Template talk:Cpulist

Requests/Wishlist

• I would like to see multiple sockets implemented so something beyond just sock= is available. Something along similar lines as the steppings, sspecs would be good where one could make a sock1= sock2= sock3= sock4= would be good. 134.134.139.73 (talk) 00:01, 5 March 2010 (UTC)

I just added this as an option, now sock= still behave as before but if you pass sock1, it takes that (and sock2, sock3), showing it without further mangling or wikilinking. Arndbergmann (talk) 21:01, 5 March 2010 (UTC)

Thanks! I noticed there is no sock4= to line up with the sspec4= step4= part4=. I also noticed the sock1=, sock2=, sock3= do not expand like sock= so I cannot set them to values like 989 and 1288 and have automatically expand. I updated the following to use the new sock1= sock2= sock3=: Core i3-330M, Core i3-350M, Core i5-430M, Core i5-520M, Core i5-540M, Core i-620M7.192.102.209.29 (talk) 20:59, 15 March 2010 (UTC)

• I would also like to see a place for graphics turbo on the nhmgfx like Arrandale that have graphics turbo ability. It might also be interesting to specify the turbo values separately such that they are machine readable and the individual speeds could be calculated/viewed somehow. 134.134.139.73 (talk) 00:01, 5 March 2010 (UTC)

• The template needs to be changed a bit for the List of Intel Xeonmicroprcessors article. We need a column for number of core. Since there are many released with 1-8 cores on each processor in the same generation, it doesn't make sense to make seperate tables anymore that are seperated by core count. The processors are all grouped together now no matter the number of cores, but it isn't easy to tell how many each one actually has. Paranoidmage (talk) 03:45, 23 June 2010 (UTC)
  • Since nobody else stepped up, I just did the trivial change to implement this and enabled the 'Cores' column for the nehalem and nehgfx table layouts. I'd also suggest dropping the useless L2 cache column for these now to make up for the lost space. Arndbergmann (talk) 19:43, 16 June 2010 (UTC)

• Also, the "Uncore Speed"column should be removed unless a value is given for it, since it is blank in all of the columns right now. It should probably also be renamed to "Uncore Frequency" or "Uncore Freq." Paranoidmage (talk) 03:45, 23 June 2010 (UTC)

• automatic colspan and rowspan, at least for the socket. User:ScotXW^t@lk 07:12, 23 March 2019 (UTC)

  @ScotXW: Meaning... what, like if the same data is entered in multiple adjacent cells, they'll automatically be joined together? I just have to ask, is that even a thing?

  This talk page is the only hit for "automatic colspan" across the entirety of Wikipedia, and "automatic rowspan" has no hits whatsoever. Presumably that's because it's simply impossible to implement, at least in tables constructed by multiple template transclusions (since they can't interact with each other). It might theoretically be feasible if the templates were converted to a Lua Module, but this isn't a technology Wikipedia currently has available to it... like, at all. -- FeRDNYC (talk) 03:00, 30 March 2019 (UTC)

• Articles such as e.g. List_of_Nvidia_graphics_processing_units#GeForce_900_series are maintained manually. They are un-encyclopedic, but happen to be very useful, as the manufacturer don't deem it useful to inform their potential customer to well! This would negate most of the hard work done by the marketing departments… However, when you fuse cells, which contain the same information together, such tables become easier to parse for the human reader. User:ScotXW^t@lk 07:35, 30 March 2019 (UTC)

P5 Bus

FSB was not created until P6 so I suggest using bus= vs. fsb= and having the header display Bus Uzume (talk) 16:11, 2 June 2010 (UTC)

It seems fsb= needs to be fully merged into the iobus= in the template anyway. Perhaps this is a good time to work on that. Uzume (talk) 16:16, 2 June 2010 (UTC)

