Talk:Qualcomm Snapdragon/Archive 1

Vapourware
This processor is widely assumed to be vapourware, nobody has ever seen one to my knowledge and Qualcomm have been caught faking Snapdragon demonstration units in the past (IE: claiming a chipset that isn't snapdragon is, then when caught saying it was just supposed to give an idea of what snapdragon might look like)

Is it even notable enough for inclusion? Hideki (talk) 14:21, 8 January 2009 (UTC)


 * I believe that Qualcomm was demoing the Snapdragon at CES 2009, so there should be detailed reports about it in a day or two. In general, I think that a major planned chip of a major industry player is notable. Dimawik (talk) 15:07, 10 January 2009 (UTC)

Snapdragon is definitely out and powering quite a number of Android phones now, among other devices. Radaghast (talk) 00:22, 15 September 2010 (UTC)

FPU
Do these CPUs have FPU or not? --217.195.52.165 (talk) 22:20, 16 June 2010 (UTC)

Model Table
I'm glad to see that the table is working out for everybody. Glad I took the time to do it. Any thoughts on improvements to the table structure? What can we change the Instruction set to that would be a good measure of processing core performance over and above clock speed? DJTachyon (talk) 08:49, 29 September 2010 (UTC)
 * It appears to me so full of "Utilizing Devices" that it turns the rest of the table rather sparse and overlong. Could use splitting out in a separate table perhaps; one with just the device model and the chip model. — Preceding unsigned comment added by 84.86.176.163 (talk) 20:55, 20 November 2011 (UTC)

"Wireless Technology" - clarification needed
Currently, this article's table features a column called "Wireless Technology". How is this to be interpreted? Are the entries meant to be understood as such that support for a particular technology in hardware is provided (e.g. HSUPA implemented in hardware)? --Abdull (talk) 10:14, 7 October 2010 (UTC)

MSM8250?
Is this Chipset really existing? I don't find any MSM8250-Chipset on the Qualcomm Website, so i think this is a mixing up with the old QSD8250, likely because almost all other older and newer Qualcomm chips are MSMs (e.g. MSM72xx, MSM8255). Microsoft says, the new HTC WP7-Devices like HTC 7 Mozart, 7 Surround, 7 Trophy and HD7 have the QSD8250.

http://www.microsoft.com/presspass/presskits/windowsphone/glance.aspx and http://www.qualcomm.com/products_services/chipsets/snapdragon.html

This Wikipedia-Entry confusing many people all around the world (because of the WP7-Launch), so please can anyone fix it? I'm not sure enough the MSM8250 isn't really existing... —Preceding unsigned comment added by 84.74.50.151 (talk) 10:37, 15 October 2010 (UTC)

Qualcomm Hexagon
Which Snapdragon SoC has Qualcomm Hexagon DSPs interegated into them? `a5b (talk) 17:18, 23 May 2012 (UTC)

--- All S3, S4 Pro, S4 Plus, S4 Prime and Snapdragon 800 and 600 have Hexagon integrated in them.

Proposed merge
I'm thinking that the article specifically on the Krait family should be merged with this article - as the content is already discussed here, it doesn't seem to merit its own page. I was going to just go ahead and redirect, but some consensus would be much better. Thanks. Noir (talk) 14:47, 2 June 2012 (UTC)
 * Neither Krait (CPU) nor Scorpion (CPU) have any merit as separate articles. Redirected them here and removed wikilinks to them as to avoid self-redirects. --Dmitry (talk•contibs) 16:33, 2 June 2012 (UTC)
 * Sounds good to me, thanks. Noir (talk) 16:51, 2 June 2012 (UTC)
 * Krait and Scorpion pages added back, added technological information that doesn't work well on this page. Removed merge template. User931 22:591, 3 June 2012 (UTC)
 * The one and same technical table you added to these respective pages can easily be placed in this article. Reverting.
 * I think we need more consensus on this before you do drastic changes like this, the page looks very messed up and this page is about the System on chip, not about specific CPU architecture. Reverting back. User931 23:35, 4 June 2012 (UTC)


 * Please do not remove MSM8974 as it was referenced in the leaked roadmap, just like most other current and upcoming S4 processors which aren't even listed on Qualcomm's website. --Dmitry (talk•contibs) 19:19, 4 June 2012 (UTC)
 * All current and upcoming S4 SoCs are listen on Qualcomms website, you just haven't taken a good look. https://developer.qualcomm.com/sites/default/files/snapdragon-specs.pdf The MSM8974 is not listed because it has been cancled. Reverting. User931 23:35, 4 June 2012 (UTC)
 * No, the MSM8974 has NOT BEEN CANCELLED. The document that you refer to lists what is currently available to OEMs. You need to provide a reference that states explicitly that it has been canceled. The fact that it is not on that list does not say anything. In a few short months or weeks, you will see the announcement of the MSM8974. — Preceding unsigned comment added by Dominik78 (talk • contribs) 01:20, 22 September 2012 (UTC)

So we can remove MSM8974 from S4 prime list as it is included in new Snapdragon 800 series. S4 prime hence should only contain MPQ8064.

