Talk:Intel Atom

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Computing / Hardware (Rated C-class, Mid-importance)
WikiProject icon This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
C-Class article C  This article has been rated as C-Class on the project's quality scale.
 Mid  This article has been rated as Mid-importance on the project's importance scale.
Taskforce icon
This article is supported by Computer hardware task force (marked as Mid-importance).

Intel Atom vs Intel Atom Centrino[edit] anyone know more details about the Centrino, or are they just the same thing? NeoDeGenero (talk) 13:59, 18 June 2008 (UTC)

I've not read up on this sort of technology for a few years now but the last time I checked the difference between an Atom and a Centrino Atom solution would mean that the the Atom solution is an Atom CPU with a third party chipset, while a Centrino solution is an Atom CPU coupled with an Intel designed chipset. Basically Centrino is a badge that gets awarded to computers with all intel components (northbridge, southbridge etc).
In fact, to quote the address you posted:

...Intel® Centrino® Atom™ processor technology, a collection of chips enabling amazing Internet experiences in pocketable devices.

--Benchamoneh (talk) 10:40, 12 August 2008 (UTC)


"A 1.8 GHz Atom processor's single thread performance is equivalent to its predecessor Intel A100 " - how's the performance of a 1.8GHz cpu possibly equivalent to a sub GHz cpu? I think some clarification might be needed. Anton (talk) 06:05, 26 April 2008 (UTC)

Atom no dual, but single core[edit]

Early tests showed Atom had 2 processors in Windows task manager.It was therefor believed the atom dualprocessor was a dual core version. However Atom Diamond ville is a single core processor! With hyperthreading enabled it may seem like there are 2 processors in taskmanager.

So far intel has said no word about releasing a Atom dualcore version anytime soon!

The Intel Atom 330 has a Dual Core processor. Here is a company that sells a board with the 330: the Intel 330 Dual Core Atom powered mini-itx board – D945GCLF2

-cheers H.E. Hall (talk) 15:27, 26 December 2008 (UTC)

Merge with MID[edit]

I think this is nonsense - you could merge Pentium into Personal computer like this.--Kozuch (talk) 19:05, 8 March 2008 (UTC)

Agreed. The Intel Atom is just one of many processors used in mobile internet devices... no reason for a merge. ǝɹʎℲxoɯ (contrib) 19:29, 8 March 2008 (UTC)
Yeah. Keep the articles separate. They are different things. -- Imperator3733 (talk) 20:12, 8 March 2008 (UTC)
Part of the problem is that the Mobile Internet Device page speaks not of mobile Internet devices, in the sense of mobile devices that can access the Internet, but about Intel's "Mobile Internet Device" platform. I guess that's why Talk:Mobile Internet Device asserts that "Intel is mentioned because they invented MIDs"; that page also says:
A Mobile Internet Device (MID) is an Ultra-Mobile PC (UMPC) initiative raised by Intel for consumers and prosumers. Most other UMPCs are designed for mobile professionals.
Unfortunately, "most" is not "all"; the iPhone, for example, weren't first targeted at mobile professionals. I don't know for whom Nokia designed the various Nokia Communicator devices, or whether the Nokia N800 - or, for that mater, the iPod touch - would not be considered sufficiently "mobile" as they support Wi-Fi (and, apparently, WiMAX in the case of the N800), but not any mobile phone Internet access mechanisms such as GPRS, EDGE, HSDPA, EV-DO, etc..
I'd argue that it might be useful to have a "mobile Internet device" page that discusses mobile devices that provide Internet access in general (and would therefore go back at least to the Nokia 9000 - which, amusingly enough, had an x86 processor in it, albeit one from AMD) - and a page about Intel's "Mobile Internet Device" reference design, which should probably be called "Intel Mobile Internet Device" to indicate that it's specifically about Intel's design. One could perhaps then argue that the latter page should be merged into this page, but one could perhaps also argue that the chips and the system are different entities and deserve separate pages. Guy Harris (talk) 20:17, 8 March 2008 (UTC)
"Intel Mobile Internet Device" and "Mobile Internet device" pages would really make sense to me.--Kozuch (talk) 21:02, 8 March 2008 (UTC)
I agree. It might be useful to have a generic page but the page about Intel specific MID platforms should not be merged in with pages that describe Intel MID processors (the A100 or the Intel Atom) like this one. If anything needs to be done about this my opinion is the MID page could stand to be renamed to Intel MID and a generic page created talking about MIDs in general beyond Intel MID platforms (and probably linking to the renamed Intel MID platform page). (talk) 16:03, 10 March 2008 (UTC)

I agree, keep them seperate. Colinstu (talk) 18:46, 10 March 2008 (UTC)

Merge templates deleted.--Kozuch (talk) 23:22, 10 March 2008 (UTC)

Atom NOT a CISC processor![edit]

Whoever wrote this page clearly knows nothing about x86. x86 is the INSTRUCTION SET. Internally since the Pentium Pro all x86 processors have been RISC, with decoders to turn CISC insructions into RISC operations. Thats really basic stuff. Read the excellent anadtech article. I'm on holiday, and do not have time to sort this nonsense out now. Sorry. (talk) 15:03, 6 April 2008 (UTC)

You are right, internally Atom is a RISC architecture based on micro-operations. On the other hand, you have to agree that IA-32 is an CISC instruction set. Thusly, Atom has to decode x86 instructions into micro-operations; a method that has remained standard to this day for x86 compatible processors. Enter ARM processors! They come with their own instruction set - which is RISC. So they do not need the translation step and save that die-space/power for more important things. Remember, backward compatibility always comes with some penalty. DanTheMan (talk) 19:46, 16 April 2008 (UTC)

