Talk:List of Intel Xeon microprocessors/Archive 1

Wrong number of memory channels: 8

All the Xeon E7-28xx v2, Xeon E7-48xx v2 and Xeon E7-88xx v2 CPUs show 8 * DDR3 memory channels. But the Intel ARK pages show that all of these CPUs have 4 memory channels, so it seems wrong. — Preceding unsigned comment added by 205.250.240.122 (talk) 11:58, 20 February 2014 (UTC)

Xeon E7 V2 has two quad channel memory controllers. — Preceding unsigned comment added by 2001:CC0:A014:99:F567:EFF5:C9A4:B1B6 (talk) 01:27, 21 February 2014 (UTC)

Untitled

The Xeon 3040 and 3050 are not Allendale core, they are Conroes with half of the L2 disabled, much like the C2D e6300 and e6400. The hardcore icon the sandman 21:51, 25 May 2007 (UTC)

Yes and no, there exists now both versions after the Allendale stepping was released in Janaury 2007, the old version is Conroe with 1/2 the LV2 disabled, the newer cores only have 2MB LV2 natively.Coldpower27 14:38, 26 May 2007 (UTC)

Changed order of entries

I just finished changing the order of the entries. The previous system (by the number of cores each CPU had), did not make much sense to me. It is now arranged by the microarchitecture (P6, NetBurst, Pentium M (Yonah), Core), and then by the level of multiprocessing supported (Xeon UP/DP and Xeon MP). This setup is also a bit more chronological than the previous one. Let me know what you think of it. Imperator3733 17:58, 29 July 2007 (UTC)

BalusC site offline

Unfortunately, BalusC has taken his site offline (what a huge loss for the internet!) and it won't be online for some time as far as I know. His site currently only displays a 'bye' message because he's emigrating from the Netherlands to Curaçao. I guess all articles mentioning his site (http://en.wikipedia.org/wiki/Special:Search?search=balusc) should be updated?? RobIII 00:20, 19 September 2007 (UTC)

Socket 604

All new quad processors with fast bus must be checked for using LGA771 or newer Socket 604. I've found some errors about it in the article. `a5b 03:07, 2 December 2007 (UTC)

Socket 604 is actually several years older than LGA771. Nonetheless, this is something to look into. — Aluvus t/c 04:00, 2 December 2007 (UTC)

True, but the old Tigerton quad and dunnington hex cores platforms use the 'older' s604 sockets. —Preceding unsigned comment added by 24.90.67.152 (talk) 15:39, 30 May 2009 (UTC)

Xeon 3XX3 processors

Does anyone have information on these processors (E3113, X3323, X3353, X3363)? Intel doesn't seem to acknowledge their existence on their website, but I've seen them listed as processor options on some servers (for example, the IBM BladeCenter HS12) —Preceding unsigned comment added by BlueLikeYou (talk • contribs) 19:21, 2 June 2008 (UTC)

New Nehalem processors

There are several new Nehalam processors I don't have the full specs on, but it looks like Intel released them today. There are the L5520 and L5506, which seem to be low power variants of the E5520 and E5506. There are the W3570, W3540, and W3520, which seem to be 130W uniprocessor chips. And finally, the L5518 and L5508, which are for embedded servers. -- Schapel (talk) 02:50, 31 March 2009 (UTC)

Also, the "Gainestown" parts are listed under "Core-based processors". There should be a new section called "Nehalem-based processors" I guess. 123.204.180.152 (talk) 16:10, 31 March 2009 (UTC)

Isn't Nehalem based on Core i7, and other processors in that section based on Core 2? I would think it would make sense to have three sections: Core, Core 2, and Core i7. -- Schapel (talk) 00:13, 1 April 2009 (UTC)

Both the 5500 and 3500 series support the Nehalem-based feature called “Deep Power Down technology,” which is enabled through a new "C6" processor power state. That could be added to the processor descriptions. Also an update may be required to older operating system software to enable the C6 power state, which isn't yet supported by the current ACPI (Advanced Configuration & Power Interface) standard. 72.73.92.194 (talk) 05:40, 5 April 2009 (UTC)

The 3520 seems to be the Core I7 920 with the addition of ECC and DCA, "Direct Cache Access.” It's the same with the 3540 and the Core I7 940, however the 3570 apparently lacks the Core I7 965 "Extreme" features. Also the 3500 series, like the Core I7, claims "binary compatible with applications running on previous members of the Intel microprocessor line." I could not find that referenced for the 5500 series. 72.73.92.194 (talk) 05:40, 5 April 2009 (UTC)