MHz or GHz

Where should the cut-off be for printing the CPU frequency in MHz or GHz? The code used to use GHz only above 10GHz, everything up to 9900MHz would get printed as a large number of MHz. I've changed the cut-off to 990 MHz / 1 GHz now, which I suppose is what people usually expect. Any other opinions? -- Arndbergmann (talk) 09:34, 7 October 2010 (UTC)

shrinking the table font

Hello,

I've been experimenting a bit with this template as I want to use it in the Nehalem_(microarchitecture) page. However, the template does net look good, since the generated table does not fit in the page and the text in the cells becomes very hard to read as it spans several columns. Since I've seen similar issues in the other pages that use the template, could someone shrink the font, that is used to draw tables? IMO adding style="font-size: 85%" would be enough since most of the technical tables in other pages (such as GPU descriptions) use it. Thanks. 1exec1 (talk) 20:25, 30 January 2011 (UTC)

gpulist

Hello again,

I've done some consistency fixes in the GPU technical tables lately. I think similar template to make all these tables more consistent would be invaluable. However, I lack skills required to create similar template myself. Could someone help me to write a cpulist-like template for the GPU tables?1exec1 (talk) 20:29, 30 January 2011 (UTC)

Please add Intel HD Graphics P3000

Can someone please add the Intel HD Graphics P3000, which is part of the Intel Xeon E3-12x5 workstation processors? It uses seperate drivers from regular HD graphics that are optimized/certified for workstation applications. I can't figure out the wikicode to add it myself. Thanks! Alereon (talk) 03:25, 13 September 2011 (UTC)

Done. Arndbergmann (talk) 07:28, 13 September 2011 (UTC)

gfxclock

Change

{{#if:{{{gfxclock|}}}|{{{gfxclock}}} {{#if: {{{links|}}}|MHz|MHz}}}}

to

{{#ifexpr:{{{gfxclock|}}}|{{{gfxclock}}} {{#if: {{{links|}}}|MHz|MHz}}}}

? For case like |gfxclock={{n/a}}. Example : List_of_future_Intel_microprocessors#.22Sandy_Bridge-E.22_.2832_nm.29. The GPU Frequency is "N/A MHz". Should be "N/A" just. Also should future-proof for GHz etc. (Forgive my english.) 75.4.146.33 (talk) 08:06, 17 November 2011 (UTC)

why not just default to {{n/a}} when the field is left empty? It seems rather pointless to require the lists to contain |gfxclock={{n/a}} for every cpu that does not have embedded graphics. Arndbergmann (talk) 13:30, 17 November 2011 (UTC)

Units/values spacing

Could the template be modified so   is used between values and units, such as in "1 GHz"? Besides looking much better, that would also be inline with the MOS:NBSP. Thoughts? — Dsimic (talk) 19:54, 3 January 2014 (UTC)

fine with me. Frietjes (talk) 16:11, 4 January 2014 (UTC)

update needed

This doc seems to be out of date compared with the template itself. I added a lake-e entry (maybe someone can check it) but am not sure how to tell if more are missing. 50.0.136.56 (talk) 23:55, 17 February 2017 (UTC)

benchmarks?

Is it appropriate to add benchmark data to the templates and lists, like from cpubenchmark.net, browser.geekbench.com, or whatever? I keep referring to those sites for benchmark info. 50.0.136.56 (talk) 23:59, 17 February 2017 (UTC)

I feel that adding anything that would make these tables wider would be inappropriate until their size can be brought under control, TBH. But benchmarks, specifically, would seem to me to be less-than-encyclopedic info. Benchmark data is what those sites are for, and Wikipedia is not a stats book. -- FeRD_NYC (talk) 23:21, 19 March 2018 (UTC)

Cores / threads

Up until the 8th generation, the presence of hyperthreading was so constant that listing the threads in the tables was a bit redundant. For example, all desktop i3's had 2 cors, 4 threads; i5's had 4 cores and i7's had 4 cores, 8 threads. But with the latest generation, the number of cores and threads is less straightforward. I propose changing "cores" to "cores (/threads)" so in recent tables, the number of threads can be added. I don't know template markup language, but I see there's already something about threads. However, it isn't used in tables such as List_of_Intel_Core_i5_microprocessors. PizzaMan ♨♨♨ 12:53, 10 January 2018 (UTC)

