WP's definition of a secondary source states it "provides an author's own thinking". Therefore the argument that these sources' authors have given their view is exactly in keeping with WP policy on secondary sources.
  • "GeekBench 4 scores, a completely new test" - Nope, Geekbench 4 is a derivative of Geekbench 3[4][5][6]
  • "Gordon Mah Ung... a tech journalist, not an "expert in the field"" - 15 years benchmarking and comparing computers is insufficient to make one qualified to discuss how to benchmark and compare computers?!
    • "Did he work at ARM or Apple" - expecting a source to have worked for the manufacturer would seem explicitly contradictory to finding independent sources.
  • "it itself is not sourced" - The expectation of a secondary source providing citations in keeping with WP policy on secondary sources. My understanding is editors are to evaluate a source, not a source's sources. Otherwise the becomes a turtles all the way down problem (see Infinite regress).
  • "If I purposely wrote a blog entry with complete lies" - You have no evidence of lies. This is a straw man argument and an informal fallacy.
More directly to the issue at hand, Geekbench 4 continues to be a controversial means of comparing desktop versus mobile CPUs, even as recently as last week.[7][8] I propose keeping the prose but updating to these sources. Dbsseven (talk) 14:03, 2 October 2017 (UTC)
How sad. You're deliberately twisting the facts to suit your narrative. Let's look at the entire quote: "Editorial commentary, analysis and opinion pieces, whether written by the editors of the publication (editorials) or outside authors (op-eds) are reliable primary sources for statements attributed to that editor or author, but are rarely reliable for statements of fact."
Neither of those sources support a statement of fact. They're purely the author's opinion. I'm allowed to state my opinion. That's generally how editorial decisions are made on Wikipedia. Are you new here? The entire website is editors arguing about content for articles. That's kind of how it works and is the entire purpose of the website.
The article you cited specifically mentions that GeekBench 4 is more accurate than GeekBench 3, and the weight of the scores have been rebalanced, which was Gordon Mah Ung's primary criticism of GeekBench 3. It sounds like we agree about this, since the article you shared supports what I previously said. They aren't the same test.
I didn't say anyone lied, I was simply providing a hypothetical example. If I wrote a blog entry or article about a topic that you weren't familiar with that was completely false, would you deem it an acceptable source, simply because it's a secondary source? How do you validate accurate sources? I don't consider blog entries of an author's personal opinion to be very accurate, and neither does Wikipedia.
I'm glad you know how to use big words like "straw man" and "informal fallacy" and quote Wikipedia articles to me in an effort to sound intelligent, but it's not working I'm afraid. Poorly attempting to dismiss my entire argument by saying it's a straw man argument and fallacy doesn't mean anything to me or most people, and is completely meaningless. But I hope saying it made you feel better! Dnywlsh (talk) 20:40, 2 October 2017 (UTC)

I agree with both of you on different points, and I really like this discussion, so please try to keep it civil. I for one doesn't like benchmarks mostly because they never will use the same tool over an extended period and in the end, they always seem to measure how good the benchmark is run, not necessarily how well the platform is performing buy any other software. In the case of benchmarks on iOS, there will quickly become an apple/pears/oranges-comaprison if the tester doesn't re-test all devices whenever a new device is tested, with an updated tool, running new OS releases. And in this particular case, Geekbench is testing the system (memory subsystem, storage, operating system, energy saving facilities, etc) not just the CPU, which this article is about. So, if you'd ask me, i'd rather skip the benchmarks, but if others would rather have them included, I'll not contest it. -- Henriok (talk) 21:22, 2 October 2017 (UTC)