Editing note: To be consistent with the rest of the article, tables should be in ascending order by frequency - not descending order. 71.254.107.104 (talk) 05:47, 2 April 2009 (UTC)

Level 2 Cache Speeds

It would be interesting to show the Level 2 cache speeds. I believe that for some of the early Xeons, L2 cache was half the core clock, and others had 1:1 cache speeds. I don't have the info. maybe someone can add it? 60.240.207.146 (talk) 08:43, 15 December 2009 (UTC)

W3680 does not have 2x QPI links

However, I found no way to edit it out in the list, see http://ark.intel.com/Product.aspx?id=47917 for source -- 195.67.83.146 (talk) 10:42, 17 March 2010 (UTC)

You could pass an iobus= argument to override the automatic entry from QPI. I've also added a new "gulftownup" type for the UP-version of Gulftown. Arndbergmann (talk) 13:53, 17 March 2010 (UTC)

Number of cores

We need to add the number of cores to the new 56xx series of Xeon processors as it's very difficult to tell how many cores each has based on the table alone. —Preceding unsigned comment added by 112.213.146.100 (talk) 18:04, 30 March 2010 (UTC)

Agreed. How about replacing the (pointless) L2 cache column with a 'Cores' column in Template:cpulist/nehalem and Template:cpulist/nehgfx? Arndbergmann (talk) 11:03, 31 March 2010 (UTC)

EIST in "Clovertown" (65nm)

According to this page http://www.cpu-world.com/CPUs/Xeon/Intel-Xeon%20E5310%20-%20HH80563QH0258M%20(BX80563E5310A%20-%20BX80563E5310P).html and this one http://ark.intel.com/products/28030/Intel-Xeon-Processor-E5310-(8M-Cache-1_60-GHz-1066-MHz-FSB) at least the E5310 supports EIST. Since the ark.intel.com site may still contain errors, can someone owning such a cpu confirm it? — Preceding unsigned comment added by 129.13.97.3 (talk) 11:43, 31 August 2011 (UTC)

Gallatin and HT

Hi, i think Gallatin have HT, but looking at ark.intel.com it returns me "number of thread: 1", for every gallatin based xeon I checked. Strange. — Preceding unsigned comment added by 91.213.255.7 (talk) 01:17, 17 January 2012 (UTC)

https://ark.intel.com/products/27280/Intel-Xeon-Processor-3_20-GHz-2M-Cache-533-MHz-FSB Shows that the Gallatin 3.2Ghz has Hyperthreading enabled. Further, Intel's page shows 2MB L3 cache, but cpu-world shows 512k L2 cache and 2MB L3 cache. http://www.cpu-world.com/sspec/SL/SL7BW.html

I actually have one of the 3.2Ghz chips, but not sure if it's a DP or MP(although I'm sure it's a 533FSB so it should be a DP). Will have to pull the system up and look.

How to interpret Sandy Bridge Turbo Boost

I understand how to interpret the Turbo stepping designation (e.g., 1/1/2/2 on the four core E52643). What I don't understand is the single clock rate entries (e.g., 2.4 GHz on the E5-2658). Does this imply: 1. All cores will run simultaniously at this rate. (A consistent designation of 3/3/3/3/3/3/3/3 would be preferred.) 2. Only one core will run at this rate. (A consistent designation of 0/0/0/0/0/0/0/3 would be preferred.) 3. This is the maximum turbo rate and the stepping is unknown. 4 Some other meaning.

HyperZink (talk) 21:05, 12 June 2012 (UTC)

I think I've found my own answer (#3 above) in http://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/xeon-e5-family-spec-update.pdf where the "Turbo bins" for the SR0LZ (E5-2658) spec are listed as: 0/0/1/1/2/2/3/3. I'll see if I can make that update and check the spec update against any similar notations on the wiki.