That backward compatibiltiy while having some costs helps ensure that DOS 1 still boots on the fastest IA32 processor out there, so it's not all bad. ;-) Regards, Andromeda451

Background is confusing[edit]

The background section is confusing, since it comes immediately after the introduction, and tracks historical rumours (even ones which have proven incorrect). Some speculation is still written in present tense, like "This seems to strengthen speculation that Diamondville is simply a lower-cost[...]". It would be better to add more info before, and move the history (if it is at all needed). Say, one should explain that "there are these two processor lines, which have different codenames, Diamondville and Silverthorne, referring to the same architecture in different variations, one for a target of Diamondville, one for target of Silverthorne" (this is just to give an idea). --Blaisorblade (talk) 15:40, 10 August 2008 (UTC)

Platform power consumption way to high[edit]

So far the Intel Atom is using the i945G-chipset, which has several times the Atom's power consumption and thus makes it unattractive as a platform. I would like to see a mention of this fact here. What good is a 2Watt processor if it needs a 10+Watt chipset? —Preceding unsigned comment added by (talk) 13:19, 13 August 2008 (UTC)

Good point. I saw a motherboard based on the Atom that had one fan which was placed, not on the Atom, but on the i945G. Andries (talk) 19:35, 13 September 2008 (UTC)
I added a little bit with references I found in another article. - Taxman Talk 20:29, 3 October 2008 (UTC)
MSI has a motherboard (model MSI IM-945GSE (-A)) using the 945GSE chipset, which uses much less power than the 945G. A few other companies seem to be offering similar products in Europe. All seem to be using the N270 single-core Atom CPU. The MSI has dual gigE Intel NICs, making it a good choice for a linux-based firewall. --Bobbozzo (talk) 06:34, 10 April 2009 (UTC)


Is the Tegra really competition. Its a ARM part, with no IA32 compatability. ARM has always low power consumption devices and most likely much better W/cycle, MIPS/W and FLOPS/W etc? —Preceding unsigned comment added by (talk) 14:47, 4 September 2008 (UTC)

I have to agree that this bit of the article confused me. The Atom is popular because it is a low power x86 processor. Surely non x86 are not competition? If they are, don't we have to open up playing field to many many more processors? (talk) 02:24, 12 February 2009 (UTC)
The last sentence of this section states that the 945GSE is lower powered, while the caption to the picture states that the 945GSE consumes more power. Either one of the references is wrong and should maybe be "i945G"(?), or the power comparisons should be better specified. (I.e. lower/more power than what?) (talk) 02:04, 12 June 2009 (UTC)

Intel is making ARM a competitor by entering into the market space traditionally dominated by ARM. That being mobile devices, smartphones, and tablets. In this space, x86 compatibility means nothing and so isn't relevant. Intel has clearly stated it's intentions to push Atom in the mobile market. See here So, Intel themselves have made ARM / Tegra performance comparisons valid and relevant to this discussion. CG - Jan 2011 — Preceding unsigned comment added by (talk) 18:33, 13 January 2012 (UTC)


Currently this article is incredibly hard to read and follow caused by a few problems. Primary of which is there is very little overview information to provide context for the reader. Instead the article plunges into technical details with full jargon. There should be much more context to ease the reader in. I'm not really so up on the details of all this but I'll try to help where I can. This platform seems poised to be more and more important, particlularly if it keeps getting improved. - Taxman Talk 20:29, 3 October 2008 (UTC)

how does the performance compare?[edit]

How does the performance of the 1.6ghz atoms currently common in netbooks compare to the "Celeron M ULV 353" used in earlier netbooks? Plugwash (talk) 15:56, 31 January 2009 (UTC)

Intel graphics lockin in future versions?[edit]

nVidia slides on

indicate that the next version of Atom will require Intel graphics, locking out nVidia's Ion platform. —Preceding unsigned comment added by (talk) 21:36, 25 February 2009 (UTC)

There is no lock-in. There are, now, some n450 (pineview) based devices running third party graphics. It will either offload certain processing to the extra graphics or disable the integrated GMA on the CPU. While you are forced to have the GMA included, on devices where only the GMA is present, a significant power savings is seen over older generation netbooks. In fact, the netbook I, personally, have (with a pineview cpu) has been tested successfully with a broadcom addin card. (talk) 01:33, 26 March 2010 (UTC)


I've found two PNGs available for the Atom logo, but I didn't upload them because I'm not completely sure about their copyright :/ Here are the URLs:

 * (Big One!)

--;) Peregrino (talk) 14:35, 8 April 2009 (UTC)

Competition, ARM, etc[edit]

Ok guys, I made some major changes to the page and would appreciate a second look by those who are knowledgeable about this CPU and platform

- Moved ARM information to competition section

- Replaced reference to brand-specific Nvidia Tegra product in the "competition" section with information on ARM's competing platform which is used by Nvidia, Texas Instruments, Qualcomm, Samsung, etc.

- Added information on competition between future Atom platform "Lincroft" and next-gen ARM Cortex-A9 based CPUs

- reworded some other sections —Preceding unsigned comment added by Winterspan (talkcontribs) 02:09, 18 May 2009 (UTC)

Hello! I just added the flag [dubious ] about the ARM performances because there are two mistakes: 1) actually it's really hard to find a comparison apple-to-apple; if you know one, just link the reference 2) ARM and Atom are really different architectures, so similar results with syntetic benchmarks could mean nothing: old Pentium 4 Prescott has similar integer performance of POWER 5 processors, but nobody thinks that you can use a Pentium 4 instead of POWER5 in mainframe applications

I know that it's a widespread belief that Cortex A8 will compete with Atom but actually there is no evidence of this, so it's more a commonplace (for someone a hope, I think) than a real fact.

