backtop


Print 67 comment(s) - last by Amiga500.. on Apr 25 at 9:31 AM

Giuseppe Amato gives another overview on the high points of AMD's next-generation CPU architecture

This article was first published in German on K-Hardware.de.

Yesterday, AMD held a press presentation in Munich, Germany to update journalists about its upcoming K10 processor. AMD's Giuseppe Amato, Technical Director Sales and Marketing EMEA, had a few minutes to talk about the architecture at length. This architecture, previously dubbed K8L by Henri Richard -- now publically called K10 -- is scheduled to be AMD's first monolithic quad-core design.

The integrated memory controller (IMC) will get a few new features in the K10 core. When utilizing multiple memory modules, along with proper BIOS implementation and mainboard routing, the IMC can access memory in 64-bit channels (72-bit if you use ECC). This way it is possible to read and write data simultaneously, or improve efficiency for irregular access patterns which increasingly occur in a quad-core environment. This feature is available on AM2+ and F+ boards; on "old“ socket AM2 and F boards the usual 128-bit dual-channel mode is available.

Due to split power planes, the IMC can be clocked down independently of the CPU cores, along with reduced voltage. This also enables CPU overclocking without touching the memory frequency, something that may appeal to enthusiasts. These features are again dependent on Socket AM2+ and F+ platforms.

Amato explained how the quad-core design benefits from the internal crossbar switch the backbone of communication inside the K10 CPU. With Intel's current quad-core design there are cases where data needs to travel over the FSB -- in AMDs case all inter-CPU communication takes place on-die.

The crossbar switch of the K10 core is already prepared for up to 8 cores, Amato boasted. Amato wouldn't give even a vague timeframe for market availability of such a CPU, though he indicated the company is prepared for whatever the market demands. Amato made clear that octo-core is far away in the future – Shanghai will not get 8 cores.

K10 will introduce a shared L3 cache while the individual cores have dedicated L1 and L2 caches. As long as requested data lies in L1, it can be directly loaded. This also works if the data lies in the L1 cache of another core, in which case the communication works via the crossbar switch. In case requested data resides in the L2 cache, it will be loaded to L1 and then invalidated in L2 as AMD has an exclusive cache design. The L3 Cache, however, is not exclusive, but allows for a shared bit to be set. If a core loads data marked as shared, it will reside in the L3 cache and can be fetched by other cores as well.

Amato also mentioned an array of power saving measures which, in sum, allow AMD to deliver a quad-core CPU in the same thermal envelope as today’s dual-core CPUs.

K10 adds the capability of independently clocking all the CPU cores. In current K8 processors (and Intel's Core 2 generation), all cores are clocked at the same level all the time -- the P-state can only be changed synchronously. In case of a compute-intensive single-threaded process, all cores must run on the highest level P-state. On K10-based CPUs, the idle cores could be switched to the lowest P-state, while others are in different states, depending on load.

This feature could possibly be abused by overclockers to overclock a single core above the specified levels. Amato clarified that AMD doesn't endorse overclocking, but acknowledges there are people interested in that. In a warranty case, AMD could detect PLL programmings out of spec which would deny the warranty. The new cores, however, have new thermal sensors, to improve overheating protection.

Amato closed the session by mentioning Shanghai as a successor to Barcelona in the server space for 2008. Shanghai will be an improved quad-core architecture, which is supposed to be socket-compatible with current Socket F platforms. Roadmaps available to DailyTech revealed Shanghai is a 45nm quad-core CPU featuring 6MB of L3 Cache.



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

umm
By SLEEPER5555 on 4/16/2007 6:37:23 AM , Rating: 0
is there no K9? did they just skip it?




RE: umm
By chaos386 on 4/16/2007 7:24:33 AM , Rating: 5
There was a K9, but its performance was a bit of a dog, so they decided to skip it.


RE: umm
By sviola on 4/16/07, Rating: 0
RE: umm
By cochy on 4/16/2007 4:01:24 PM , Rating: 2
quote:
There was a K9, but its performance was a bit of a dog, so they decided to skip it.


I think I got that exact same response when I asked that question.

I still have another question though. I was under the impression that Barcelona was based on a "K8L" core. Now AMD is talking about K10 cores. Can anyone clarify for me?


RE: umm
By johnsonx on 4/16/2007 5:07:02 PM , Rating: 2
My impression is that K8L has simply been renamed to K10. The 'original' K9 and K10 projects, which were new architectures largely distinct from the K8, have been scrapped (which is not to say that nothing was learned, or that nothing from those projects made it into the 'new' K10). The upcoming evolution of the K8 architecture that has been referred to as the K8L may or may not have been ever referred to as such by AMD internally, depending on who you listen to. The public rename to K10 is clearly meant to describe the significant differences from K8; K8L hardly brings to mind a significantly improved core. K9 was skipped for the obvious reasons: it isn't just that it sounds similar to "canine", K9 and K-9 are often used in place of the word.

Some sources now suggest that K9 is the internal code name for the dual-core K8, which may well be true now, but that doesn't mean K9 wasn't originally for a different project. This may be nothing more than a non-doggy explanation for calling the 'new' K8 architecture K10.

Just for the record, I'm not claiming to 'know' any of this, and I have no inside info. I've just read a fair amount about this over the past few years, and the above is the picture I've formed after separating the wheat from the chaff.