Thank you Henriok. I am glad we seem to be making progress and I agree we should keep this civil and focus on the content. And I agree that in general these benchmarks are not encyclopedic. However they appear to be actively edited, so if they are to be included appropriate context should be given.
Let me try this again then, taking the same quote and shifting the emphasis, "Editorial commentary, analysis and opinion pieces, whether written by the editors of the publication (editorials) or outside authors (op-eds) are reliable primary sources for statements attributed to that editor or author, but are rarely reliable for statements of fact." Within this context the prose can be included, with appropriate cites, with the modification "Hardware reviewers have commented that these benchmark results are not comparable..." Again I would suggest that if the 2015/2016 cites are controversial there are new cites to the same effect be used.[8][7] Dbsseven (talk) 21:45, 2 October 2017 (UTC)
It's a little off the topic of this discussion, but I agree with Henriok that this list of iOS device benchmarks is not really a good fit for this article about the Apple's mobile application processors. It could be moved to a new section of List of iOS devices, where the context makes more sense. For that matter, I think an even better idea would be to move it to its own article — perhaps something like List of iOS devices sorted by benchmark — as I think there is enough content to sustain its own article. The new article could be linked from the "See Also" sections of both Apple mobile application processors and List of iOS devices. —RP88 (talk) 22:08, 2 October 2017 (UTC)
@RP88: I think this is an excellent suggestion. Benchmarks are best understood in the context of the device, as the device can certainly alter the SoC's performance characteristics (ie. throttling due to thermals). However, as I noted in my reply above, I still think if benchmarks are going to be included anywhere, appropriate context should be given. Dbsseven (talk) 22:36, 2 October 2017 (UTC)
This discussion seems to be idle. If there are no objections or further thoughts, I propose to move the benchmarks to the List of iOS devices, revising the note and updating the citations. Dbsseven (talk) 13:01, 11 October 2017 (UTC)
It's idle because my viewpoints were mocked and attacked lol. Do what you want. I accept defeat. So sad to see Wikipedia declining like this with overreaching and power-hungry editors. Dnywlsh (talk) 02:59, 12 October 2017 (UTC)
I have no objections. You may want to drop a quick note to Talk:List of iOS devices before moving the section (but that talk page looks pretty sleepy, so you probably won't get a significant reply). —RP88 (talk) 13:16, 11 October 2017 (UTC)

References

  1. ^ "Tested: Why the iPad Pro really isn't as fast a laptop".
  2. ^ "Apple's A10 Fusion, benchmarking, and the death of macOS".
  3. ^ https://www.pcworld.com/article/2844214/meet-gordon-mah-ung-pcworlds-new-hardware-guru.html
  4. ^ https://www.xda-developers.com/geekbench-4-how-the-processor-ranking-changed-under-the-new-more-accurate-benchmark/
  5. ^ http://www.geekbench.com/blog/2016/08/geekbench-4/
  6. ^ https://www.xda-developers.com/geekbench-ceo-fireside-chat-pt-2-oems-cheating-benchmarks-custom-cores-honest-manufacturers/
  7. ^ a b https://www.pcper.com/reviews/Processors/Apple-A11-Performance-Review-iPhone-8-Plus-Taking-Desktop
  8. ^ a b https://9to5mac.com/2017/09/22/iphone-8-geekbench-test-scores/

non-mobile parts

The T series parts, particularly the T2, are not mobile (or mobile specific) parts. Should these be moved elsewhere, or should the article be moved to a more general name? Dbsseven (talk) 20:05, 14 December 2017 (UTC)