Actually, the Westmere Core-i5s (List_of_Intel_Core_i5_microprocessors#"Clarkdale"_(MCP,_32_nm_dual-core)) were only dual-cores with HT enables, while the predecessors and successors were quad-cores without HT. User:ScotXW^t@lk 07:10, 23 March 2019 (UTC)

Column ordering

I really want to update these templates for consistency of column-ordering, as the generational subtemplates all use not only different columns but different order of the columns, and it's become a real mess. When multiple generations are included on the same page, the inconsistent ordering is confusing, and leads to hatnotes like the one that's been placed on List of Intel Core i5 microprocessors since August 2015.

But adopting a consistent column ordering requires that there be an agreed-upon ordering to follow. Working off that i5 list, as a working set I suggest:

• Model
• sSpec
• Cores
• Freq
• Turbo
• L2
• L3
• [GPU
• GPU Freq]
• [Mult.
• Memory
• Voltage]
• TDP
• Socket
• I/O Bus
• Date
• Part#
• Price

Where some or all columns in the sets delimited with [ ], especially, will be present/omitted as a group for some/many generations, so it makes sense to keep those columns together. (Any column is subject to being omitted, or included; this is about ordering, not visibility.) But the first 7 and last 5 columns appear to be present on every list, so I'd like to see if we can have them always appear in the same places on every list. The columns in between can vary when appropriate, but they should do so with a consistent ordering.

There are more cpulist-supported columns not present on the i5 page, and I haven't yet accounted for those here. I need to survey the rest of the generational lists to get a fuller picture. So, consider this a WIP — I'm not yet ready to make any template changes as of yet. I just wanted to open up discussion, if anyone has input, about the column ordering and the need for consistency among these various list templates. -- FeRD_NYC (talk) 23:08, 19 March 2018 (UTC)

"gfxmodel" needs rework

Currently you add things like gfxmodel=10eu; then it should lookup the used sub-template, e.g. "cpulist|haswell" and then link you to Intel_Graphics_Technology#Haswell instead of just Intel_Graphics_Technology; however, Intel_Graphics_Technology#Haswell does not add much more information; more information is actually found at

• Intel_Graphics_Technology#Capabilities
• Intel_Graphics_Technology#Capabilities_(GPU_video_acceleration)
• Intel Quick Sync Video
• maybe somewhere else?

User:ScotXW^t@lk 07:31, 30 March 2019 (UTC)

Bug on Intel Atom List + UHD 615, UHD 617

Hello, 1) On the Pineview section of the List of intel Atom microprocessors, there is "span id=xxx" displayed in the first column. 2) Would it be possible to add UHD Graphics 615 and 617 ? As they are not on the template, "(?)" are added to the Amber Lake-Y section of List of Intel Core i5 microprocessors. Thanks ! - LThecross (talk) 20:20, 12 April 2020 (UTC)

Intel ARK links

The links for ark.intel.com using |ark=12345 are currently linked to https://ark.intel.com/content/www/us/en/ark/compare.html?productIds=12345 which does provide the user a clickable link through to the proper ARK page, but it's not ideal. Link template should be changed to take advantage of the redirect system in place:

from: ark_support@intel.com (2020-09-23)

Doing a pattern replace of

https://ark.intel.com/content/www/us/en/ark/products/<id number here>.html

will trigger our existing redirection/canonization capabilities that were based on our previous capabilities,

but now require a .html on the end. In this example, the redirector can nudge

https://ark.intel.com/content/www/us/en/ark/products/27199.html to

https://ark.intel.com/content/www/us/en/ark/products/27199/intel-celeron-processor-900-mhz-128k-cache-100-mhz-fsb.html