RE: umm
By johnsonx on 4/16/2007 6:09:46 PM , Rating: 2
Marcus Pollice confirms some of what I suggested here:

http://www.dailytech.com/article.aspx?newsid=6918&...

I'd often suspected that K8L was actually just a low-power mobile K8, and that the use of that term to describe the next-gen K8 was a media mistake.


RE: umm
By paydirt on 4/16/2007 7:25:01 AM , Rating: 3
yes, AMD is a humane and animal friendly company. They are not in the business of manufacturing or disposing of K9's.


RE: umm
By clayclws on 4/16/2007 8:18:24 AM , Rating: 1
They removed "K9" because its ounded like "canine" ? Been reading about it across the internet, but is it really true?


RE: umm
By paydirt on 4/16/2007 9:02:10 AM , Rating: 4
Sounds like cute, cuddly, furry friend in the U.S. and potential meal in parts of Asia. Too confusing to consumers.

The K10. Now with improved packet sniffing and instruction prefetching and shares its cache of toys.


RE: umm
By dnd728 on 4/16/2007 2:09:18 PM , Rating: 2
I guess K12 and prolly K13 are out of the question too...


RE: umm
By Fritzr on 4/21/2007 6:26:34 AM , Rating: 2
Doctor Who is likely the reason for skipping K9. K9 the robotic dog was one of his many companions :)


RE: umm
By clemedia on 4/16/07, Rating: -1
RE: umm
By Griswold on 4/16/07, Rating: 0
RE: umm
By Daven on 4/16/2007 11:45:04 AM , Rating: 3
There is some confusion surrounding the K9 name. Here are some links for reference:

http://www.theinquirer.net/default.aspx?article=37...
http://en.wikipedia.org/wiki/AMD_K9

The wikipedia entry says that K9 has been officially confirmed as the dual core version of K8 (the original AMD 64 architecture). K10 is definitely the upcoming barcolona architecture. K9 was also supposed to be an 8-issue core that was canceled at some point.

Read the articles and make up your own mind for whatever it's worth. However, the answer as Griswold stated as 'No' may not be correct. And please don't respond No again unless you have at least a link to something stating that K9 never existed or was cancelled or whatever.


RE: umm
By Daven on 4/16/2007 12:30:57 PM , Rating: 2
Here is another link from Cnet News.com stating that AMD is doing away with K nomenclature. The article also states that dual-core processors released in 2005 are referred to as K9 processors lending more credibility that K9 processors are dual-core K8's.

http://news.com.com/AMD+hatches+new+naming+plan+fo...


RE: umm
By bldckstark on 4/16/2007 12:32:10 PM , Rating: 2
No.


RE: umm
By Marcus Pollice on 4/16/2007 5:36:48 PM , Rating: 2
Yes there were all these stories about K9 and K10 cancelled, there being no K9, K9 as a name for dual-core K8 and K8L as a name for Barcelona-architecture now dubbed K10.

Here are the facts:
K8L = K8 Low Power = Turion 64
K10 is now apparently the official name and it fits way better than K8L given all the additions. AMD employees used it during the presentation themselves.


RE: umm
By Targon on 4/16/2007 11:45:52 PM , Rating: 2
If you think about it, K9 brings to mind a dog. Now, if you say that a chip performs like a dog, that would be seen as a negative thing, so it's pretty clear why AMD marketing would avoid letting that name be used, even internally.

So, even if dual-core is what WOULD have been K9, the name would never be used for the above reasons. K10 on the other hand works as a code name.


RE: umm
By GlassHouse69 on 4/17/2007 3:05:13 AM , Rating: 2
yes. however, they stopped that info from comming out as k9 was a stupid marketing name.

also, teh original k9 details were off a bit from the projected k9 so I think this also has something to do with it.

k9 sux compared to dualcore.


RE: umm
By SmokeRngs on 4/18/2007 11:28:00 AM , Rating: 2
I will state this from the beginning, I have no proof of any reason why they would have skipped K9 and gone straight to K10. I believe it's a mix of reasons already mentioned, though.

I'm sure there were many architecture changes made during design that may have been the intention to call K9. However, due to the differences between the expectations of K9 and what is now called the K10, AMD may have decided to skip K9 and go to K10.

K-9 is obviously a reference to dog in the US (and possibly other parts of the world) which is a reference most marketing departments would want to avoid.

"Skipping" a product number is nothing new. Notice how Arctic Silver mysteriously went from AS3 to AS5. It is rumored that was because 4 is an unlucky number in parts of Asia if my memory serves me correctly. On a tangent, how many of you have seen audio CDs that mysteriously have a missing track 13 or track 13 is nothing but silence. It's a superstition. I know of a couple of music performers that have admitted that is the reason they skipped that track number.

A combination or mixture of reasons I have stated or others have given could have resulted in this "skipping" of a product number.

Does any of this actually mean anything? Not a damn bit. The architecture is the same no matter what number or name they give it.


RE: umm
By Marcus Pollice on 4/19/2007 8:41:27 AM , Rating: 2
Especially since architecture codenames are not even mentioned in marketing, but only in technical discussions and publications where such stupid associations with no technical merit are not really common. Leaves only fanboys and those make up bad stuff also on good names.

So I guess the K9 mystery will remain unsolved.


"And boy have we patented it!" -- Steve Jobs, Macworld 2007

Related Articles
AMD 2007 Server Roadmap
December 14, 2006, 9:47 AM













botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki