Talk:List of AMD Athlon 64 processors

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

Incorrect information[edit]

The recent changes by 210.9.3.180 seem to have resulted in incorrect information about the "Newcastle" processor details being entered. Could someone independently verify the information at http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/30430.pdf and confirm whether these changes should be reverted? The information on page 8 seems to confirm the oringinal data. Mhowkins 21:33, 25 Aug 2004 (UTC)

Instruction set[edit]

Is it possible to get a list of whether these have just sse2 or if they also have sse3? I guess a column called sse3 with 'yes' or 'no' marked. Or I dunno, I think all 90nms except winchester have it


"Multiplier" column[edit]

I took out the "Multiplier" column, partly to see if anybody screams.  :-)

I had asked, in a different AMD page, for a definition of Multiplier. It turns out the nothing more than the ratio of the speed (Hz) of the FSB to the CPU itself (the old column 2 divided by the old column 4). This struck me as clutter--not really worth a column--which I'm trying to purge from the AMD pages.

To return to the "screams" comment, if Multiplier turns out to be an absolutely critical piece of information, I'll be happy to revert it (at this point, I've only changed the first table). But before reverting, we would really at least need a pointer to the definition of Multiplier, and maybe even an explanation of why it's important.

DanielVonEhren 23:18, 20 Feb 2005 (UTC)

FSB/HyperTransport[edit]

Would anyone mind if I replaced the current FSB column with a HyperTransport column like the List of AMD Sempron microprocessors page has? IMO, the HT speed is important enough to be listed in the tables instead of in a footnote, and the way K8 does FSB is so non-traditional that it's probably best to leave it out of the tables and either stick it in a footnote or just leave it to be explained on the main Athlon 64 page instead, which is where descriptions of the architecture should belong. --Jgp 05:52, 16 Jun 2005 (UTC)

Well, it's been two days and no one's complained, so I'll go ahead and make the change... --Jgp 22:43, 18 Jun 2005 (UTC)

--Megalon 13:59, 22 May 2006

Added HyperTransport and voltage for Athlon FX-62 --Megalon
Changing "HyperTransport" to "HT" (with a note of what it stands for) to save column space. --Vossanova o< 15:40, 27 February 2007 (UTC)[reply]

chronology?[edit]

I suggest the first field be the year a given model number was first produced in mass quantity.

Dual Cores[edit]

The X2 3800+, 4200+, and 4600+ are all from the Manchester core, and have only 512 KiB cache/core. They are not on the Toledo core, which is only the 4400+ and 4800+ and have 1024 KiB cache/core. Blitz Tiger 23:36, 10 May 2006 (UTC)[reply]

No, the Toledo core can have both 2 × 512 KiB and 2 × 1024 KiB versions. This is sourced. Jgp 03:47, 11 May 2006 (UTC)[reply]
Right, but the 2 × 512 KiB version has the other half of its cache on-die, but disabled, yes? Shouldn't there be a note to that effect? —BorgHunter ubx (talk) 11:33, 11 May 2006 (UTC)[reply]


Dates[edit]

The release dates for the AM2 Windsor and Orleans cores appear to be incorrect, seeing as how it's 2006-06-02, and they're not available anywhere.

2006-05-23 was the official launch date. Furthermore, due to the way shipments trickle into stores and OEMs, tracking initial availability dates would be unverifiable. jgp 23:24, 2 June 2006 (UTC)[reply]

Nakes cores[edit]

Not sure where to suggest this so decided to do it here. I think it might be an idea to include pictures of naked cores on the appropriate processor articles (Athlon 64, X2, Opteron). Someone with the time could perhaps go to one of the places where you're likely to find people who have removed the IHS like extremetech or dfi-street or something (search for naked Athlon 64 or something if you don't know where to go) and ask for a photo taken by a member who is willing to contribute it with an appropriate license Nil Einne 16:32, 19 June 2006 (UTC)[reply]

