amd 90nm processors when?

S

stimlaphine

n00b
Joined
Jul 11, 2004
Messages
45
When will the athlon 64 winchester chips based on 90nm be released? And what new features/capabillities will they have over their 130nm predeccesor? And how much can i expect to pay for them when they are released?
 
The mobile equivalent to the Winchester, Oakville, has already begun shipment to OEMs and ODMs worldwide. I would estimate availability of Winchester parts to begin in the October, November timeframe.

As for features, the Winchester will add the following features to the feature set:

full SSE3 instruction set implementation, with the exception of the HyperThreading-specific instructions
improved hardware data prefetch mechanism
more write-combine buffers
improved on-die memory controller
on-die thermal throttling (a la Pentium 4)
Black Diamond low-k dielectric process
reduced execution times on some instructions

This should all account for about a 5% improvement at any given clock frequency, when compared to a Newcastle Athlon 64.

Don't know about pricing, but I think that it will be better than what we're seeing now.
 
With a smaller feature size, AMD can fit more chips on a wafer. This helps make things cheaper. So, if yields are good, it's not at all strange to expect price cuts. Yield rate is the big variable here.
 
xonik said:
The mobile equivalent to the Winchester, Oakville, has already begun shipment to OEMs and ODMs worldwide. I would estimate availability of Winchester parts to begin in the October, November timeframe.

As for features, the Winchester will add the following features to the feature set:

full SSE3 instruction set implementation, with the exception of the HyperThreading-specific instructions
improved hardware data prefetch mechanism
more write-combine buffers
improved on-die memory controller
on-die thermal throttling (a la Pentium 4)
Black Diamond low-k dielectric process
reduced execution times on some instructions

This should all account for about a 5% improvement at any given clock frequency, when compared to a Newcastle Athlon 64.

Don't know about pricing, but I think that it will be better than what we're seeing now.
Click to expand...

wow. sounds like a steal if you can get them for prices cheaper than a newcastle and still get all those improvements.
 
We'll see what happens when we get there.

Cautionary Statement

This release contains forward-looking statements, which are made pursuant to the safe harbor provisions of the U.S. Private Securities Litigation Reform Act of 1995. Forward-looking statements are generally preceded by words such as "plans," "expects," "believes," "anticipates" or "intends." Investors are cautioned that all forward-looking statements in this release involve risks and uncertainty that could cause actual results to differ materially from current expectations. Risks include the possibility that the company will not achieve its current product and technology introduction schedules; that adoption of AMD64 products, including upcoming dual-core processors, by tier-one OEMs and other customers will not occur as expected; that solutions providers will not timely provide the infrastructure, including operating systems and applications, to support the company's AMD64 technology; and that Intel Corporation's pricing, marketing programs, product bundling, new product introductions or other activities targeting the company's processor business will adversely affect the company's sales plans. We urge investors to review in detail the risks and uncertainties in the company's Securities and Exchange Commission filings, including but not limited to the Annual Report on Form 10-K for the year ended December 28, 2003, and the Quarterly Report on Form 10-Q for the quarter ended March 28, 2004.
Click to expand...
 
GodsMadClown said:
We'll see what happens when we get there.
Click to expand...

Standard disclaimer. I can't figure out what you meant by posting it.
Are you disputing that parts are already shipping or that the desktop parts won't ship?
 
I'm not disputing that AMD is shipping 90nm mobile parts. I'm saying that we just don't know what the yeilds will be for 90nm desktop parts. Sorry for the silliness.
 
No problem...just didn't know if you had some info that we hadn't heard yet.
Yields will almost certainly be problematic if history is any indication.
Thanks for clearing that up. :)
 
Well Im gonne Bump this just to keep it in focus. I really want to know when the first person who is willing to tell about it get his/her hands on one... ( And naturally also what he/she managed to do with it)
:)
 
GodsMadClown said:
Yield rate is the big variable here.
Click to expand...

Yield rates will be much higher than current A64 rates. The A64 has a HUGE die >150mm^2 on 130nm. On 90nm i read that the die is about 110mm^2 which is much closer to the optimum die size (~1cm^2). Yield issues are mainly caused by inconsistencies in the silicon (defective areas) since we cannot make 100% pure silicon. The shrink in the die area will dramatically improve yield rates over current A64 rates.

Defective silicon is not the only cause of low yeilds, but its is a MAJOR factor.
 