Snapdragon S3 and HTC Sensation
Table includes this phone at 1.5 GHz. In fact, only the Sensation XE runs at 1.5 GHz, the original Sensation clocks in at 1.2 GHz (but still using Snapdragon S3).

Source: HTC Sensation — Preceding unsigned comment added by 90.229.34.175 (talk) 07:03, 19 June 2013 (UTC)

About recent edits
Why delete the MSM7225AB even sources supplemented and why remove those note regrading QRD?C933103 (talk) 01:14, 23 June 2013 (UTC)

style="font-size" in tables
Have you ever thought about visually handicapped persons? If a table is to wide, there are other solutions: narrowing columns and/or deleting columns (Wikipedia is an encyclopedia and not a collection of datasheets). --DrSeehas (talk) 17:40, 2 September 2013 (UTC)
 * Per WP:FONTSIZE, font-size changes are completely acceptable on templates & tables. Additionally, your constant removal of them are in violation WP:STYLEVAR. 「gu1dry」 ⊤ • ¢ 18:03, 2 September 2013 (UTC)
 * I read in WP:FONTSIZE: "Editors should avoid manually inserting large and small fonts into prose. Increased and decreased font size should primarily be produced through automated facilities such as headings or through carefully designed templates. Additionally, large tables may require a decreased font size in order to fit on screen..." I read in this context "may require" as "if no other solution is available". I have mentioned two other solutions. So you should first think about these before using "font-size". Again: Wikipedia is an encyclopedia and not a collection of datasheets.
 * Manual of Style is very long. Please explicate which section I am violating. --DrSeehas (talk) 07:35, 3 September 2013 (UTC)

We're still trying to solve the issue of over-crowded utilizing devices columns for this article and others like it. If anyone is interested in commenting, please join us at Talk:MediaTek/Archives/2015 Charwinger21 (talk) 09:04, 10 November 2014 (UTC)

Unifying dates
 This conversation has been moved from my talk page. --  Ohc  ¡digame!¿que pasa? 02:18, 5 September 2013 (UTC)

you kindly unified date formats. I tried the same, but User:Gu1dry fights a private edit war against me because Gu1dry was blocked twice after I reported User:Gu1dry to the administrators and as a payback User:Gu1dry reverts nearly all my edits :-( Maybe you could unify the dates in Snapdragon (system on chip)? Thank you. --DrSeehas (talk) 17:18, 4 September 2013 (UTC)
 * I wouldn't revert your edits if you didn't violate the WP:MOS. If I was "fighting a private edit war against you", I could find plenty of your "contributions" that were in violation of the WP:MOS. I'm really getting tired of this slander & harassment. 「gu1dry」 ⊤ • ¢ 17:36, 4 September 2013 (UTC)
 * Gu1dry does have a point in that there is no prohibition of short month names in ref section, the big proviso is that this be consistently applied. Looking at Gu1dry's version, I see absolutely no consistency, so my harmonisation is the right way to do it. There is absolutely no MOSNUM violation with DrSeehas' edit [redacted]. --  Ohc  ¡digame!¿que pasa? 17:40, 4 September 2013 (UTC)
 * The article is consistent. All dates are in dd Month YYYY format except the the accessdates are in dd Mon YYYY format (which are acceptable per the WP:DATEFORMAT) to make the reference section much easier to read. 「gu1dry」 ⊤ • ¢ 18:05, 4 September 2013 (UTC)
 * You have to understand the world of User:Gu1dry: e.g. 11 times 2013-MM-DD is consistent with dd Month YYYY and dd Mon YYYY. ;-) --DrSeehas (talk) 18:26, 4 September 2013 (UTC)
 * This cheekiness is starting to really get annoying. Please stop trolling. Out of the 126 references in the article there were EIGHT references that had the accessdate formatted different the remaining references, but instead of converting the non-consistent accessdate dates to the consistency, you created a completely new consistency, which violates WP:STYLEVAR (the fourth time I had to mention that exact paragraph to you or a discussion including you). 「gu1dry」 ⊤ • ¢ 18:49, 4 September 2013 (UTC)