X2 3600+[edit]

I have searched, yet could find no evidence of a Athlon 64 X2 3600+ on the 939. It is going to be released on the AM2, and as such should not be under the Manchester as well as Windsor.—Preceding unsigned comment added by 141.151.82.145 (talkcontribs) 18:41, August 1, 2006

No idea who wrote this or when (hint, sign your comments), but I couldn't find any evidence either so I removed it. --Vossanova o< 18:50, 11 November 2006 (UTC)[reply]
I have added the relevant information on that comment from the page history. — Aluvus t/c 22:15, 11 November 2006 (UTC)[reply]
The information of X2 3600+ were both added by me. According to Chinese websites I've searched, the S939 Manchester X2 3600+ CPUs were sold to PC manufacturers in China such as Lenovo and Tsinghua Tongfeng. The AM2 X2 3600+ were available in Hong Kong since August 2006. When I search for X2 3600+ OPN at http://www.amdcompare.com, however, the official website returned "No results found". An article in Chinese about release of AM2 X2 3600+ could be found at http://www.hkepc.com/bbs/hwdb.php?tid=651258&tp=amd-x2-3600-2&rid=651569 --Wrightbus 10:24, 16 November 2006 (UTC)[reply]
I've re-added the S939 X2 3600+ CPU as I found relevant information and clear pics form a Chinese website, http://www.pcpop.com/doc/0/146/146851.shtml . --Wrightbus 10:47, 16 November 2006 (UTC)[reply]

Brisbane 3600+[edit]

According to a CPU pricelist webpage from Asrock [1], it's reported that Brisbane 3600+ (ADO3600IAA5DD) exists. It could be added into the list when verified. --Wrightbus 07:43, 6 December 2006 (UTC)[reply]

  • Confirmed but the postfix is DL instead of DD.--Wrightbus 21:31, 22 January 2007 (UTC)[reply]
  • Confirmed DD exists along with DL.--Wrightbus 04:48, 20 February 2007 (UTC)[reply]

Unlisted Part Numbers[edit]

I have just bought an Athlon 64 CPU which I think is a Venice core but the part number ADA3200DAA4BW is currently not in this list. An unofficial source on the AMD Forums suggests this is a new core stepping that uses less lead to comply with EU legislation. List of new part nos. : ADA3800DAA4BW AMD Athlon 64 3800+ 2400Mhz Socket 939 1.35-1.40 V 89W ADA3500DAA4BW AMD Athlon 64 3500+ 2200Mhz Socket 939 1.35-1.40 V 67W ADA3200DAA4BW AMD Athlon 64 3200+ 2000Mhz Socket 939 1.35-1.40 V 67W ADA3000DAA4BW AMD Athlon 64 3000+ 1800Mhz Socket 939 1.35-1.40 V 67W ADA4800DAA6CD AMD Athlon 64 X2 Dual-Core 4800+ 2400Mhz Socket 939 1.35-1.40 V 110W ADA4600DAA5CD AMD Athlon 64 X2 Dual-Core 4600+ 2400Mhz Socket 939 1.35-1.40 V 110W ADA4400DAA6CD AMD Athlon 64 X2 Dual-Core 4400+ 2200Mhz Socket 939 1.35-1.40 V 110W ADA4400DAA5CD AMD Athlon 64 X2 Dual-Core 4200+ 2200Mhz Socket 939 1.35-1.40 V 89W ADA3800DAA5CD AMD Athlon 64 X2 Dual-Core 3800+ 2000Mhz Socket 939 1.35-1.40 V 89W SDA3400AIO3BX AMD Sempron 3400+ 2000Mhz Socket 754 1.40V 62W SDA3300AIO2BX AMD Sempron 3300+ 2000Mhz Socket 754 1.40V 62W SDA3100AIO3BX AMD Sempron 3100+ 1800Mhz Socket 754 1.40V 62W SDA3000AIO2BX AMD Sempron 3000+ 1800Mhz Socket 754 1.40V 62W SDA2800AIO3BX AMD Sempron 2800+ 1600Mhz Socket 754 1.40V 62W SDA2600AIO2BX AMD Sempron 2600+ 1600Mhz Socket 754 1.40V 62W SDA2500AIO3BX AMD Sempron 2500+ 1400Mhz Socket 754 1.40V 62W

