AMD's K9 chip is now the K10 chip????

[kenneth9265_3]

Here is an interesting article from the inquirer.....

In the finest Intelesque naming traditions, the dual core K8 is now K9, and the next real core is K10. If the addition of MMX to the PII core rated an extra I, the addition of an extra K for another whole core is a forgivable PR gaffe.

I know.... I know.....The inquirer.


http://theinquirer.net/?article=18967

 

[Guden Oden]

Who cares if it's K9 or K10 or K471190042735? AMD isn't using K-whatever as product names anymore, haven't done so for half a decade.

 

[3dilettante]

I hope they do a bit more with the dual core K8's than just paste another one on the unused port that has always existed on the crossbar.

If it's just that, then I wouldn't call it a K9. Then again, it's just a code-name, and AMD can do whatever fancy gymnastics with that as it wants.

It wouldn't be the first time AMD has done this, as the original K8 design that was being kicked around sounded to be a much more dramatic departure from K7.

 

[akira888]

Do you know of any more removed changes from the K8 design other than the non-stack based floating-point unit (TFP)? Even as late as 1H 2003 IIRC AMD still was making noises to the effect that it would be added in eventually...

Although with SSE2 support it isn't such a big deal I guess.

 

[3dilettante]

http://www.chip-architect.com/news/2001_10_02_Hammer_microarchitecture.html

This is rather old, but it is a nice sampling of the speculation surrounding the K8 program at the time. There was discussion that the supposed preliminary K8 wound up having elements being passed on to K9, which is now K10.

Considering the massive changes being bandied about, I can see why they may have delayed it. Things like a L0 cache as well as much wider instruction issue width along with a look-ahead unit were a significant departure from K7.

However, much of the speculation is based on patents that seemed to match the development time of K8 like some Athlon patents matched the development time of K7. There were apparently several preliminary K8s that were junked before the loss of TFP.