SoftwareSimian (talk) 13:30, 24 September 2020 (UTC)

@SoftwareSimian: Implemented a change to the template to take advantage of this URL magic. If you see anything break, let me know and revert. —Locke Cole • t • c 06:58, 5 July 2021 (UTC)

Locke Cole: The ark parameter is not documented. Please modify the documentation to explain the ark parameter and any other undocumented parameters. Also, I discovered, because of an error in List of Intel Core processors, that using the ark parameter and also a reference link for the model parameter results in a links in links lint error. Please document this as well as a "don't do that". —Anomalocaris (talk) 05:53, 25 May 2023 (UTC)

Release price

I noticed this at Skylake (microarchitecture) but the entire addition of the "release price" is complete trivia. There is no mention of historic release prices for monitors, games, toys, or almost any other product lists I can think of. I'd guess the only source about these would be Intel itself and even then it's the sale in I presume the US market alone that is being reported. I highly doubt there's been a lot of secondary sources discussing release price points for different CPUs. No one has bothered to list this trivia at any of the early List of Intel processors where the historic release price could at least be useful. Instead, it only shows up in the tables that start at the 7th generation core which is when this table routine started. Ricky81682 (talk) 10:04, 4 January 2023 (UTC)

Requests

Dear maintainers, in order to improve the template, I propose the following functional and design changes:

• Add a caption option, per WP:ACCESSIBILITY.
• Add a collapsible option, per WP:ACCESSIBILITY.
• Enable row sorting, per WP:ACCESSIBILITY.
• Add header scopes, per MOS:DTAB.
• Remove the section option, per WP:ACCESSIBILITY.
• Add Brand column (for Core, i3, i5, i7, i9. etc.).
• Simplify the Model nmumber column to Model.
• Add Prefix or Suffix column (depending on the generation), to enable sorting between the different product lines (S, H, HX, P, U, etc.), per WP:ACCESSIBILITY.
• Change the Brand , Model, and Prefix/Suffix row cells to row header cells.
• Remove the sSpec number, and Part number(s) columns, per WP:NOTCATALOG.
• Change Frequency column to Clock rate.
• Group the cache columns (L1, L2, L3, L4) in a Cache "colgroup".
• Add a Node column for litography size (45 nm, 32 nm, etc.)
• Simplify the Release price (USD) column to Release price

Let me know what you think. Thanks! ~ Arkhandar ^{(message me)} 13:49, 21 January 2023 (UTC)

Frankly I don't know why this template exists, I'd rather have all uses of it be completely deprecated and deleted, and go back to the tried-and-true regular wikitables instead. Templates like these just complicate things by, a looooot. Updating the template being one of the big issues here. You don't get that issue with regular tables. — AP 499D25 (talk) 13:21, 1 May 2023 (UTC)

I've used the cpulist template for many years, but I'm starting to agree that it's becoming more trouble than it's worth. Each generation adds more parameters, and with the introduction of "performance" cores and "efficiency" cores, I've found that making my own wikitable based off the existing tables is less of a hassle. --Vossanova ^o< 13:46, 1 May 2023 (UTC)

Another major problem I have discovered with these templates, is their incredibly inefficient resource use. I discovered it after performing the merge of List of Core 2, i3, i5, i7, i9 articles into List of Intel Core, even though the page is only 500 KB in raw size right now (the max. size is 2 MB), when the page is loaded, the post-expand include size limit (also 2 MB) is easily exceeded, resulting in every single template from the "Penryn", "Penryn-3M" (medium-voltage, 45 nm) section onwards breaking, or not displaying correctly, in this permalink.

Had these been all wikitables, with the raw page size being the same or more overall, this issue would have not occurred at all. Thus it wouldn't be necessary to split up the article to 5 CPU generations per list article or something like that. — AP 499D25 (talk) 03:16, 2 May 2023 (UTC)