And only the A and S series parts are application processors. I vote for either 1) a more general name or 2) splitting into four articles, one for each series. Guy Harris (talk) 21:01, 14 December 2017 (UTC)
I don't think dividing this list article into 5 separate list articles is a good idea. I'm inclined to leave things as they are for now, but I could probably be convinced to support a rename if a suitable name was suggested. A title like Apple ARM-based processors seems simultaneously too general (as Apple has used ARM in several chips not currently in this article) and too limiting (as I would want to include a hypothetical A13 chip in this article even if Apple switched the A-series to the Intel architecture). A name of Apple mobile processors would not resolve Dbsseven's objection regarding the T2. A name connecting Apple, processors, and iOS might work, but I'm not sure if the W series run a derivative of iOS (like the A, S, and T series do) —RP88 (talk) 05:02, 15 December 2017 (UTC)
The A series run iOS, but the S series run watchOS, and the T1 runs something that apparently calls itself "Watch OS", although I doubt it includes all of watchOS's UI code - it may be more like "the watchOS build of Darwin as a small embedded UN*X" plus some frameworks from watchOS - so they're not all running iOS.
And if Apple designed an A15 ARM application processor, put it into a Mac as the main CPU, and ran macOS on it, it wouldn't be running iOS, either.
So I wouldn't go with anything that includes iOS, even if we ignore the W series.
"Apple has used ARM in several chips not currently in this article" And, for at least a few fleeting moments, when the iPhone XI comes out, with an A12 processor, they will do so again. :-) The fact that it's an incomplete list of Apple ARM-based processors shouldn't make that name inappropriate; perhaps those chips should be in this article, unless the idea is that it should contain only chips that Apple designed (whether including Apple-designed CPU cores or not), and the chips in question had minimal design work by Apple.
"I would want to include a hypothetical A13 chip in this article even if Apple switched the A-series to the Intel architecture" So "Apple-designed processors" or "Apple-designed SoCs and SIPs"? Guy Harris (talk) 05:45, 15 December 2017 (UTC)
I like "Apple-designed processors". I think trying to break this into five articles is unnecessary and would lead to 5 stub articles.Dbsseven (talk) 18:50, 15 December 2017 (UTC)
bump. RP88 and Guy Harris, just want to check in before moving the article to "Apple-designed processors". This is a pretty popular article and want to make sure consensus has been reached. Dbsseven (talk) 17:08, 28 December 2017 (UTC)
Fine with me. Guy Harris (talk) 18:16, 28 December 2017 (UTC)
I don't particularly like that name as it somehow seems more general than the current scope of the article, but I'll readily admit I don't have a better name suggestion. —RP88 (talk) 00:15, 29 December 2017 (UTC)

Apple-designed processors?

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Perhaps we should redirect to Apple-designed processors (or move that here)? — stickguy (:^›)— || talk || 18:47, 22 June 2020 (UTC)

That was already moved here, but, unfortunately, the talk page wasn't moved at the same time.
I'm inclined to request that an administrator move the talk page, because the only current items on the page are related to the consequences of the half-completed move. Guy Harris (talk) 01:45, 23 June 2020 (UTC)
At the time, Apple Silicon was occupied with a redirect to a new draft namespace page, that redirect having a speedy delete tag (and it was deleted soon after). I figured suggesting it on a talk page would be better than changing the tagged page unilaterally - didn't realize it would cause its own issues. I have no objection with replacing this talk page with Talk:Apple-designed processors. — stickguy (:^›)— || talk || 02:56, 23 June 2020 (UTC)
I support renaming this article "Apple Silicon" and removing the other page. -- Henriok (talk) 10:45, 23 June 2020 (UTC)
Good, because that's what happened (except that Apple-designed processors wasn't removed, it was made a redirect to Apple Silicon - and, after some repair work, Talk:Apple-designed processors was moved to Talk:Apple Silicon and a new Talk:Apple-designed processors was created to redirect to Talk:Apple Silicon). Guy Harris (talk) 17:03, 23 June 2020 (UTC)
Yes. I liked it! I actually thought about doing just this late last night (Sweden time, when the State of the Union broke) but i fell asleep instead, and woke up to this! Marvelous! Good job! -- Henriok (talk) 19:00, 23 June 2020 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Requested move 24 June 2020

The following is a closed discussion of a requested move. Please do not modify it. Subsequent comments should be made in a new section on the talk page. Editors desiring to contest the closing decision should consider a move review after discussing it on the closer's talk page. No further edits should be made to this discussion.