HyperZink (talk) 21:31, 12 June 2012 (UTC)

there are cpus listed under the sandy bridge listings, instead of below the sandy bridge listings on their own ivy bridge listing

........... — Preceding unsigned comment added by 94.66.125.112 (talk) 00:03, 2 September 2012 (UTC)

10 CORES

According to Asrock specifications Xeon E5-2643 v2 and Xeon E5-2687W v2 have 10 cores, 25mb lv3 cache there is no proof they are wrong.Processors can be better binned thats why they have higher clock rate.--Emperor-Overlord100 (talk) 14:15, 14 July 2013 (UTC)

If Indeed E5-2643 v2 is 10 cores and running at 3.5Ghz, then it makes no sense of the Processor numbering scale, since it's slightly faster then E5-2687W AND draws less power! — Preceding unsigned comment added by 220.128.212.69 (talk) 05:15, 17 July 2013 (UTC)

2643 v2 with 10 cores is obviously wrong. V1 of the E5-2643 is 4 cores running at 3.3Ghz, so to retain the same xeon naming it should have similar core count. so 6 core and running at 3.5Ghz should be correct. definately not 10 cores. The leaked spec is obviously wrong. — Preceding unsigned comment added by 59.120.147.113 (talk) 13:19, 17 July 2013 (UTC)

It could draw less power because it can have lower turbo,disabled hyperthreading and/or is better binned it doesnt need to have similiar number of cores because sandybridge-EP had lower number of cores altogether and was manufacted at 32nm and Ivy is 22nm .Just because someone thinks something is wrong doesnt make it true only hard data comes from Asrock for now , so j request for you to stop changing it until processors are released if J am wrong J will then change it back . — Preceding unsigned comment added by Emperor-Overlord100 (talk • contribs) 06:33, 21 July 2013 (UTC)

The leak showing 10 cores is likely incorrect. The clock speeds and the TDP do not make sense for a 10 core part when compared to the rest of line up. Combine that with the fact that the core count was derived from the cache/core, which we know Intel varies from part to part, and your assertion is not as solid as you seem to think. As you are now out voted 3 to 1 on this issue, I request that you stop changing the table.2601:1:9C80:82:D061:F76E:11C6:3334 (talk) 15:15, 21 July 2013 (UTC)

10 cores at 3.2ghz is confirmed so its plausible they could achieve 3.4 and 3.5ghs as well,E5-2643 v2 could have lower turbo or disabled.J have solid proof and you have no proof Asrock is wrong. HT.http://www.xtremesystems.org/forums/showthread.php?285956-Coolaler-shows-us-just-how-pathetic-the-i7-4960X-will-be.../page3</ref> — Preceding unsigned comment added by Emperor-Overlord100 (talk • contribs) 18:36, 23 July 2013 (UTC)

3.2 GHz with 10 cores is a 150W part, you're not going to increase clocks almost 10%, at the same time decreasing TDP by almost 15%. No, you do not have solid proof, the 10 core assumption for 2643 v2 is based on a list that showed a part with 25MB cache. The link you just posted shows a 6 core part with 25MB cache, a 130W TDP and 3.3 GHz clocks. http://imageshack.us/a/img841/8817/cam00101o.jpg

I'm reverting your edits, and if you change them again, I'll request that this section be locked. — Preceding unsigned comment added by 2601:1:9C80:82:4B4:BCDB:4173:7946 (talk) 21:16, 23 July 2013 (UTC)

Haswell formatting

Can someone fix up the Haswell section formatting so that it doesn't look like something from another article?  Tabanger  03:25, 25 July 2013 (UTC)

Missing references for Turbo modes

The article seems to be missing references for the turbo modes, especially for the Xeon E3 and E5 families. I was unable to find them in the existing references. I also tried searching on ark.intel.com, which was what was linked to. They also don't appear there, or have been removed. Laureled (talk) 16:22, 2 April 2014 (UTC)

Intel calls them "turbo bins", please have a look at pages 8–11 of this PDF, for example. As there doesn't seem to be an easy way for adding it as a reference, I've added it into the "External links" section. — Dsimic (talk | contribs) 05:57, 2 December 2014 (UTC)

Xeon-SP Turbo modes

Page 8 of the Anantech article on the Xeon-SP vs. AMD shootout has a detailed list of Turbo behavior in Xeon-SP CPUs. It's the best overview of the detailed behavior I could find, including performance in AVX2.0 and AVX512 operating modes.

See https://www.anandtech.com/show/11544/intel-skylake-ep-vs-amd-epyc-7000-cpu-battle-of-the-decade/8 -- 5.57.21.155 (talk) 13:17, 16 July 2018 (UTC)

Misc. new E5 v3 CPUs

Found a couple more, but missing either core count or are an early sample. Will be back with more if I find them.