 * There should not be any issue with "long" or "short" months so long as it's consistent, but the date format is not consistent throughout. Even User:Gu1dry's latest preferred version is a mess. Without counting "May" as a spelt-out month, I count more than half of the access dates amongst citations having spelt-out months. The number of short months is considerably less than half of the number of access dates in that article. That's hardly consistency in anyone's book. This version, before anyone appeared to have started fighting over date formats, shows that there is no consistency. Anyway, there is no rule that says access dates have to be included, or that their format can or should be different from publication dates. It just looks weirdly quixotic to have publication dates spelt out and access dates in their three-lettered form. IMHO, you have a choice of having all short dates, or all long dates, but we should keep dates the same throughout. Also, it's not all that difficult keeping an article's dates fully aligned. When there is a tool such as WP:MOSNUMscript, the work becomes a piece of piss. I've just updated the article's citations and aligned all the dates. Regards, --  Ohc  ¡digame!¿que pasa? 02:10, 5 September 2013 (UTC)
 * Gu1dry has reported me for edit warring.
 * @Ohconfucius: Sorry for implicate you in this issue as you are now the next target for Gu1dry's edit warring. --DrSeehas (talk) 14:52, 5 September 2013 (UTC)

In-ref italicisation
There are currently an excess of populated parameters that are incorrect in this article. Another editor seems to own this article and keeps on warring with me to retain their preferred version that does not comply with MOS:ITALICS. I'm not going any further down this road but only to state Engadget is considered a periodical, hence italicised (the Engadget article itself reflects this); LG Hong Kong, Huawei, Sony Mobile Developer World are not "works" and therefore ought not to be italicised. --  Ohc  ¡digame!¿que pasa? 10:06, 5 September 2013 (UTC)

Merging identical cells
Please, merge identical tables cells. Artem-S-Tashkinov (talk) 05:50, 16 September 2013 (UTC)

Snapdragon 400 Discrepancies
Quallcom states Snapdragon 400 as

Krait 200 up to 1,7 GHz (https://developer.qualcomm.com/discover/chipsets-and-modems/snapdragon)

Krait 300 up to 1,7 GHz (http://www.qualcomm.com/snapdragon/processors/800-600-400-200/specs)

Which one is correct?

Then it says in the table

8930AA is up to 1,7GHz but both 8930AA-devices, namely the HTC First & One Mini, only clock up to 1,4GHz.

(http://www.htc.com/us/smartphones/htc-one-mini/#specs)

(http://www.htc.com/us/smartphones/htc-first/#specs)

188.96.80.87 (talk) 13:02, 16 November 2013 (UTC)


 * It seems both Krait (200) and Krait 300 are present in SoCs under the S4 brand name. Krait 300 are likely ones introduced around the time of the Snapdragon 600 for obvious reasons, but don't take that as a rule of thumb. Part numbers are also potentially confusing e.g. APQ8064 is Krait (Krait200), whereas APQ8064T (and various other suffixes) are Krait 300, including some marketed as S4 Pro. I'm really not sure about the MSM8960T though; there are conflicting stories on the web and I can't seem to find a reliable source proving one way or the other.Nathen_1 (talk) 00:45, 1 January 2015 (UTC)

Rounding of frequency numbers
Currently the specs in the article has some of the frequency numbers rounded and some of them not, for instance the MSM8974AC is listed as "up to 2.5", when in reality it will only boost up to 2.45 GHz. The reason for this is fairly obvious, since it is caused by Qualcomm generally reporting rounded numbers.

Sources for the 2.45 boost value of MSM8974AC, if anyones in doubt:

http://www.anandtech.com/show/7783/qualcomms-snapdragon-801-msm8974ac-the-new-32bit-flagship-until-805

http://www.trustedreviews.com/opinions/snapdragon-805-vs-801-vs-800

http://www.tomshardware.com/reviews/snapdragon-801-performance-xperia-z2,3777-2.html

http://arstechnica.com/gadgets/2014/02/meet-qualcomms-snapdragon-801-a-slightly-faster-snapdragon-800/

http://www.androidauthority.com/snapdragon-801-vs-800-353448/

http://gizmodo.com/qualcomms-snapdragon-801-is-2014s-must-have-mobile-ch-1529553545/all

http://www.phonearena.com/news/Qualcomm-announces-Snapdragon-801---45-faster-image-processing-in-devices-by-April_id53015

And explanation of Qualcomm's rounding of numbers:

http://www.anandtech.com/show/7846/the-difference-between-snapdragon-800-and-801-clearing-up-confusion

This obviously doesn't just apply to MSM8974AC, but to basically every SOC in the article, for instance the Snapdragon 800 chips are currently listed at 2.26 GHz, instead of Qualcomms rounded 2.3 GHz value (http://www.qualcomm.com/snapdragon/processors/800).

Question is should the article use Qualcomms rounded numbers, the more exact non-rounded numbers or should it just not care about consistency and just use whatever (since it'll probably be hard to find a consistent set of numbers)? — Preceding unsigned comment added by 129.78.233.210 (talk) 04:25, 2 June 2014 (UTC)

Windows 8?
Sure not Windows RT?  Sky 6t  05:43, 28 July 2014 (UTC)

Semi-protected edit request on 9 March 2015
There's a typo regarding Snapdragon 820: "to be available for sampling in H2 2016" is wrong - should be 2015; see the reference.