The result of the move request was: moved to Apple-designed processors. Consensus is that this is the best title. – wbm1058 (talk) 01:15, 28 July 2020 (UTC)


Apple Silicon → Apple silicon – Officially used in press releases without the capitalized "silicon". Compare with Retina display example at Talk:Mac_Mini#Requested_move_4_August_2018. The "silicon" is a descriptor as with "display". RS seems to be split as to whether it is capitalized at the moment. - CHAMPION (talk) (contributions) (logs) 07:36, 24 June 2020 (UTC) —Relisting. Steel1943 (talk) 18:14, 13 July 2020 (UTC)

  • Unnecessary and needlessly bureaucratic, as his "Move back to the old title, per Master Of Ninja" references Master Of Ninja's comment above, which preceded the undiscussed move by three days, it's clear they mean the old hyphenated title. —Locke Cole • t • c 04:52, 30 June 2020 (UTC)
  • Support - Name the article how Apple calls it. We've been struggling to name this article for years, and I think we should take the lead of those who actually gave a name to it. There must be some kind of Wikipedia rule to not interfere with those who is the authority to actually name something. -- Henriok (talk) 19:20, 28 June 2020 (UTC)
  • Apple-designed processors is the correct article title. "Apple Silicon" is a marketing invention, and we don't do marketing. "Apple silicon" is never correct because it is neither a generic term nor the correct brand name. "Silicon" as a generic term to refer to processors is a niche neologism, not appropriate for a Wikipedia title, which should use common and easily-recognized terminology. 1.132.109.209 (talk) 00:59, 29 June 2020 (UTC)
  • As I attempted to explain yesterday, I think the article title should, at least for the moment, be reverted to Apple-designed processors, as that was the existing stable title of this article up until a week ago. Once that's done, I would be open to discussing ideas to improve it from there. In other words, I think we should return to the stable base condition first, and then entertain suggestions for change, rather doing it all in mid-air, so to speak. — Jaydiem (talk) 14:34, 30 June 2020 (UTC)
  • Revert to old title as per Jaydiem. Even as a brand name, "Apple silicon" wouldn't be appropriate, since that's specifically used for Macs, and this article covers a broader range of processors. Anterras (talk) 19:00, 2 July 2020 (UTC)
  • Not necessarily something that would count as a reliable source, but a quick Google search for "Apple silicon" on site:apple.com does show most of the references to "Apple {S,s}ilicon", as a phrase referring to the Apple-designed chips (as opposed to, for example, "Apple Silicon Engineering", referring to their chip designers, at jobs.apple.com) do, indeed, seem to specifically refer to Apple SoC's for Macs. Guy Harris (talk) 19:54, 2 July 2020 (UTC)
  • Revert to "Apple-designed processors" and leave it there: as I said in my comment on User:Anterras's vote, a quick look at Apple's Web pages seem to suggest that they're using the term "Apple {S,s}ilicon" only to refer to the SoC's they're planning to use in Macs. Perhaps they never bothered with a new term to describe what they used in iPhones and the like, as there was no need - it's not as if they transitioned from Somebody Else Silicon to Apple Silicon in a change that required developers to target a new instruction set when developing for iPhones/iPads/iPod touches. Guy Harris (talk) 20:43, 2 July 2020 (UTC)
  • Revert to "Apple-designed processors" – If the term "Apple silicon" gains better currency, capped or not, we can discuss further; no rush on that. Dicklyon (talk) 01:29, 27 July 2020 (UTC)

Meta

Procedural suggestion: I'm going to repeat here the gist of a comment I posted a few minutes ago in the section below this one; please look there for the complete explanation behind it. ~ I submit that the most appropriate way to handle this confusing situation — multiple moves, counter-moves, and discussions thereof all colliding with each other over the course of a week — is to "undo" back to a week ago and then start over. In concrete terms, this means: (1) cancel this Requested Move; (2) immediately move the article back to the stable, existing title it had as of a week ago, namely "Apple-designed processors"; and (3) then open up a fresh new Requested Move to discuss whether to leave that stable title unmolested for the time being, or to change it to something else, and if changed, to what. — Jaydiem (talk) 23:35, 29 June 2020 (UTC)