Thank you! =)

Lenovo skylight review provides some information on power consumption. I have been using ARM processors in the embedded space for over 10 years and they provide better MIPS per watt than any other processor. Take a look at coremark benchmarks for the ARM processor a Cortex-A9 scores 11.5, a intel-i5 is 7.9 for comparison. The total system design including memory bus, video, etc all play a role, so you should not take any benchmark as gospel. However this Dubious tagging smacks of FUD to me. I am not sure what a conclusive reference would be to make you remove this tag? —Preceding unsigned comment added by (talk) 17:34, 8 January 2011 (UTC)
How can you compare MIPS (or MIPS per Watt) when they're different ISAs/architectures? (talk) 23:20, 22 January 2011 (UTC)
Could you elaborate? Why isn't it possible? The comparisons of performance-per-watt on different architectures is indeed the point of these benchmarks. 1exec1 (talk) 15:24, 23 January 2011 (UTC)
You cannot compare MIPS between different ISAs, especially between RISC and CISC architectures. 02:14, 17 December 2011 (UTC)
Explained in layman terms: RISC means reduced instruction set, so there are less instructions and they're more basic. You can't directly compare MIPS (millions of instructions per second) between ARM and Intel because, in general, ARM is going to take more instructions to do the same work as Intel. What you can compare are benchmarks for the same piece of software, across different type of software. In any case, until very recently, Atom and ARM were practically not comparable, because they were in entirely different price, performance and power consumption points (there was no overlap, you couldn't compare for example power consumption for the same price/performance because no Intel processor was so cheap as any ARM processor and no ARM processor had such high performance as any Atom). Now ARM is increasing performance, mainly through multi-core processors, and Intel is looking at reducing power consumption and price in order to enter the same mobile/tablet market, so comparisons are starting to make sense. In any case, as I said before, MIPS are not directly comparable. A 1GHz ARM with ARMv7 instruction set might be similar to a 450MHz Pentium III or something like that in performance, and we're talking about integer performance. An typical ARM11 at 700Mhz with ARMv6 instruction set could be similar to a 266Mhz Pentium II. If we get into floating point or SIMD (SSE for intel, Neon for ARM), the differences are even much more in favor of Intel. On the other hand, Intel Atom processors still consume considerable amounts of power, more than ARM processors. — Preceding unsigned comment added by (talk) 08:04, 22 August 2012 (UTC)
Now that Intel is targeting tablets and has released Android Oak Trail (why they are still "prototyping" Oak Trail devices while simultaneously releasing Ivy Bridge devices to the press is beyond me) Atom tablets to the press, those android benchmarks are probably pretty relevant in comparison. So maybe you can include Quadrant and CoreMark results? Quadrant is supposed to compare whole subsystems accurately (GPU, memory, CPU - but single threaded), coremark specifically benchmarks the cores (multithreaded). PS> just to tempt you to look the numbers up, stock (no overclock) Cortex A9s quadrant benchmarks span about 1250-3050, the Oak Trail Atoms score ~1950. --— robbie page talk 13:56, 5 June 2011 (UTC)
The first Intel Atom processors to be included in mobile phones have a 3489 quadrant score: Quoting from that article: "happened to have Quadrant installed on it, so we launched it up and discovered that the Intel Atom Z2460 chip on the Lenovo K800 scored a 3489. Naturally, we’d like to stress that benchmark numbers don’t necessarily reflect real-life use scenarios, but they do give us an idea of what kind of performance we can expect from a handset. Since Quadrant is only optimized for single-core devices, we’re not surprised at all to see that the Intel Atom Z2460 scores a lot higher than the quad-core NVIDIA Tegra 3 powered devices." And this is still a 2012, 32nm process Atom processor (so I guess practically a hack using existing Atom designs). — Preceding unsigned comment added by (talk) 08:17, 22 August 2012 (UTC)


LPIA links here, but is not explained. --Nomeata (talk) 20:57, 30 June 2009 (UTC)

Low Power Intel Architecture ? —Preceding unsigned comment added by (talk) 23:50, 21 December 2010 (UTC)


The bit about availabilty "Atom processors are not yet available to home users or system builders, although they may be obtained preinstalled on some ITX motherboards." the cpu's have been availabe to home builders from the, there are dozens of montherboards availabe, and have been for at least a year.

Clarified as "not available to home users or system builders as separate processors" (talk) 03:19, 6 September 2009 (UTC)

Yakacm (talk) 11:20, 13 July 2009 (UTC)

Atom + NVidia 9400M = ION is incorrect[edit]

Mainboards with an NVidia chipset and an onboard NVidia 9400 video card are marketed as ION - but that's not limited to Intel Atom CPUs, NVidia announced it for other processors, including the VIA Nano -- (talk) 22:25, 4 August 2009 (UTC)

New unlisted cpu: Intel(R) Atom(TM) CPU K510 @ 1.66GHz[edit]

This CPU was seen in the field but is not listed on Intel's Spec Finder site:

cpu family      : 6
model           : 28
model name      : Intel(R) Atom(TM) CPU K510   @ 1.66GHz
stepping        : 10
cpu MHz         : 1662.309
cache size      : 512 KB
Cores: 2
Hyperthreaded: yes