TheArchitect said:
Yield issues are mainly caused by inconsistencies in the silicon (defective areas) since we cannot make 100% pure silicon. The shrink in the die area will dramatically improve yield rates over current A64 rates.

Defective silicon is not the only cause of low yeilds, but its is a MAJOR factor.
Click to expand...

Later in the life of a process, yes.
Early on, the defect rate can easily be 100 times (per unit volume) that of a very mature process however. The vast majority of that delta is due to new processes needed alot of tweaking to get the electrical, chemical and mechanical characteristics right. The majority of that is done internally, AMD has been producing 90nm samples for in house use for atleast 9 months now. But it'll probably be another 6-9 months before 90nm has a defect (per unit volume) on par with 130nm.
We'll still see more dies per waffer at the outset of full 90nm production, but that's not a direct reflection on the defect rate.
 
xonik said:
The mobile equivalent to the Winchester, Oakville, has already begun shipment to OEMs and ODMs worldwide. I would estimate availability of Winchester parts to begin in the October, November timeframe.

As for features, the Winchester will add the following features to the feature set:

full SSE3 instruction set implementation, with the exception of the HyperThreading-specific instructions
improved hardware data prefetch mechanism
more write-combine buffers
improved on-die memory controller
on-die thermal throttling (a la Pentium 4)
Black Diamond low-k dielectric process
reduced execution times on some instructions

This should all account for about a 5% improvement at any given clock frequency, when compared to a Newcastle Athlon 64.

Don't know about pricing, but I think that it will be better than what we're seeing now.
Click to expand...


how long before the mobile variant reaches an etailer?
 
The Mobile Athlon 64 produced under the 90 nm process has already shipped to OEMs. I can't name any retailers off the top of my head, but I bet someone has them. If not, they will come very soon.
 
xonik said:
The Mobile Athlon 64 produced under the 90 nm process has already shipped to OEMs. I can't name any retailers off the top of my head, but I bet someone has them. If not, they will come very soon.
Click to expand...

cool... and just to reiterate, they will have all the neat stuff the desktop 90nm will have?
 
Yes, just a different codename, and lower input voltage and power dissipation.
 
( ´|` up ´|` )
And where are they now?
And has any1 acctually had their mitts on one?
 
xonik said:
The mobile equivalent to the Winchester, Oakville, has already begun shipment to OEMs and ODMs worldwide. I would estimate availability of Winchester parts to begin in the October, November timeframe.

As for features, the Winchester will add the following features to the feature set:

full SSE3 instruction set implementation, with the exception of the HyperThreading-specific instructions
improved hardware data prefetch mechanism
more write-combine buffers
improved on-die memory controller
on-die thermal throttling (a la Pentium 4)
Black Diamond low-k dielectric process
reduced execution times on some instructions

This should all account for about a 5% improvement at any given clock frequency, when compared to a Newcastle Athlon 64.

Don't know about pricing, but I think that it will be better than what we're seeing now.
Click to expand...

So all those features will be in the new 90 nano s939 3500'+'s?? I didn't know they were having SSE3 implemented. Can anyone link me to a new review of the 90 nano counterparts which talks about these new features? Anandtech didn't make any mention of them in their review.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2242&p=1
 
burningrave101 said:
So all those features will be in the new 90 nano s939 3500'+'s??
Click to expand...
No new features or architectural changes have been implemented in the new 90nm A64. There was a lot of false information doing the rounds before the 90nm parts appeared and it was all wrong. The 90nm A64 is just the same as the older 130nm part but on a smaller feature size and therefore a smaller die size. Maybe they'll make some changes in a future stepping but the 90nm chips around right now are no different and no faster (clock-for-clock) than the older parts.

But I still think they look like an excellent choice if you're thinking about an A64 purchase. Slightly cooler running and slightly better overclocks than 130nm A64. Ok, that doesn't sound that impressive but compared to the horrors Intel produced on 90nm it's actually pretty good. :)
 
JNavy89GT said:
I disagree
I have tested these parts side by side and found performance improvements.
http://www.madshrimps.be/?action=getarticle&articID=230

there is no SSE3 as of yet. Next core stepping will have that. I did find good gains in gaming though as you can see in my review :)
Click to expand...

Nice review JNavy89GT. So your saying SSE3 will be in the next stepping huh? I wonder how long that will be? I'm gonna be putting togethor a new system around Christmas.

The Anandtech review also showed consistent wins by the 90 nano s939 chips against the 130 nano versions of the same speed so there is definitely a slight performance difference.
 
You must log in or register to reply here.
Back
Top