No. —Locke Cole • t • c 04:49, 30 June 2020 (UTC)
I'm reminded of the saying, attributed to Einstein, that it's difficult to solve a problem at the same level of thinking at which it was created. I'm more inclined to step back, get the bigger picture, and go from there, rather than to barrel ahead on a clearly screwed-up path. Evidently, you're not. Good luck with that. ¯\_(ツ)_/¯ — Jaydiem (talk) 14:13, 30 June 2020 (UTC)
Since apparently you need things explained to you, let me take a stab at it: the discussion above has been running for 6 days, requested moves are usually closed at the 7th day. There is no rule that says a requested move MUST only be for the page the nominator intended, and in fact, if you just do a !vote count of the comments so far, there's a pretty clear majority in favor of moving it back to the old title (note that my !vote was neutral, but supported a move away from the current title). Your misguided bureaucracy would drag this out for another week. See also WP:RMCOMMENT and WP:MOVED. Also, per Wikipedia:Requested moves/Closing instructions#Determining_consensus: Therefore, if a page has been moved from a long-standing title, and it is not possible to move the page back to its original title during the discussion, the default title will be the title prior to the contested move. For example, if an article is created at Soda can and stays there for years prior to being WP:BOLDly moved to pop can, and a move request is filed leading to a decision of "no consensus", the article must be moved back to its longstanding title. This is the case even if the original page was placed at pop can or fizzy drink can or orangutan-flavored soft drink can, as long as soda can took over through consensus and can be determined to be the actual long-standing title. If this results in no consensus, the closing admin should move it back to the previous "stable" title. —Locke Cole • t • c 23:56, 30 June 2020 (UTC)
You're quite humorous with those multiple barbs about "misguided bureaucracy" that you've thrown around, particularly as you go on coughing up such a long and pointless stream of obtuse bureaucratic minutia that's totally unnecessary in this context. As I have tried repeatedly now to explain, reverting the undiscussed move of June 22 is routine procedure for this kind of thing: the traditional protocol is "Be Bold, Revert, Discuss". You're going nuts because you fail to grasp that the guidelines about page moves in the midst of a discussion are meant to avoid someone trying to prematurely force a conclusion, not to prevent anyone from reverting an undiscussed change back to an established baseline in order to properly set the stage for the discussion. The protocol is not "Be Bold, Discuss Without Reverting, then Revert-But-Only-If-That's-the-Final-Outcome-of-the-Discussion". My suggestion would have simplified things, not complicated them. I wish you had been able to understand that. Anyway, the point is largely moot by now, so nevermind. — Jaydiem (talk) 05:43, 1 July 2020 (UTC)
You're quite humorous with those multiple barbs about "misguided bureaucracy" that you've thrown around Which is precisely why I simply said "No" the first time. But apparently that was an invitation for Einstein quotes and claims of barrel[ing] ahead on a clearly screwed-up path. So I explained the path to you, and now it's considered a long and pointless stream of obtuse bureaucratic minutia that's totally unnecessary in this context despite the fact that the result will ultimately be the same, and no need for a second move discussion because this one should be closed as having consensus for the old title. And strictly as an aside, your "recap" below did not help my impression of you when you singled my revert out as the first thing worth talking about (despite chronologically being the last thing to happen; I also provided justification for the move below). Now is there something else I can help you with, or are we done here? —Locke Cole • t • c 17:13, 1 July 2020 (UTC)

Unrequested move 27 June 2020

Apple Silicon was boldly moved to Apple designed processors on 27 June 2020.

There are two issues with this move:

  1. it wasn't discussed;
  2. "Apple designed" should be "Apple-designed", as it's a compound modifier. and the unhyphenated version could be parsed as a sentence making the (true) claim that Apple (the corporation) designed processors (namely the processors listed on this page).

For the first item, the move should be proposed here so that it can be discussed.

For the second item, no, an editor can't move the page back to "Apple-designed processors", because there's a redirect in the way, but, if we do get a consensus to undo the "Apple-designed processors" -> "Apple Silicon" move, rather than to further move to "Apple silicon", the way to do that move is to request a move, so that an administrator can move the redirect out of the way to allow the move to be done.

See Wikipedia:Requested moves#Undiscussed moves for some comments on the consequences of Just Renaming and a note that such a move might Just Be Undone, and Wikipedia:Requested moves/Controversial for moves that might be controversial (such as this one).