Schreib70 (talk) 10:44, 9 March 2015 (UTC)
 * Yes check.svg Done with thanks, Nici  Vampire  Heart  11:28, 9 March 2015 (UTC)

Microarchitecture in infobox
In the infobox, microarchitecture is stated to be A8/A15, which is incomplete and seems a bit out of place. Qualcomm design their own cores along with ARM's IP cores, and they use more than just A8 and A15; in fact as far as I know they haven't released an A15 part. Are there any objections to removing this line, or perhaps suggestions to improving it? A simple list of used microarchitectures could be a bit unwieldy.Nathen_1 (talk) 16:09, 27 March 2015 (UTC)


 * I see that the current infobox template is for processors like Krait (CPU), but I also can't find any infobox templates specifically for SoCs like Snapdragon. I think there is probably a better template we could use alltogether.


 * Disclosure: I have a financial connection with Qualcomm. CorporateM (Talk) 19:04, 27 March 2015 (UTC)

Inaccurate/Speculative Snapdragon 820 Information
Much of the information in the Snapdragon 820 section is either wrong or unknown. Also, the references lead to dead links. It is not 8 cores, it is 4: https://www.codeaurora.org/cgit/quic/la/kernel/msm-3.14/commit/arch/arm/boot/dts/qcom/msmthulium.dtsi?h=LA.HB.1.1.1_rb1.10&id=7392a87d90c84b2cdf96a9726311b110214d8c5b

All of the listed devices use the Snapdragon 810 chip rather than 820. I'm not sure what the Hexagon DSP frequency will be, but 1000MHz appears to be speculation. Based on the source code, the top frequency appears to be 1.6 GHz rather than 2.857 GHz https://www.codeaurora.org/cgit/quic/la/kernel/msm-3.14/tree/arch/arm/boot/dts/qcom/msm8996.dtsi?h=LA.HB.1.1.1_rb1.10

I have attempted to correct the misinformation, but my changes have been reverted and I'm guessing that if I try again, it will be overwritten. — Preceding unsigned comment added by 71.14.176.152 (talk) 18:06, 26 April 2015 (UTC)

810 overheating/throttling
I have added a paragraph covering the rumors and press coverage of the 810 and specifically how it has been rumored to run hotter than competitors and previous generation chips. Please edit it if the paragraph isn't objective enough or I have missed notable press coverage. — Preceding unsigned comment added by Theoriginalandrewia (talk • contribs) 05:11, 27 May 2015 (UTC)

Draft
I have a COI/financial connection with Qualcomm and would like to help improve this article, bringing it up to Good Article status eventually. The current page, from my perspective, is an indiscriminate collection of specifications data, whereas the ideal page would be a prose-style summary of the chip-family's history, focused on reliable, secondary sources.

In keeping with WP:COI, I've prepared a draft for consideration and review at:
 * User:CorporateM/Snapdragon

Most of the draft article is very positive, as the chips do enjoy a good reputation; the exception being the controversy around the Snapdragon 810. Certainly that is the area where my conflict of interest is most relevant, given that Qualcomm's view differs so greatly from an NPOV representation of the total body of literature. I think did a fair job covering it, but I did take a shot at a bit of a different version under the Benchmarks section that focuses more on secondary sources.

Thanks in advance to anyone that takes the time to fairly review and consider my suggested replacement. CorporateM (Talk) 00:09, 23 June 2015 (UTC)


 * Thank you . I like your draft and think you summarized my edit well without making it too minimal.  Two recent 810-related developments that might be worth mentioning are the OnePlus Two's decision to use a Snapdragon 810 v2.1 underclocked to 1.8GHz and Anandtech's findings that that 810 v2.1 made several performance improvements to be competitive with Samsung's Exynos 7420, but "both variants of the Snapdragon 810 still see the unfortunate characteristic of ultimately forcing all threads off of the A57 cluster to stay within TDP limits in high load conditions".  Theoriginalandrewia


 * Thanks ! Do you happen to have a link to the AnAndTech article you're referring to and where the OnePlus Two information came from? I only saw the AnAndTech story that was more or less damning. CorporateM (Talk) 00:54, 23 June 2015 (UTC)

Here you go. http://www.anandtech.com/show/9388/comparing-snapdragon-810-v2-and-v21 https://forums.oneplus.net/threads/oneplus-2-cpu-qualcomm-snapdragon-810-v2-1.316786/ https://forums.oneplus.net/threads/oneplus-2-rumors-we-are-mad.316975/ Theoriginalandrewia


 * I read through your draft and it is a big improvement over the current article. I am not an expert on Snapdragon, but but article read well. I think your incorporation of Theoriginalandrewia's prose on the overheating issue was of due weight and better balances the article. From a technical perspective, I think one could usefully include some of the material from the Architecture section of the current article into your draft. There is good information there, although I haven't checked if the sources are reliable, etc. I'd support bringing your draft over to replace the current article. --Mark viking (talk) 04:36, 23 June 2015 (UTC)


 * I went through the architecture section and added one tid-bit, but most of that was primary sources.