Xeon E5-2623v3 3.00GHz CM8064401832000
Xeon E5-2652v3 ES, 2.3GHz, 10/20, 25MB L3 cache --Azul120 (talk) 03:11, 18 June 2014 (UTC)

8.4.x Haswell E7 processors - L3 Cache numbers are missing

The two tables for Haswell E7 processors have missing L3 cache numbers. Entries read 'MB' but doesn't specify the number. Composerananth (talk) 02:55, 25 October 2014 (UTC)

Broken Link for Xeon E7-4870 (10 core)

Hello, the link for Xeon E7-4870 does not work anymore: http://ark.intel.com/products/53579 returns "The server encountered an internal error or misconfiguration and was unable to complete your request." Proper link is http://ark.intel.com/products/53579/Intel-Xeon-Processor-E7-4870-30M-Cache-2_40-GHz-6_40-GTs-Intel-QPI. I'm not sure how to fix this as table is generating links autonomously.

Update: ark=53579/Intel-Xeon-Processor-E7-4870-30M-Cache-2_40-GHz-6_40-GTs-Intel-QPI has worked. I have update main article with this new link until Intel fixes it.

Jhla5146 (talk) 11:05, 4 August 2015 (UTC)

Hello! That's been some kind of a glitch, the short URL works fine now and I've restored it. — Dsimic (talk | contribs) 06:56, 2 September 2015 (UTC)

Hyperthreading

Processor that enable hypertreading should reference the number of logical core (thread). IMAO, the terme logical core should be prefered to the therme thread that can be confusing with https://en.wikipedia.org/wiki/Thread_%28computing%29 — Preceding unsigned comment added by 213.174.103.142 (talk) 09:01, 19 January 2016 (UTC)

Memory Types

Shouldn't the column "Memory" list all types of memory that are supported? For instance an E5-2695v3 can use 3 types of DDR4 Memory (DDR4-1600/DDR4-1866/DDR4-2133) and not just one, see http://ark.intel.com/products/81057.

I don't see there being much point in listing all types of memory that are supported. It is a given that any CPU that supports DDR4 memory will support all speeds up to and including the maximum. So if its maximum is DDR4-2133, we know that it will support DDR4-1600 and DDR4-1866 as well, however you would not get the best memory performance then; but this is not useful information. However, showing the maximum IS useful information, as the Xeon E5-26xx v3 processors can support speeds up to DDR4-2133 (68.3GB/s), but the Xeon E5-26xx v4 processors can support speeds up to DDR4-2400 (76.8GB/s). - Extec286 (talk) 23:12, 7 April 2016 (UTC)

Broadwell-H Xeons (E3-12xx v4) AVX2, FMA support?

Broadwell-H (Xeon E3-12xx v4) don't list AVX2 FMA as supported? I thought the Broadwell architecture/core supported AVX2 and FMA? Did Intel disable this for these Xeons or is it just missing in the list? VictordeHollander (talk) 12:25, 28 April 2017 (UTC)

Xeon Gold 5122 Memory Speed

According to the ARK (https://ark.intel.com/products/120475/Intel-Xeon-Gold-5122-Processor-16_5M-Cache-3_60-GHz) the Xeon Gold 5122 actually has a memory speed of 2666 MHz and not 2400 MHz. The other Xeon Gold 51xx series do seem to all have 2400 MHz memory speed. It is strange that that single processor doesn't fit the pattern. Even stranger the following page (https://www.intel.com/content/www/us/en/products/processors/xeon/scalable/gold-processors/gold-5122.html) which is an official Intel page says it uses DDR4-2400 but has a maximum memory speed of 2.67 GHz... which of all of these is correct? — Preceding unsigned comment added by Thaimin (talk • contribs) 17:16, 7 September 2017 (UTC)

I sent feedback to Intel and they have fixed the second non-ARK page and confirmed that "The 2666 MHz specification that you see listed on the Xeon Gold 5122 page on the ARK site as well as on the intel.com content page is correct for that processor. It is true that all the 51xx series use 2400 MHz memory, except for the 5122. The 5122 also has two AVX-512 FMA units, which makes it a bit different than the other processors in the series." I have updated the Xeon processor list for this processor. Thaimin (talk) 19:39, 12 September 2017 (UTC)

External links modified

Hello fellow Wikipedians,

I have just modified one external link on List of Intel Xeon microprocessors. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

• Added archive https://web.archive.org/web/20100304141848/http://ark.intel.com/ProductCollection.aspx?familyID=594 to http://ark.intel.com/ProductCollection.aspx?familyId=594

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 18 January 2022).

• If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
• If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 18:07, 25 December 2017 (UTC)