See Wikipedia:Requested moves/Technical requests for information on how to request that this be moved back to "Apple-designed processors" if that's the consensus. Guy Harris (talk)

Agree. I think the original move to "Apple Silicon" however is controversial as it is misnamed. Would recommend most people actual contribute to the discussion above, rather than make new moves until this all gets sorted out. Master Of Ninja (talk) 10:42, 28 June 2020 (UTC)
I didn't make that move, but whoever did was correct, except for the missing hyphen. See the section above for why. 1.132.109.209 (talk) 01:01, 29 June 2020 (UTC)

— — —

OK, folks, let's recap. Locke Cole's June 27 move from Apple designed processors [sic] to Apple Silicon was undiscussed — which is acceptable, if it's interpreted as a nearly-immediate reversion of Abobeck11's June 26 move from Apple Silicon to Apple designed processors [sic], which itself appears to have been undiscussed. However, I think Guy Harris and others commenting here are overlooking the fact that what started this cascade of moves and counter-moves was ViperSnake151's June 22 move from the previously stable title Apple-designed processors to the new title Apple Silicon, which was—you guessed it—undiscussed. Abobeck's move, then, should be construed as I believe it was intended: a straightforward reversion of ViperSnake's move (I think we can presume that the missing hyphen was an accident). Under the traditional "Be bold, revert, discuss" editing protocol of Wikipedia, neither ViperSnake nor Abobeck did anything wrong here. The complex tangle of multiple moves, and multiple Talk-page sections about those moves, arose because after ViperSnake initially changed the title, no one properly reverted that move before opening a discussion about it. Now, we're caught in a rather sticky web of confusion. Ugh.
⋮
In my view, what really, really ought to be happening here is this: (1) We roll back all the title changes to the point in time just before when ViperSnake made the initial change a week ago. (2) We close out both this section and the stillborn Move Request above it; I would recommend a complete "strikeout" overmarking of both, to emphasize this. (3) We start over with a fresh discussion of whether to change the previously-stable title Apple-designed processors to something else, and if so, to what. — Jaydiem (talk) 23:15, 29 June 2020 (UTC)

"(I think we can presume that the missing hyphen was an accident)" I don't - it may have been an attempt to avoid the existence of an "Apple-designed processors" redirect left behind by the move to "Apple Silicon". Pro tip: if there's a redirect in the way of a move, e.g. there's some history preventing it from just being treated as an undo of a move, ask an administrator to do it for you. Guy Harris (talk) 23:32, 29 June 2020 (UTC)
@Guy Harris: You’re right, I shouldn't've said "accident". My point was simply that Abobeck's move appears to have been intended as a conventional reversion of ViperSnake's, rather than as a separate, subsequent move, unrelated to the previous one. — Jaydiem (talk) 04:47, 30 June 2020 (UTC)
There was already a requested move in progress when that subsequent move was made without discussion. Per the template at the top of the article page right now: Please do not move this article until the discussion is closed. I simply reverted back to the status quo as it existed when the RM (directly above) was started. —Locke Cole • t • c 04:49, 30 June 2020 (UTC)
Thanks @User:Jaydiem for summarising the situation. What I indeed was referring to was reverting back to the pre-22 June edit where the original title was "Apple-designed processors". I do also note that the subsequent move was procedurally wrong as there was a move discussion in progress. - Master Of Ninja (talk) 12:02, 30 June 2020 (UTC)

Sorry guys - I assumed the change to Apple Silicon was also an unrequested move and definitely didn’t look right with a capital “S” in silicon. I also didn’t check the talk page which my mistake. Let’s start a new discussion for the proposed move to Apple-designed processors. Abobeck11 (talk) 16:20, 30 June 2020 (UTC)

@Abobeck11: Re: "I assumed the change to Apple Silicon was also an unrequested move" — Indeed it was. Your reversion was entirely in order, per WP:BRD, and it should have been recognized and respected as such. It had nothing whatsoever to do with the entirely separate discussion that somebody had started to examine whether the "S" in "Apple Silicon" should be capitalized — a question which would be relevant to the article regardless of its title. Alas, not everyone gets that. ¯\_(ツ)_/¯ — Jaydiem (talk) 19:08, 30 June 2020 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Archives