I guess this "K" series is very new and should be added to the page, but I cannot find any public docs on it yet. —Preceding unsigned comment added by Benefros (talkcontribs) 16:44, 1 April 2010 (UTC)

Some reports are also saying this is a BIOS/firmware bug that is misreporting a D510 as a K510. If that is true, it would explain why you cannot find any details on it. That said, Intel has not been very forthcoming with the Lincroft-based Z6xx line either. (talk) 12:35, 19 August 2010 (UTC)

Socket? Self-install / upgrade capability?[edit]

Many other Intel processors such as the Celeron, Pentium, Core i and Xeon series can be installed and upgraded in-situ, by an unskilled person, without soldering, using a motherboard socket or slot. This article implies that the Atom cannot be used in this manner, that it requires soldering directly to the motherboard and cannot be upgraded without electronics skills, but the article could be improved by a knowledgeable person making this directly clear (or, if untrue, explaining how unskilled upgrades can be performed). Andrew Oakley (talk) 09:56, 12 July 2010 (UTC)

it does say it, it also shows the chip on a board. also not all of the others can be upgraded, a lot of others in subnotebooks and macs in particular use pentium celeron or core are also bga not pga chocobogamer mine 10:39, 12 July 2010 (UTC)

Atom D525 Graphics[edit]

Is there any knows limit to the pixel resolution of the internal Atom D525 Graphics? I can not find graphics specs at Intel's processor site. Would be interesting to know if the resolution is limited to 1920 x 1080 and 2048 x 1152 screens can not be used in native mode. —Preceding unsigned comment added by (talk) 04:09, 24 August 2010 (UTC)

Are there Linux drivers for PowerVR graphic core (GMA 500 + GMA 600)?[edit]

PowerVR -> "Intel uses the SGX 535 as its GMA 500 and GMA 600 integrated graphics for their Atom platform"

As far as I am informed, there are only some very crappy Linux drivers for this hardware. The manufacturer himself Imagination Technologies doesn't care for Linux at all. Simply google for linux driver PowerVR or for SGX 535 linux drivers

The Linux drivers by intel are reported to be quite good, but how mature are the drivers for the PowerVR?

Missing Atom Serie[edit]

E6xx -

full Atom info - — Preceding unsigned comment added by (talk) 13:36, 24 October 2011 (UTC)

for Atom D2500 was specified incorrect GPU speed, please edit[edit]

d2500 has 400mhz gpu according to new d2550 may has 640mhz gpu — Preceding unsigned comment added by (talk) 13:03, 22 February 2012 (UTC)

Done. Feel free to edit it yourself next time though. --Manguene (talk) 19:17, 2 April 2012 (UTC)

x86-64 instruction set[edit]

The information in the section about 64bit support is outdated and probably refers only to older models. I have a netbook with N2600 Cedarview (MSI U180) and 64bit Ubuntu runs fine.

Intel Atom reverts[edit]

Please explain why you seem to believe that erasing valuable and difficult to obtain information is useful to Wikipedia and humanity at large? Is it that you simply can't be BOTHERED checking before you erase? If so - retire from Wikipedia. Seriously - let the work be done by others that will put more effort into researching changes rather than simply erasing other's work

  1. You are NOT doing a "public service" by simply reverting pages arbitrarily
  2. Without either fact-checking or providing any written (eg talk) justification for doing so - you're simply undoing other people's work for no useful purpose
  3. I'm curious as to your motivations: Why are you doing this? - what is your self-justification for your actions? Is it just so that you can be a hipster and state "I edit pages for Wikipedia"? Perhaps you should instead simply mention that you erase other people's work - rather than actively contributing new information.
  4. Do a higher quality job. Provide justifications / rule breaches as part of your process of reverting pages.
  5. How much damage have you caused by doing this? How many pages have you "reverted" after work by other people?

In the specific case of the Intel Atom CPU:

  1. The information I have supplied is accurate
  2. There is nothing (that I can see) in the information that is in breach of Wikipedia rules

Please explain why you reverted my changes to this page. — Preceding unsigned comment added by Pragmatool‎ (talkcontribs)

It is up to you to justify adding unsourced material, not for others the check everything added. Please see WP:RS and WP:OR. You are welcome to add properly cited and sourced information. HumphreyW (talk) 23:35, 27 August 2012 (UTC)

Let me understand this - you seem to be indicating that you will revert any page that includes unsourced material? Is this your justification for reverting my changes? — Preceding unsigned comment added by Pragmatool‎ (talkcontribs)

It is Wikipedia policy to not allow unsourced material. Please read the policies. The standard for inclusion in Wikipedia is verifiability, not truth. HumphreyW (talk) 00:22, 28 August 2012 (UTC)

I'd appreciate answers to the following. These are important questions - and directly lead to an understanding of your function at Wikipedia.

  1. Is your "process" or function at Wikipedia to view recently updated pages and instantly revert them if they do not provide citations? Note that you appear to be doing so without any explanation or justification of your action.
  2. Please explain the significant number of pages that are marked as "citation needed" for content.
  3. I have a recently purchased N2600 Atom powered laptop. Intel states that an N2600 supports 64 bit (my reference on the page). However - it does not support 64 bit OS'es. The BIOS indicates that "EMT64" is "not available". There is no way through the BIOS to turn it on. Other websites have indicated that this is because Intel have disabled the 64 bit support. I have requested an answer from Intel. Clearly Intel is lying on their main website by not indicating that some CPUs will have their functionality limited in this manner.