Added Steppings[edit]

Perhaps someone will want to move the steppings from each core's header, but I thought it was just as important as the size of the CMOS process used to make the chip, so I put it in the header. At least, if you would rather the stepping not be in the header, I ask that you still keep the steppings on this page, as I think they were the only major piece of information lacking on this page, and I have wanted th esteppings to be listed on this page with the corresponding cores for a while now...

Also, I did not put the steppings for every processor because I wasn't aware of what the steppings were for the mobile processors nor for some of the Energy Efficient and/or Small Form Factor cores.

And one last thing: I got all of my information from the Athlon 64 and Athlon x2 wiki pages.

missing processor?[edit]

I have an Athlon 64 3000+ 1800MHz 1024KB cache on socket 754 (89W). (cpu family 15, model 4, stepping 8)

it seems it's not listed —The preceding unsigned comment was added by 151.52.92.106 (talk) 00:14, 9 December 2006 (UTC).[reply]

X2 3800+ EE SFF[edit]

According to a Chinese website (http://tech.sina.com.cn/h/2006-10-10/0016120235.shtml) , the Windsor X2 3800+ EE SFF with OPN ADD3800IAA5CU exists, along with postfix IAT5CU. --Wrightbus 05:18, 11 December 2006 (UTC)[reply]

Windsor? Brisbane? Huh?[edit]

It would be helpful for an explanation of the difference between cores, or a proper link to one. All the information I was able to find, Wikipedia or elsewhere, was a difference between 90nm and 65 nm manufacturing, which I happen to know (if I'm not mistaken >_>) makes for more transistors, but "so what?" performance-wise, is a question I'm still left with. Cheezmeister 20:33, 6 March 2007 (UTC)[reply]

Since this is just a list, it isn't really the place to describe and compare the differences between the processors in length. The proper link would be Athlon 64 X2, but that also compares the cores as bulleted data, not quite in layman's terms. That could stand to be changed. --Vossanova o< 21:25, 6 March 2007 (UTC)[reply]
Brisbane is a die shrink of Windsor, with a few differences. It has higher latency when accessing memory and specifically accessing L2 cache, which puts performance a touch below the older parts. Being on a smaller process does not mean it can have more transistors (in fact, it has fewer than some Windsors, due to L2 cache size), but that with a given number of transistors each die can be smaller. So more dice fit on one wafer, and manufacturing costs are lower. Because the transistors are smaller, the total heat output is also reduced a bit. — Aluvus t/c 23:54, 6 March 2007 (UTC)[reply]
Right you are. Belated thanks for the clarification and that link there, very helpful it was.Cheezmeister 17:08, 3 April 2007 (UTC)[reply]

ADA3600DAA4BV[edit]

Never heard of this before. Guess it was an OEM product. Came across one pic of it here [2]. Was it only sold in a certain market places? E.g. China Nil Einne 07:57, 25 May 2007 (UTC)[reply]

Actually I see this discussion has already taken place above. Nil Einne 17:16, 2 June 2007 (UTC)[reply]

Ask for help (about 5000CU, 5000CS and 5200CS)[edit]

Do 5000CU, 5000CS and 5200CS both have 65W versions? I found many websites listed these processors, but I still couldn't find any relevant pics showing their OPNs.