 * I added the AnAndTech piece. For OnePlus, Wikipedia requires secondary sources that are independent from the events. So we couldn't use OnePlus itself as a source for OnePlus using Snapdragon. If someone like AnAndTech did some benchmarking and published an article with several hundred words devoted to the Snapdragon in that product, then I think we'd have a Wiki-compliant addition. CorporateM (Talk) 16:47, 23 June 2015 (UTC)


 * As with some of your previous work, it takes some effort to review because of your start-from-scratch approach. I would prefer if we could make these improvements incrementally. I agree that the prose-based approach you've taken is superior to the current article but in starting from scratch, you've left out a lot of the information that is in the current article:
 * Original model numbers for pre-800 parts
 * Semiconductor technology
 * CPU instruction set
 * CPU cache
 * Wireless radio technologies
 * Memory technology
 * Availability dates
 * Utilizing devices
 * Snapdragon S1, S2, S3 and S4 families
 * Summary description of capabilities and architecture of each series
 * Hexagon DSP capability
 * Discussion of similar (competing) platforms
 * You are asking us to replace 7 years of work by over 100 presumably independent editors with something put together in a couple days by a single editor with financial ties to the subject. I'm not ready to agree with this. ~Kvng (talk) 14:48, 25 June 2015 (UTC)


 * Hi . I don't mind going through it section by section or step by step if you're willing to vet it that way with me. My first suggestion would be to trim almost the entire article though, because most of the page is piles of specifications tables that are un-encyclopedic. The Architecture, History and Lead sections are mostly unsourced or primary sources. There's only a paragraph or two of encyclopedic material cited to secondary sources in the current article. I disagree with your depiction, because it always takes the concentrated effort of a single editor to make highly-ranked pages. It's very rare to get GA/FA quality pages out of 100 editors over seven years. However, it's always better to discuss COI-created draft material thoroughly and have multiple editors editing it, if there are editors willing to spend the time. CorporateM (Talk) 16:15, 25 June 2015 (UTC)


 * First off, you have to be a bit more specific when you make an WP:UNENCYCLOPEDIC accusation. As to how to proceed, I suggest you can either make further improvement to your draft until we have consensus that wholesale replacement of the existing article with your draft improves things, or you can start improving the existing draft incrementally and we'll use WP:BRD to work through it. Because of the COI issue, if you choose the incremental approach, changes should be smallish and slowish to allow other editors time to review. ~Kvng (talk) 16:37, 25 June 2015 (UTC)
 * I wouldn't mind doing that and that approach would be more aligned with Wikipedia's principles, but it's not currently allowed by WP:COI. The reason community consensus is that COI editors not edit in article-space is because of the many cases where they have made even small changes that manipulate the text to support their agenda and those changes either go un-noticed or are too tiresome for other editors to revert. The drawback of current policy however is that it's not practical to make incremental improvements from Talk, which is the concern you're raising. Lets see what other editors say, if anything. I'm happy to contribute in whatever manner the community is most comfortable with, but what one editor is most comfortable with another will scream foul over. Sorry for being so difficult!  CorporateM (Talk) 17:03, 25 June 2015 (UTC)
 * While some editors prefer an incremental approach, I prefer the complete draft version approach (of course the two are not mutually exclusive: CorporateM can easily post a request to change X to Y in the article talk page and at the same time change X to Y on his draft version in his userspace).


 * What I do when I evaluate the draft:
 * Put the old and the new up on my screen side-by-side.
 * Carefully read the new, looking for any sort of bias or unsupported dubious claims. CorporateM, of course, knows that my continued cooperation requires that he do his best to write an WP:NPOV article.
 * Carefully read the old, looking for well-sourced negative material that satisfies WP:WEIGHT that might have disappeared in the new.
 * Read all the talk page discussion to see what issues other editors have.
 * If the draft needs any changes to make me happy with it, we discuss that, and then finally I cut-and-paste replace the current article with the draft, making a point of saying that I take full responsibility for all edits made in my own name.
 * Finally, I stick around for a while and help with any cleanup or changing any sections that other editors have a problem with.


 * I think there is room on Wikipedia for both approaches. I am about to start evaluating the draft and the existing page. Any comments about someone not liking something in the draft would be most helpful.