So: until this is resolved: LET THE INFORMATION STAND on the basis of:

You still need to provide a WP:RS. It makes no difference how true or accurate your experience is because Wikipedia policies simply do not allow for inclusion of unsourcd material. If you feel this is wrong then please argue the policies enacted within wikipdeia on the policy pages. HumphreyW (talk) 02:35, 28 August 2012 (UTC)

Humphrey: please answer the preceding questions that you ignored. These are not difficult - and there are only two of them. For simplicity I have reproduced them here:

  1. Is your "process" or function at Wikipedia to view recently updated pages and instantly revert them if they do not provide citations? Note that you appear to be doing so without any explanation or justification of your action.
  2. Please explain the significant number of pages that are marked as "citation needed" for content.

I also add the following two questions to challenge your judgement of "reliable"

  1. Indicate why a deep expert exchange forum is not "reliable" - with a screenshot of the issue - especially when in concert with my own direct experiences.
  2. Alternatively - Please provide a list of "reliable sources" or some way to objectively measure the "reliability" of a source. I believe your process thus far to have been heavy-handed, destructive and extremely subjective. I would like to introduce some objectivity into an analysis of your actions. — Preceding unsigned comment added by Pragmatool‎ (talkcontribs)
Reliable sources are as described in the policy: WP:RS (please take the time to read these policy links, they are important to help you understand how Wikipedia works).
Screen shots provided by random people on the Internet are not reliable sources. For example there could be many reasons for the error. The user could have misconfigured the BIOS setting. The manufacturer could have installed the wrong BIOS. The poster could simply be faking. Or the whole thing could indeed be true and accurate as you say. But the whole point is there is no way to judge that based upon unreliable sources.
I will give you a few days to find a reliable source for your claim(s). If there is no forthcoming reliable sources (as per the WP:RS guidelines) then there is not really an option here, the claim, and any listed non-RS references, must be removed.
Also please sign your talk page postings with four tildes, like this: ~~~~. Thank you. HumphreyW (talk) 05:48, 28 August 2012 (UTC)

I have asked you (twice) to answer two relatively simple questions. Both times you have ignored my request - without any explanation. Indeed - much of your "work" appears to be done with very little to no explanation other than some blithe references to global Wikipedia policy.

ANSWER THE QUESTION(s) below - from there I'm happy to proceed based on your answers. These are reasonable questions directly related to the function that you appear to be trying to perform in relation to the Intel Atom web page.

  1. Is your "process" or function at Wikipedia to view recently updated pages and instantly revert them if they do not provide citations? Note that you appear to be doing so without any explanation or justification of your action.
  2. Please explain the significant number of pages that are marked as "citation needed" for content.
  3. You appear to be indicating that you claim dominion over this page and its ownership. You're indicating that you own the pages. That they're "yours" to determine their final status. Do you claim ownership of the page?
  4. When you say "I will give you a few days" - you're indicating that you will (again) remove my entirely reasonable updates. Is this your intention?

As mentioned - there are at LEAST two people (myself and another) having an issue with the Atom processor and it's lack of 64 bit support. You appear more than happy to ignore clear screen shots of "no 64 bit support" - which you can see with your OWN EYES. Yet - you still want to remove the information as "not having a reliable source".

I doubt that you'll find Intel, Asus nor other laptop manufacturers will want to admit that they're disabling it in the BIOS (or removing it on the Atom CPU). I have asked them, and will continue to press Intel and Asus for answers - but it'll take longer than "a few days".

Be aware - I'm now downloading Wikipedia - and all of its page delta log information.

I want to determine just how much damage you've been doing by unwinding other people's updates. Pragmatool‎ (talk) 04:33, 29 August 2012 (UTC)

HumphreyW is only doing his best, so please have good faith in him. I will say, as someone who works in this content area, that specific models of the Atom do have 64-bit support missing. It is too much to say that the whole series has no support on basis of a single model, and nor can you say that Atom has 100% 64-bit support either.--Jasper Deng (talk) 04:37, 29 August 2012 (UTC)
My current research indicates that it's possibly not the chips - but the vendors (Asus et al) that are permanently switching 64 bit off. Regardless of the mechanism - some Atom powered laptops are only 32 bit, yet Intel's website indicates "64 bit support". Clearly this is misleading to the public for Wikipedia to simply follow the Intel corporate line.
My questions related to HumphreyW's activities stand. Humphrey has refused to answer questions related to his activities. The erasures of content based on purely on whether they have "verifiable" sources or not is not necessarily a positive action for Wikipedia as a whole. As evidence of the negative impact I put forward the following:
  • Humprey's talk page: Samba erasures - "You are reverting my efforts to clean up the samba school article. I am trying to clean up this article because it is merely copied from the portuguese Wikipedia article and at least 70% of it completely meaningless.". While I can't find the original articles - clearly this user was frustrated at Humphrey's arbitrary deletion of what I believe to have been entirely factual and useful information.
  • Intel Atom - 64 bit erasures; no justification of action - simply remove the new information; a 'net loss' of factual data.

I am presently analysing the Wikipedia-wide impact of Humphrey's erasures through analysis of this page history dump:

Frankly: my "faith" in his "goodness" is broken. I'm questioning these activities under the microscope of "has the nett reliable information published by Wikipedia been increased or decreased by Humphrey's actions?"

Pragmatool‎ (talk) 06:36, 29 August 2012 (UTC)

For the record: here's the Intel statement on 64/32 bit. Asus are yet to respond. This is as close as I've come to a 'verifiable source' at this point - and it's hopelessly wishy washy.