I ported the header from the old talk page, but I'm not sure how to quickly pack that page's talk page and its single archive into a template. Ehler (talk) 00:07, 23 June 2020 (UTC)

See above. I'll request that an administrator do whatever magic is needed to move Talk:Apple-designed processors here, and then either move the archives myself or, if I can't, ask an administrator to do it. Guy Harris (talk) 01:45, 23 June 2020 (UTC)
All done, by the aforementioned administrator. Guy Harris (talk) 04:24, 23 June 2020 (UTC)

Large gap in this talk page's edit history

If you look at the edit history of this-here talk page right now, you'll find a yawning thirty-month gap between 2017-12-29 (when Apple mobile application processors was moved to Apple-designed processors) and 2020-06-22 (when Apple-designed processors was moved to Apple Silicon). It seems a shame to lose two-and-a-half years of editing history, so I'm curious: How feasible would it be for someone (I presume an Admin) to recover the missing history and splice it into the gap from which it's missing? ~ Thanks, — Jaydiem (talk) 19:50, 30 June 2020 (UTC)

@Anthony Appleyard: Was there any history there? If so, is there a way to get it back? Guy Harris (talk) 20:08, 30 June 2020 (UTC)
  • The likelihood that there truly was absolutely nothing said in the talk page for thirty straight months — and that such a silence just happened to begin and end at the precise instant of each of two page moves — is infinitesimal. It's a virtual certainty that something got mucked up in the process of moving/recreating/splicing multiple incarnations of the talk page, starting on Monday of last week, when ViperSnake151 moved the article page and somehow the talk page didn't come along with it. — Jaydiem (talk) 04:20, 1 July 2020 (UTC)
  • @Anthony Appleyard: See, for example, the first two comments in the "Archives" section; they aren't in the history, but User:Ehler added the first one at 00:07, 23 June 2020 (UTC) and I added the second one at 01:45, 23 June 2020 (UTC). This edit, by User:Anthony Appleyard, with the edit comment "text-merge talk pages", apparently put those comments, and some comments from the "Apple-designed processors?" section, back in, presumably merged from some other page, without the history being merged. There may well have been no other talk page edits that had to be merged - perhaps the page was sufficiently non-controversial that nobody had anything to say until the recent moves. Guy Harris (talk) 05:22, 1 July 2020 (UTC)
  • What I suspect happened here is:
    • At 21:34 UTC on 22 June 2020, User:ViperSnake151 moved "Apple-designed processors" to "Apple Silicon", but didn't move the talk page. At that point, there was still a "Talk:Apple-designed processors" page, and there was no "Talk:Apple Silicon" page.
    • At times between 00:07 UTC on 23 June 2020 and 02:56 UTC on 23 June 2020, a "Talk:Apple Silicon" page was created, the "Apple-designed processors?" and "Archives" sections were created in it, and further comments were added to them.
    • At 04:09 UTC on 23 June 2020, "Talk:Apple-designed processors" was moved by User:Anthony Appleyard to "Talk:Apple Silicon", discarding the old "Talk:Apple Silicon" page.
    • At 04:10 UTC on 23 June 2020, User:Anthony Appleyard took the comments that were in "Talk:Apple Silicon" before it was discarded and put them into the new "Talk:Apple Silicon" (which is the page that was "Talk:Apple-designed processors").
  • (I'm not sure what happened with User: stickguy's 18:47 UTC on 22 June 2020 comments. Those appear to have been made before the 21:34 UTC on 22 June 2020 move to "Apple Silicon", but maybe that wasn't 21:34 UTC.) Guy Harris (talk) 09:46, 1 July 2020 (UTC)

The edits you couldn't find were deleted and buried under the same talk page. After I moved this from Talk:Apple Silicon to Talk:Apple-designed processors, I restored the deleted edits at Talk:Apple Silicon and history-merged them into this page's history. Yes, there really were no talk edits between 16:11, 29 December 2017 and 18:47, 22 June 2020. — wbm1058 (talk) 01:38, 28 July 2020 (UTC)

Evolution of Apple "A" series chart

Henriok, you recently made a significant update to the Evolution of Apple "A" series chart. The colors to indicate whether an A-series chip is still in production looks good. However, can you explain what the dotted lines in your update now are intended to indicate? I am particularly confused by the dotted line between the A6X and the A8X. The A7 was such a leap in performance and capabilities that it replaced both the A6 and A6X across Apple's mobile phone and tablet lines. If the solid lines are supposed to indicate technological inheritance and the dotted lines are supposed to indicate product successor, then the the product successor of the A6X should be the A7 and the product predecessor of the A8X should also be the A7. I can't think of what a "dotted line" relationship between the A6X and A8X would be. In the original chart I think the solid lines were product successor/"major" tech inheritance and the dotted lines were "minor" tech inheritance. —RP88 (talk) 22:07, 13 August 2020 (UTC)

RP88, I could never figure out what the previous dotted vs straight lines was supposed to indicate. Your explanation makes some sense, but since the article is technical in nature, I think marketing/product relationship would be the wrong focus. In my chart, straight is technological implementation, i.e. they share the same core michroarchitecture. The dotted is something looser and undefined, like marketing, product succession or packaging. The X-processors are obviously related on their own terms, but they absolutely share more technicalities with their non X siblings, so the lines connecting X:es should not be as prominent.
There is a notes field in the chart left blank. I tried to write something there explaining the chart but I couldn't make it look good.
The chart also could/should be expanded to include the three types of A5 and the two types of A7 and A9. They are as distinct as the A12X and A12Z are from one another.. but they didn't get a new cool name, just a not as cool new product code (which A12Z didn't get). But WOW! the chart is hard to edit :O -- Henriok (talk) 07:49, 14 August 2020 (UTC)
OK, under the theory that two chip columns indicate a chip's primary packaging (compact or larger), the solid lines are technical inheritance, and the dotted lines are now the "softer" marketing / product succession, I've updated the chart to connect dotted lines between the A6X and A7, the A7 and A8X, and removed the dotted line from the A6X and A8X under the theory that Apple described and marketed the A7 as being powerful enough to be the spiritual replacement for both the A6 and A6X, but with the next generation (i.e. A8 and A8X) Apple went back to separate chips for the tablets. I also made some edits that might make the chart code easier to edit by making the element layout easier to decode. —RP88 (talk) 00:07, 15 August 2020 (UTC)

Proper noun?

Why is it a proper noun here (Apple Silicon) and not a common noun (Apple silicon)? Does Apple marketing matter? --Mortense (talk) 15:30, 22 September 2020 (UTC)

I believe Apple marketing also capitalises it as "Apple silicon", it's just that media has been calling it "Apple Silicon" thus making it the more common form. IMO it should be treated as a common noun in the body of articles. --17jiangz1 (talk) 13:25, 18 November 2020 (UTC)

Apple capitalizing the "s" in "silicon" in "Apple silicon"?

If I search for "apple silicon" on Apple's Web sites other than developer.apple.com, support.apple.com, jobs.apple.com, discussions.apple.com, podcasts.apple.com, and apps.apple.com, they largely seem to use "Apple silicon":

If I search developer.apple.com, the first page of results largely seems to show "Apple silicon"; the same is true of support.apple.com.

jobs.apple.com, however, shows "Apple Silicon" on the first page of results. Some of that is the title of the organization "Apple Silicon Engineering", but others appear to refer to the chips as "Apple Silicon".

discussions.apple.com, podcasts.apple.com, and apps.apple.com are biased by third-party content, i.e. users who think it's called "Apple Silicon", podcast authors who think it's called "Apple Silicon", and app developers who think it's called "Apple Silicon", so that may reflect what some of the world outside Apple has decided it's called.

So that seems to support what User:17jiangz1 said in Talk:Apple-designed processors#Proper noun? - "I believe Apple marketing also capitalises it as "Apple silicon", it's just that media has been calling it "Apple Silicon" thus making it the more common form." Guy Harris (talk) 21:30, 22 January 2021 (UTC)

Retrieved from "https://en.wikipedia.org/w/index.php?title=Talk:Apple_silicon/Archive_1&oldid=1068819835"