 * Full disclosure: I have no financial or other ties with any product using snapdragon or with CorporateM. I do a lot of engineering work with ARM processors, but haven't had the occasion to use this particular one. Guy Macon is the real, legal name that I was born with and I am listed at the Wikimedia Identification noticeboard as haven proven my identity.. --Guy Macon (talk) 19:22, 27 June 2015 (UTC)


 * Process aside, I did compare the two articles side-by-side and was not convinced that there was an improvement in the from-scratch version. I enumerated specific concerns in my reply above. Please ping me if improvements are made to the draft you'd like me to take another look. ~Kvng (talk) 15:11, 28 June 2015 (UTC)


 * I am not comfortable with the idea of me doing any cut-and-paste replace until either Kvng is happy with the replacement or someone makes a compelling argument that what Kvng wants is bad for the article (and I have seen no evidence of that). I do like to help paid COI editors who follow the rules in order to encourage that sort of behavior in others, but this one doesn't seem ready. I am going to wait and see while the proposed changes are discussed. --Guy Macon (talk) 20:43, 28 June 2015 (UTC)

Kvng pointed out that he felt a lot of the material in the current article was missing from the draft. I pointed that almost all the content in the current article is original research, primary sources, and an indiscriminate collection of specifications data. It is not missing - it's just stuff that does not need to be preserved. I think the natural next step in the discussion would be for me to ask Kvng to be more specific about what content that is both sourced and significant from the current article is missing, or for me to start proposing specific edits, rather than a draft. As previously discussed, my first edit to suggest would be to stub the article of all the original research, primary sources, and excessive specs tables. Normally the first thing I do when I come across an article like this as a volunteer is some cleanup of that kind of material. CorporateM (Talk) 21:03, 28 June 2015 (UTC)


 * In my opinion, the long section containing tables full of different models is a bit overwhelming for an encyclopedia article, and yet the information is not something I would want to see lost, so I have created List of Qualcomm Snapdragon devices. This should allow us to trim down the list of models in this article to give just a quick overview, with a "Main article:" link to the actual list. --Guy Macon (talk) 21:09, 28 June 2015 (UTC)


 * I dislike list articles personally, but there is community support for them, so that seems like the right approach. CorporateM (Talk) 21:23, 28 June 2015 (UTC)


 * Thanks I have removed the material from this article. ~Kvng (talk) 13:26, 30 June 2015 (UTC)


 * I have fixed broken section links that resulted from this split. ~Kvng (talk) 14:53, 30 June 2015 (UTC)


 * Thanks guys! There are a lot of editors constantly updating the technical specifications, so I'll wait for the dust to settle to see if there is any outcry before suggesting more changes for discussion. The next most prominent thing that jumps out at me are the product lists. CorporateM (Talk) 15:27, 30 June 2015 (UTC)


 * It looks like an improvement to me, so I am implementing the edit request. As always, I take full responsibility for any edits I make, even if I am doing them on behalf of a COI editor. Also, I have no COI here; the processors I use are on the extreme low end. --Guy Macon (talk) 15:44, 29 July 2015 (UTC)

More cleaning up
The current article has a "Similar Platforms" and a "See Also" section that appear to be primarily made up of links to competing products. When the source material contains a critical comparison between a product and others, sometimes this is included in a Reception or Comparison of section, but I believe lists like this are usually seen as linkbait and plugs. They are often created by vendors that each add themselves to a list. I do think it is sensible to have a See Also section to the articles devoted to the Snapdragon components:
 * Scorpion (CPU)
 * Adreno (GPU)
 * Krait (CPU)
 * Qualcomm Hexagon (DSP)

However, I'd like to request we trim the plugs/product lists.

CorporateM (Talk) 16:08, 9 July 2015 (UTC)


 * It looks like most of the links in the existing See also section could be moved to Similar platforms. The new links you suggest are or should be already linked from the article body. I agree that the Similar platforms section is linkbait and one could argue that the function of this topic is already covered by ARM-based chips. Quite fine with me to have no See also if that's the way it works out. ~Kvng (talk) 14:25, 12 July 2015 (UTC)
 * That's sensible.  CorporateM (Talk) 17:17, 12 July 2015 (UTC)
 * I thought we were going to remove the linkbait entirely as oppose to putting it in the Similar Platforms section? On a side-note, I do see a couple Comparison of articles where Snapdragon is included, but these are not. That would be a much more appropriate place for this kind of list, but I don't have the technical knowledge to do this. Maybe will be able to offer some guidance on whether these are relevant for the vendors listed under "Similar Platforms". CorporateM (Talk) 17:15, 15 July 2015 (UTC)
 * I did a little work to make sure everything was covered in the templates and have now removed the Similar platforms section. The two comparison articles you call out are also linked from the templates. ~Kvng (talk) 18:02, 15 July 2015 (UTC)

Draft revisited
indicated here that his concerns have been primarily addressed by the creation of a separate article List of Qualcomm Snapdragon devices. By his request, I've also culled through the current article to make sure the proposed draft doesn't over-ride properly sourced material. I also did a few quick searches to see if poorly sourced material could be better-sourced. I got the impression (please correct me if I'm wrong) that with that discussion concluded, it was now a good time to re-visit incorporating the draft.

There is still quite a bit of material in the current article not included in the draft, but much of it is trivial and promotional information such as demos Qualcomm did at conferences, original research I could not find sources for, etc. At the very least I think the draft would be a good starting place for GA review and/or ongoing incremental improvement. CorporateM (Talk) 18:12, 16 July 2015 (UTC)


 * It looks like an improvement to me, so I am implementing the edit request. As always, I take full responsibility for any edits I make, even if I am doing them on behalf of a COI editor. Also, I have no COI here; the processors I use are on the extreme low end.


 * I couldn't find the right file for the image titled "A Snapdragon SoC from Qualcomm" so I commented it out. CorporateM, could you take a look at it? --Guy Macon (talk) 15:57, 29 July 2015 (UTC)

Please stop re-adding or modifying the Specifications section
Please stop restoring or editing the Specifications section. It has been removed because it is: Thank you. sovereign°sentinel (contribs) 04:26, 14 August 2015 (UTC)
 * not stable (often edited by IPs, sometimes incorrectly)
 * only relies on non-independent sources
 * only contains information about current or recent products
 * is redundant to content in List of Qualcomm Snapdragon devices

Recent edit
Someone recently added benchmark tests that are cited to a dylricho.com URL. dylricho.com is self-described as "also my personal portfolio!". Suggest trimming/reverting due to inappropriate sourcing. CorporateM (Talk) 16:03, 9 September 2015 (UTC)


 * Do you think I can bother you with this? David King, Ethical Wiki (COI) (Talk) 19:20, 5 October 2015 (UTC)

had a 50% market share of a $7.9 billion
"had a 50% market share of a $7.9 billion smartphone processor market."

This is the source cited: http://www.forbes.com/sites/briancaulfield/2012/07/18/no-factories-no-phones-no-fuss-how-qualcomm-grabs-wireless-profits/#68ffb28353b1

Couldn't find the information in the source. Uziel302 (talk) 12:33, 9 May 2016 (UTC)


 * I am affiliated with Qualcomm. The support for that sentence can be found on page 2, near the end of paragraph 7, about halfway down the page, where it says: "Qualcomm grabbed 50% of the $7.9 billion smartphone processor market in 2011." (link). It's easy to miss if you didn't notice there were multiple pages. Cheers. CorporateM (Talk) 17:29, 9 May 2016 (UTC)
 * Thanks. Uziel302 (talk) 19:04, 24 May 2016 (UTC)

Updates
Hi. I am affiliated with Qualcomm and a heavy contributor to this page, having brought it up to GA status a couple years ago. A few things have changed since then:
 * The 200 series is no longer associated with the Snapdragon brand
 * There is a new Snapdragon logo
 * The infobox keeps getting removed for some reason
 * Some poor language has been added in recent edits like "rush out," "ditching," and "solutions"
 * Several new products like the 700 series have been added to the Snapdragon family

I have placed numerous proposed edits here with bold indicating suggested new content, cross-outs for trims, and a couple notes in bold red text about moving content about the 200 series to history. As an interested editor with a conflict of interest, I wanted to solicit any feedback from disinterested editors. CorporateM (Talk) 16:04, 26 March 2018 (UTC)


 * CorporateM, I don't see anything controversial about the changes, so I am going to go ahead and make them. --Guy Macon (talk) 18:05, 17 April 2018 (UTC)


 * I didn't change the logo. File:NewSnapdragonLogo.jpg doesn't exist, and File:Qualcomm Snapdragon logo.png is lacking a detailed fair use rationale copyright information, and whether it is low resolution. --Guy Macon (talk) 18:19, 17 April 2018 (UTC)


 * The reference at "Sun, Leo (29 March 2015). Will Intel Corporation's Curie Conquer the Internet of Things and Wearables Markets?. Fox Business. Retrieved 14 May 2015." returns a 404 not found. it can be found at lots of other sites such as . but I would prefer a link to Fox. --Guy Macon (talk) 18:37, 17 April 2018 (UTC)


 * The claim "By 2015, Snapdragon was used in most non-Apple smartphones." needs a citation. --Guy Macon (talk) 19:29, 17 April 2018 (UTC)


 * Eight dead links total, and one subscription required (if we can find a non-paywalled version, use that). --Guy Macon (talk) 20:36, 17 April 2018 (UTC)


 * Thanks . I uploaded the new logo image and updated the infobox with it (since logos aren't allowed in draft-space) and replaced the dead URLs with live ones wherever possible. I also deleted the uncited comment about most non-apple smartphones, which apparently has been on the page for years (originally with cites that did not support it). CorporateM (Talk) 22:43, 17 April 2018 (UTC)

Semi-protected edit request on 29 April 2018
223.24.173.16 (talk) 10:50, 29 April 2018 (UTC)
 * Red question icon with gradient background.svg Not done: it's not clear what changes you want to be made. Please mention the specific changes in a "change X to Y" format and provide a reliable source if appropriate. Nici  Vampire  Heart  14:35, 29 April 2018 (UTC)

Arbitrary markety milestone with no justifying cite
It included the first 1 GHz processor for mobile phones.