We have received your request. From here, we can confirm that the instruction set for the Intel® Atom™ Processor N2600 is 64-bit. However, please keep in mind that the compatibility with the Operating Systems will be also related to the motherboard or chipset. Normally on mobile systems as notebooks, the system manufacturers will limit the features and usage of the processor. Also, you should contact the manufacturer for the proper technical support and advise on the supported Operating Systems.

I'm guessing this is because Intel sells these "32 bit" only processors at a lower price to manufacturers. The Intel website (and previously the Wikipedia entry) were misleading / not revealing the full truth to state "64 bit" as the instruction set. Pragmatool‎ (talk) 00:11, 30 August 2012 (UTC)

Are you aware that the paragraph you inserted still has no WP:RS to back the claim? HumphreyW (talk) 00:40, 30 August 2012 (UTC)
Are you aware that the information you have just erased is accurate?
  • I am still awaiting feedback from Asus at this point. They confirm that "Sorry, there is no drivers of win7 64bit for X101CH temporarily. So we suggest you install win7 32bit. Thanks for your understanding." - however I am continuing to press them for a clear answer
  • Please note and read Jasper Deng's comments above where he states: "nor can you say that Atom has 100% 64-bit support"
  • Please also note that his other statement of "specific models of the Atom do have 64-bit support missing".
  • Reference my own experiences with a laptop that I own
  • Reference the other link that I had posted from a (very) well regarded technical information source that indicated an identical problem.
  • Reference Intel's statements (above) indicating that system manufacturers will limit the features
  • Reference Asus's web page that does NOT list 64 or 32 bit support -
All of this - means that this (accurate) statement I posted should stand for at least a few weeks. Each email back and forth between Asus and others seems to take at least a day. Therefore - while the information I've posted is accurate - obtaining a WPRS is difficult at best - because the various corporations involved are probably less than keen to admit to limitations in their products. I'm not someone posting the bizarre "10 questions" items to deface Wikipedia. Instead - I have a fact that I can show (if required) with screenshots. Your bizarre keen-ness to revert my change rapidly is leading me to question your impartiality. Can you state any affiliations (shares, relationships, etc) - with manufacturers / organisations involved in this?

Pragmatool‎ (talk) 08:25, 31 August 2012 (UTC)

You still need an RS. Statements made on Wikipedia talk pages do not constitute an RS. HumphreyW (talk) 08:29, 31 August 2012 (UTC)
Yes, and I'm trying to find an RS - however this process is taking time. Perhaps you can add value to this process by searching for and finding an RS for the 64/32 issue as found on my laptop and corroborated by Jasper and others. Simply 'undoing' a "I can see it on my laptop before my eyes" fact is not creating value. Please help - not tie up this process in Wikipedia red tape. Pragmatool‎ (talk) 10:30, 31 August 2012 (UTC)
I also note that you do not state that you have no affiliation with Intel / Asus / manufacturers in general....Pragmatool‎ (talk) 10:31, 31 August 2012 (UTC)
I would suggest that the text be removed until a RS is found, rather than leaving it there against the policies of Wikipedia. Once you have a RS then there should not be any objection to its inclusion from anyone (assuming it is deemed relevant to the article). The truthfulness, or otherwise, has never been questioned, so your claim that it must stay simply because it is true bears no weight.
Also, if you think I have some conflict of interest then you can report me at the noticeboard. HumphreyW (talk) 11:52, 31 August 2012 (UTC)
I would suggest that there is strong evidence for two facts: The 64/32 bit fact that I've stated (albeit without RS), and the fact that an RS appears to be very difficult to discover due to the reticence of corporations to reveal limitations in their products. I agree this is "against the policies of Wikipedia" however. I am seeing no denial of these facts - merely a difficulty in finding an RS. You seem to be limiting your contribution to this conundrum by simply bot-like reverting pages. Rather than that I would request that you help find an RS as a way to both close this discussion and positively contribute to knowledge as a whole. Pragmatool‎ (talk) 00:37, 1 September 2012 (UTC)
Since you want to insert the information, the burden is on you to find the reliable sources.--Jasper Deng (talk) 00:57, 1 September 2012 (UTC)
And I am doing so - however Internet searches haven't as yet provided anything other than circumstantial evidence (ref the Asus page above that does not mention 64 bit support) or evidence that HumphreyW doesn't view as an RS (links I've provided previously). I'm in discussions with Asus at the moment regarding this - but am yet to obtain anything concrete from them other than a "suggestion" that installing 32 bit is best due to "a lack of drivers". These discussions are ongoing - but each request / response cycle takes a day or more. I am not stating that HumphreyW should "find an RS" - but simply indicating that he's restricting his contribution to simply undoing my edits. I'm suggesting that a better way to close this discussion is to find an RS of some form - and that his efforts (should he choose to) could be used to contribute an RS rather than undo an edit. Indeed - I think a "pure role" of simply undoing edits based on simplistic rules of "does this edit have an RS or not" should best be left up to a bot, and have humans put their efforts towards activities that can't be done by bots (eg finding an RS). In short: his role need not simply be "undoing" - but also "adding value". This change in contribution type is activity is up to him, but I believe that this is the fastest approach to finalising the 64/32 bit debate.Pragmatool‎ (talk) 22:31, 2 September 2012 (UTC)
Please note - in reading the news this morning I was directed to the following article on Wikipedia: This has a "citation needed" section to it - something that presumably should be erased following the RS guidelines. To follow this other example I'm happy to update the Intel Atom 64/32 information to "citation needed" while I search for an RS.Pragmatool‎ (talk) 22:31, 2 September 2012 (UTC)
One vendor / laptop manufacturer's response to my requests for 64 bit support for the (laptop I own) (an Intel Atom powered eee laptop) is as follows: "The model (name of model) EeePC does not support a 64-Bit OS. ". This is the closest I've found to a definitive vendor statement that at least one Atom powered laptop does not support 64 bit - most probably due to a deliberate limitation by the vendor of the functions available on the chipset. I can provide the original wording on request. This is the closest to a 'reliable source' that you're likely to get in the near term, as getting a laptop vendor to state "we deliberately restricted our laptop to 32 bit when the chipset supports 64 bit" is highly unlikely.Pragmatool‎ (talk) 00:54, 14 September 2012 (UTC)
I have removed these claims of lack of 64-support. It appears to be only an affect of the entire system and not the CPU that lacks support 64-bit here. So it seems the CPUs do support 64-bit, but if your laptop/tablet/whatever maker decides to disable 64-bit then this is not a CPU shortcoming. HumphreyW (talk) 07:04, 16 September 2012 (UTC)