OPNs need to verify:

  • ADO5000IAA5CU
  • ADO5000IAA5CS
  • ADO5200IAA6CS

--Wrightbus 15:31, 10 June 2007 (UTC)[reply]

Mobile Athlon 64 X2 info missing?[edit]

Athlon 64 X2 TK series, where are they? --202.71.240.18 09:59, 15 August 2007 (UTC)[reply]

They are listed here for whatever reason. — Aluvus t/c 23:08, 15 August 2007 (UTC)[reply]

LE-1600/20[edit]

I didn't add the LE-1600 and 1620 yet. AMD has conflicting sets of info on their pages. The price list says they're 65nm but the comparison chart says they're 90nm. It also says they have F3 stepping, which would make them more likely Orleans than Lima. When the two pages have the same info, or if someone gets one of these, feel free to add them. --Vossanova o< 17:33, 8 October 2007 (UTC)[reply]

The postfix "DH" indicates that they belong to "Orleans" with stepping F3. --Wrightbus 11:17, 13 November 2007 (UTC)[reply]
Actually they are Windsor dual-core with one core disabled. There's also a DG OPN for the same core as Socket F (embedded) Opteron. --Denniss (talk) 14:13, 11 May 2008 (UTC)[reply]

Brisbane 5600+[edit]

I added this processor last night and now somebody have removed it. My source is here: http://www.hardwareschotte.de/hardware/preise/proid_9142433/preis_AMD+Athlon+64+X2+5600+ADO5600DDBOX I dont know how to add links yet. Hope somebody can get more confirmation on this and add it to the list. —Preceding unsigned comment added by 213.161.191.155 (talk) 15:46, 30 October 2007 (UTC)[reply]

Windsor X2 6400+[edit]

The ADX6400IAA6CZ part number refers to the standard multiplier-locked version of the Windsor core X2 6400+, not the unlocked Black Edition. I have one and it came in a green and white box complete with heatsink/fan. The IHS is etched with the ADX6400IAA6CZ part number and the box barcode has the ADX6400CZBOX part number. Unless someone can provide the accurate part number for the Black Edition (which undoubtedly does exist) I propose deleting the words "Black Edition" from the X2 6400+ entry in the table of Windsor processors. 83.104.249.240 (talk) 21:35, 31 May 2008 (UTC)[reply]

ADX6400IAA6CZ IS the part number of the 6400+ Black Edition, there's no locked standard edition. ADX6400CZBOX is the boxed partnumber and ADX6400CZWOF is the boxed partnumber without fan/heatsink. See also here. --Denniss (talk) 21:52, 31 May 2008 (UTC)[reply]
That's very interesting but there is most definitely a locked standard edition because I have one. The multiplier can not be changed from 16×. Thank you for the Tom's Hardware reference. Unfortunately, I'm not a native German speaker, though I understand enough to see that on page 6 the chip was overclocked simply by raising the clock frequency from its nominal 200 MHz. If I had access to a sample with an unlocked multiplier I would overclock it by changing the multiplier instead of the clock. 17× would give 3.4 GHz and 17.5x would give 3.5 GHz. That is, after all, the point of the Black Edition. Furthermore, the Black Edition is sold without a cooler as the AMD stock one would be inadequate when overclocked. My sample, on the other hand, is only intended to be used with the supplied cooler. The warranty specifically precludes the use of any other cooler than the one provided. I've found references to both the Black Edition and the one I have on AMD's site but the information is scant and the More Product Info link is self-referential. 83.104.249.240 (talk) 04:06, 1 June 2008 (UTC)[reply]
Look at the OPN, they just replaced WOF with BOX and the rest was left intact. The only difference between the two is the first without heatsink/fan and the latter with. Neither of them has a multiplier locked but only open downwards, raising the multiplier is not possible contrary to the newer Black Edition whose are open upwards and downwards. BTW - don't trust anything on the AMD website unless somehow confirmed by others or written innofficial documents - they have just too many errors on their site (5000+ Black Edition is shown as G1 instead of the correct G2, the AthlonX2 Business class are shown as AM2+ but they are AM2 and many other errors). --Denniss (talk) 14:57, 1 June 2008 (UTC)[reply]
OK. Thanks for the information. So I have a Black Edition processor but no black box to prove it. It's more than a little disappointing that AMD's own site can't be trusted! Also may I suggest that the term Black Edition be defined somewhere in the article? I'm very confused as to what it actually means now. I'm also at something of a loss to understand the point of having an unlocked multiplier which can only be changed downwards – far better just to enable Cool'n'Quiet, I would have thought. How is that meant to appeal to enthusiasts? Also, the one I bought was cheaper than the one in the black box without the fan. Seems as though the black box is more valuable than the fan! Crazy. 83.104.249.240 (talk) 23:33, 3 June 2008 (UTC)[reply]
Black Edition usually means it's something special for enthusiast gamers/overclockers. Yours was the first one but somehow limited in the multiplier control (only downwards). All newer BE have multi free in both directions. Changing the multiplier downwards is good if you want to overclock via the 200 MHz HT base clock but already reached the OC limit of your CPU: just adjust your multi downwards and look what's your memory capable of by raising even more. --Denniss (talk) 00:50, 4 June 2008 (UTC)[reply]