Intel broke all frequency records for years and years with their power-profligate Pentium IV, which sure look good until it


 * Intel to 'ditch' Pentium 4 core after Prescott — May 2004

Intel will today tear its Pentium 4 and Xeon roadmaps to shreds and announce the cancellation of its 'Tejas' and 'Jayhawk' processors and their successors. Both chips represent what was to have been the next generation of 90nm Pentium 4 and Xeon, respectively, and were scheduled to arrive in the Q2 2005 timeframe.


 * Is Intel's Prescott P4 too hot to handle? — July 2004

We should all know by now that Prescott, Intel's 90 nanometre desktop processor, did not deliver the clock-for-clock performance gains that Northwood did over Willamette. However, because of its much longer [ed: more wasteful] pipeline, Intel's newest core should scale much higher in frequency than its predecessor [ed: but not without first bursting into flames].

The reviewer wasn't too charitable when he said: "I don't think I need to comment on these numbers. Prescott processors warm up much more during active work than their predecessors. Note that we measured the CPU performance during the tests carried out in an open testbed. I am scared to imagine what happens to Prescott when we close the system case ..."

My editorial comments.

This was entirely self-inflicted by Intel in a marketing push to crow about giant GHz numbers—and it worked at the sales level for a while—but it ultimately lead to their competitor's best ever era (the glory days of the AMD Opteron).

So we definitely need a cite here that Qualcomm crossed the 1 GHz barrier as meaningful industry progress, and it wasn't merely achieved by lighting their own pants on fire, to win today and lose tomorrow.

(Yes, the Register is the least reliable source, but Intel had everyone else cowed into submission, which was a big part of the whole problem.) &mdash; MaxEnt 19:16, 11 September 2018 (UTC)

Benchmark tests
This section talks about the underperformance of the infamous SD810, then finishes with "Additionally, the 835 and 845 that followed performed substantially better." Duh?! What followed the 810 (circa 2015, I think) wasn't these; rather the 820/821 in 2016 and yes indeed, they performed better because they were based on one generation newer ARM cores, and the overheating GPU issues weren't apparent on these chips. We should fix this glib passage. Actually, we're just plain short on benchmarking comparisons between families as well as chips within a family. The GPU's changed as well. Sbalfour (talk) 23:55, 10 November 2018 (UTC)

Subject
Subject Jack40125 (talk) 00:01, 13 July 2019 (UTC)

Requested edits
I would like to trim an unsourced and promotional sentence regarding the performance of a Snapdragon semiconductor and replace it with sourced content cited to more recent independent benchmark tests in PC World/AnAndTech on the Snapdragon 865. The changes I am requesting are indicated in strikeout and bold below. CorporateM (Talk) 14:32, 9 October 2020 (UTC)

Additionally, the 820/821, 835 and 845 performed substantially better. Qualcomm led the industry in sustained GPU performance and perf/W from the release of the 820 until the release of the Apple A12. '''The Snapdragon 865's memory was improved in a later update. A 2019 benchmark test by PC World found that the 865's multi-core performance on the "default" setting was 30 percent higher than the 855 and comparable to the 855 Plus. '''

References


 * Seems like a good improvement. Done. --Guy Macon (talk) 04:59, 21 October 2020 (UTC)

888
Please add some content around the 888. Cveld (talk) 21:27, 6 April 2021 (UTC)

CPU Infobox
I just changed the infobox from the "company" template to the "CPU" template. I think this makes much more sense, but I'm not 100% sure on how I should organize everything. For example, listing the DSP, ISP, and modem under the category "co-processor" doesn't make perfect sense, but it was the best that I could find. There's also quite a bit more information we could add with the new infobox, if I or anyone else gets the time to add it. If anyone has any ideas on organizing the infobox or more information, let me know. RealSwiftCoderJoe (talk) 01:55, 2 March 2023 (UTC)


 * I work for Qualcomm. You have a good point that the company infobox isn't quite right here. There isn't SoC infobox, so this might be the next best thing. That being said, I suggest some small tweaks (see draft below). This would move stuff from "co-processors" into a more appropriate label "Cores" (since they aren't all processors) and updates the latest brand names (Qualcomm now uses Snapdragon 7, instead of Snapdragon 700), among other tweaks. I also want to update the logo image in the infobox, but I guess my account has to be 4 days old first. Hayleyolson (talk) 17:21, 23 March 2023 (UTC)
 * I have made the changes you requested to the Infobox, and I think they make much more sense. Feel free to update the logo image when you have the privileges to do so, but make sure you don't violate Non-free content! I have no idea how a logo may or may not fall under free use. RealSwiftCoderJoe (talk) 19:18, 26 March 2023 (UTC)


 * Thanks!! I just uploaded the new logo using the fair use rationale for logos. Hayleyolson (talk) 16:22, 13 April 2023 (UTC)