Humphrey - Your simplistic approach to "improving" Wikipedia by doing NOTHING but undoing other people's edits is quite probably doing considerable harm (removal of useful information) as it is good (removal of added information that's incorrect). There will be NO place within the Wikipedia hierarchy of pages that suitably informs the reader of the unsupported nature of 64 bit on some Atom-powered machines. Most readers - up to and including people that are reasonably familiar with CPU architectures - will RELY on sources such as Wikipedia for information. As such, I believe the following criteria should be set for content that covers topics / issues such as this one.

  • Information is supported by some form of evidence - even if not an "RS". This evidence could be (for example):
    • Screenshots of "64 bit mode not available", as well as an
    • An emailed statement statement from a laptop manufacturer that the laptop does not support 64 bit mode
  • That the information does not mislead the reader
  • The information is applicable to a wide variety of people (not fine-tuned to a tiny group)

Given these criteria I would state that your "improvements" to the article:

  • MISLEAD the reader that any Atom-powered laptop will support 64 bit (as - in your eyes - it is not the fault of the CPU that the laptop not support 64 bit). By removing the information regarding "if enabled" you give the reader no option but to assume that all Atom-powered laptops will support 64 bit
  • Ignore the evidence that I have supplied. Whilst not an "RS" - it is CLEAR evidence that would be supported by any legal court that 64 bit is not enabled in some laptops. By removing the information for lack of an RS you appear to be requiring MORE evidentiary support than would be required by a criminal court.
  • I require you at this point to state any affiliation with Intel or any of Intel's suppliers. I have asked this previously - and you have ignored the request. It is time to state either "no affiliation" - or some. Your keen-ness to alter this article to remove the information supplied is at best questionable.
  • In the meantime - I will alter the article to provide the reader with useful information - that some Atom-powered systems will not support 64 bit.Pragmatool‎ (talk) 13:11, 28 September 2012 (UTC)
You have no RS, as you stated, and yet you claim it is okay to put non-RS and non-supporting references into the article. So are you implying that the WP:RS and WP:V policies need to be changed? If so, then this article is not the place to do it. Wikipedia is not a criminal court; your comparison adds no weight to your argument. Also as I stated previously I won't be commenting on your "requirement" for any personal information. If you think I have a COI you are invited to bring it up a the admin noticeboard. HumphreyW (talk) 14:44, 28 September 2012 (UTC)
No RS - I would also state that much of the article's other content does not have an "RS" - but instead have references to magazines, online websites and sites such as "Tweaktown". Apparently this is content you approve of, as you have not removed this content. I've supplied as much evidence from reliable sources as these, yet for some reason you are not disputing this other article content - only mine. Please explain why you differ in your treatment of content from "Tweaktown" as from "StackExchange" and from from evidence such as emails from a laptop manufacturer and Intel themselves.
COI - I have no idea if you have a COI. This is a quite reasonable concern of mine given your dedicated interest in allowing only your own views on the Intel page. Your statement that you won't be supplying "personal information" indicating that you're not affiliated with Intel seems bizarre at best. This is hardly your name, country, address, SSN or any other information which would uniquely identify you. "HumpreyW" is far more personally identifiable (British, first name Humphrey, Last name starts with W) - yet you've supplied this readily. Please indicate your affiliations to ensure that your edits are not the subject of a COI. Pragmatool‎ (talk) 05:22, 29 September 2012 (UTC)

Third opinion[edit]


To help resolve this dispute, I'd like to offer a third opinion.

To HumphreyW
I reviewed the long history of dispute in the article and I understand that all the disputes can be summarized into the rejection of this edit. After examining it, I have to say:
  1. As you know Wikipedia articles are meant to be "representing fairly, proportionately, and as far as possible without bias, all significant views that have been published by reliable sources". (WP:NPOV) In that regard, Intel's quantified statements about its own CPUs' technical specifications meets the inclusion criteria.
  2. User:Pragmatool‎ has included an Intel source that confirm Atom N2600 features Intel 64 (also known as x86-64 or just x64) and says:

    "64-bit computing on Intel® architecture requires a computer system with a processor, chipset, BIOS, operating system, device drivers and applications enabled for Intel® 64 architecture. Processors will not operate (including 32-bit operation) without an Intel 64 architecture-enabled BIOS."