Brisbane 4400 ADO4400IAA5DU[edit]

Here's another mystery OPN if anyone has info on it. Ham Pastrami (talk) 20:10, 22 February 2009 (UTC)[reply]

And another one:ADO5000IAA5DU. No real idea what this is, maybe they fell off the Turion production line. http://www.asrock.com/support/cpu.asp?s=AM2 lists them as G2 with 65W TDP. --Denniss (talk) 11:22, 23 February 2009 (UTC)[reply]


2650e ?[edit]

Tigerdirect is selling Athol 64 2650e based systems... the JA wikipedia has it listed as a Lima type system, if so can someone add it? Hobart (talk) 19:34, 27 July 2009 (UTC)[reply]

Linux /proc/cpuinfo vs. cpuid; memory support[edit]

There really should be a paragraph (or article) detailing the algorithm to convert between Linux's decimal(?) cpuinfo family and model identifiers and the named steppings (presumably identifiable by 'model' when something noteworthy has changed?).

In relation to this, it appears that later S939 chips support RAM based on 128Mx4 chips and earlier ones don't; some demarcation between the two would be very useful and is likely what many visitors to this and the main Athlon 64 article are looking for. I was unpleasantly surprised after assuming a "Family 15, Model 47" S939 Sempron could eat the cheap RAM... no such luck, unless it was a BIOS issue.—Preceding unsigned comment added by 155.212.34.122 (talk) 20:56, 5 January 2010 (UTC)[reply]

ADA3400DAA4BY[edit]

750n3 Just wanted to say that the ADA3400DAA4BY has a spot in the page where it says these processors were only available in HP and IBM systems. This fact is wrong. They were also available in Emachines' T6420 Media center PC. Back then when this PC was released Emachines was owned by Gateway which was the year 2002. — Preceding unsigned comment added by 32.165.85.22 (talk) 19:32, 29 May 2011 (UTC)[reply]

External links modified[edit]

Hello fellow Wikipedians,

I have just added archive links to 4 external links on List of AMD Athlon 64 microprocessors. Please take a moment to review my edit. If necessary, add {{cbignore}} after the link to keep me from modifying it. Alternatively, you can add {{nobots|deny=InternetArchiveBot}} to keep me off the page altogether. I made the following changes:

When you have finished reviewing my changes, please set the checked parameter below to true to let others know.

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.—cyberbot IITalk to my owner:Online 06:25, 10 February 2016 (UTC)[reply]

8th or 9th Generation?[edit]

The lead states that the Athlon 64 CPUs are 8th generation, but the main article it links to (Athlon 64) says they're 9th generation. And then, later on, the main article says they're 8th generation. Could someone knowledgeable please clarify this?SPA5CE! talk about it 13:42, 26 October 2023 (UTC)[reply]