    A hover-on description by Intel in the same page re-confirms that Intel 64 subsystem is responsible for 64-bit computing. Including these information in a Wikipedia article, with the support of the aforementioned source, is allowed.
  3. I cannot help but take issues with the factual accuracy of edit summaries. For example: "This article is about the Atom chip. Removing WP:OR." WP:OR does not apply when a source exists, even a source of disputed reliability. In addition, the first part of the comment is irrelevant at best.
To Pragmatool‎
  1. As I explained above, I believe that most of your contributions are allowed in Wikipedia even in absence of source. Therefore, it best if you let it go.
The statement "best if I let it go" and "most of your contributions as allowed" seem confusing. Are you indicating (as I suspect) that the contributions are permissible and *should be included*? If so - I am happy to refrain from further edits (indeed - my entire purpose was to include the useful information that I have provided). I presume that HumpreyW will similarly agree not to remove my edits and remove the information.Pragmatool‎ (talk) 05:08, 29 September 2012 (UTC)
Hi. "Most of your contributions" refers to what explained in HumphreyW section above. To put other ways, I believe "64-bit support" section can be kept but its last sentence (see #3 below) and its citation should be deleted. As for your edit in the table, I think they are redundant but no pressure at all. Best regards, Codename Lisa (talk) 12:51, 29 September 2012 (UTC)
  1. I do not comment on the reliability of Please consult with Wikipedia:RSN if you wish.
  2. I take issues with your sentence: "This renders the CPU in these cases effectively 32 bit." This in parts contradicts with Intels statement that "Processors will not operate (including 32-bit operation) without an Intel 64 architecture-enabled BIOS". In other parts, it is your own synthesis of published material. I am sorry, but this is not allowed in Wikipedia.
I agree "renders the CPU ..." sentence is incorrect. A better sentence is something to the effect that the entire system is limited to 32 instructions only. Something similar to: "In this case the system as a whole supports only 32 bit instructions". Pragmatool‎ (talk) 05:08, 29 September 2012 (UTC)
Also - would you be able to provide a URL for the "admin boards" that Humphrey references above. I will take him up on his request to raise the question of his COI with admins, as I believe that this requires resolution at some point.
Hi. HumphreyW was talking about two different noticeboards. The first board is Wikipedia:RSN. The second board is about COI, but you must not do this. HumphreyW said a lot of wrong things; this one is not only wrong but also dangerous for you. Forget about COI. Always discuss the contribution, never the contributor. Just be polite and kind to everyone and everything will be alright. Best regards, Codename Lisa (talk) 12:51, 29 September 2012 (UTC)

Best regards,
Codename Lisa (talk) 22:36, 28 September 2012 (UTC)

Please note
I believe HumpreyW should still be required to state any affiliations with Intel or related companies. My affiliations have been with Microsoft, and to a lesser extent with Linux and Open Source - however I have no hardware affiliations whatsoever. His response should surely be a simple statement similar to mine above, and I am still concerned that he may hold these affiliations without stating them.
I also believe that while some of Humprey's edits appear to erase "damage" - he appears to be limiting his role at Wikipedia to removing information (some of it valuable / useful - as mine is) rather than correcting, following up, or otherwise "enriching" the information. I take your "third opinion" as an example of this high-value contribution. I would (strongly) suggest that Humphrey expand his "vocabulary of worth" to include enriching activities as well as simply "undo" activities.Pragmatool‎ (talk) 05:08, 29 September 2012 (UTC)
Hi. No, no, my friend. I think both you and HumphreyW should drop this line of discussion immediately. Always discuss the contribution, never the contributor. Best regards, Codename Lisa (talk) 12:51, 29 September 2012 (UTC)
Fine - then let me state this for the 'final record': I've previously assumed that the "edit war" that Humprey started was finalised when I provided as much RS information as I could (this was prior to me discovering that significant portions of the Atom page reference sources such as "TweakTown" which are only slightly more reliable than StackExchange).
Humphrey then removed by edits again after a delay of a week or two - restarting the edit war by (yet again) enforcing his own opinion.
I'm not particularly happy with my early experiences of adding value to Wikipedia:
  • I find out a particularly annoying piece of information about Intel Atom (Intel states 64 bit - my eee laptop only supports 32 bit)
  • I add this information to the page - it's immedately removed by the 'veteran' Humphrey's "undo" behaviour with terse justification
  • I add the RS that Humphrey requests - he undoes it on the basis that it's not strong enough. He does not
  • Cross-check my reference against the strength of the other references on the site
  • Does not offer additional value - such as likely RS's beyond the 'undo' action
  • In general Humprey has made my early experience with Wikipedia extremely discouraging, and seems to point towards an attitude of "my viewpoint is the only correct viewpoint in the world" from a "wikipedia veteran"
In finalising - can someone update the page to add back the 64 bit support, and to add the "yes (if enabled)" section to the table.
I will be watching this page every so often. You can be sure that I will escalate well beyond the group that I have engaged with this time should Humprey (or any sock puppet) remove my (useful) information once again.Noasshats (talk) 01:31, 1 October 2012 (UTC)

Newly Updated Intel 64 Section[edit]

I have assembled what I hope are enough citations to put this issue to rest. There is now a section on 64-bit support that contains multiple cites and attempts to clearly differentiate between different issues (which CPUs contain Intel 64 hardware support, what role drivers play in 64-bit support, etc.). I realize that not all the cites used would be considered RS by all editors, but this is important information, and I believe the cites used are the best currently available. Therefore, I ask that any editor who has an issue with the sources please discuss here on Talk first and not delete anything without further discussion.

Shelbystripes (talk) 05:16, 25 March 2013 (UTC)

Some References Appear Rather Blatently Misrepresented[edit]

For eg: The Core 2 Duo SU7300 outperforms the dual-core Nano.[35]

Appears completely unsubstantiated after reading the reference. — Preceding unsigned comment added by (talk) 19:25, 24 January 2014 (UTC)