backtop


Print 31 comment(s) - last by coldpower27.. on Aug 17 at 12:47 AM


AMD's official pricing on the 24 new processors
AMD launches revision F Opterons for real

AMD has officially announced its revision F Opteron 8000, 2000 and 1000 series of server and workstation class processors to take on Intel’s previously released Xeon Woodcrest processors. AMD Opteron revision F processors are architecturally similar to AMD’s socket AM2 consumer processor architecture. Revision F processors introduce DDR2 to AMD’s Opteron—a feature Intel introduced a few years back to its Xeon lineup. Opteron revision F processors also have a new socket too; this time around AMD has abandoned the traditional pin-grid array layout for a land-grid array layout originally used by Intel with its LGA775 socket. AMD’s land-grid array socket has 1,207 pins.

The new AMD Opteron lineup includes six speed grades including x210, x212, x214, x216, x218 and x220 which are 1.8, 2.0, 2.2, 2.4, 2.6 and 2.8 GHz respectively.  All models are dual-core only. Three different models are available too, including regular, HE and SE variants. AMD Opteron HE variants are low powered models that consume 55 watts of power. Opteron SE models are the flagship and have a 125 watt TDP. All new AMD Opteron processors have 1MB L2 cache per core.

Distributor pricing for the new AMD Opteron 2200 series starts at $252 for the Opteron 2210 and tops out at $1153 for the Opteron 2220 SE. AMD MSRPs are slightly higher, although vendors are able to sell under the MSRP so long as they do not undercut the distributor price.  Opteron 8200 series start out at $864 for the Opteron 8212 and tops out at $2621 for the Opteron 8220 SE. Pricing for the Opteron 1200 series have not yet been revealed. Availability is expected immediately as distributors have had revision F processors for weeks.


Comments     Threshold


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

Hm..
By bozilla on 8/15/06, Rating: 0
RE: Hm..
By Targon on 8/15/2006 7:55:32 AM , Rating: 3
With HyperTransport and a better multi-processor design, quad processor and up easily show that AMD is in the lead in terms of performance. Prices will also drop quite a bit once AMD finally starts to release 65nm based parts.

If AMD can compete(even if they arn't in the lead) with a 90nm fab process, that will allow them to hold on until 65nm and K8L are finally released.

Another thing to consider is that if a system uses a 64 bit OS and applications, AMD's performance is actually a bit better than Conroe. The design improvements in K8L alone without going quad-core would be enough to catch up if not beat Conroe, so the real question EVERYONE should be asking is how long will it take before we see the new chips?


Not so fast
By DallasTexas on 8/15/06, Rating: -1
RE: Not so fast
By armagedon on 8/15/2006 8:56:03 AM , Rating: 3
God with his unquestionable thruth has just spoken !!! Everyone else, kneel.


RE: Not so fast
By redeyedfly on 8/15/2006 9:30:58 AM , Rating: 2
its because he is from dallas texas. everyone in dallas thinks they are gods.


RE: Not so fast
By shabodah on 8/15/2006 9:15:11 AM , Rating: 2
Go take your intel fanboyism over to tomshardware with the others. If you have any idea how processors work, you should know that bandwidth is the limiting factor of Intel's current platform. Because there is enough bandwidth for 2P servers, you don't see them run into the wall they do above 2P. Forget the terms "IMC, and Hyper-transport" for a moment and just read "more bandwidth." On the flip side, it sure would be nice of AMD could take more advantage of all the bandwidth they have (which they don't by far). Too extreme examples of design limitations.


RE: Not so fast
By DallasTexas on 8/15/06, Rating: -1
RE: Not so fast
By Viditor on 8/15/2006 10:30:12 AM , Rating: 2
quote:
what if Intel puts TWO 1066 FSB on their chipset?

It's still a bandaid...even with Intel's dual FSBs, they still have only a single pipe to the memory from the NB for both of them. So while BW to and from the NB is increased, true FSB BW isn't...


RE: Not so fast
By SilthDraeth on 8/15/2006 9:24:20 AM , Rating: 2
AMD is buying ATI.


RE: Not so fast
By Tsuwamono on 8/15/2006 9:34:55 AM , Rating: 1
Cream? what are you 4? Some AMDs beat the E6600 is SOME tests. There is no CREAMING going on. The A64s didnt CREAM the P4s. They were just faster. And conroe is merely faster then the older technology.

I dont see why all these fanboys are freaking out because New technology is beating old technology by a large margin in some cases. Grow up buddy. Go back to your ranch and milk some cows or something because you obviously dont belong on a tech site.


RE: Not so fast
By DallasTexas on 8/15/06, Rating: -1
RE: Not so fast
By Viditor on 8/15/2006 10:32:39 AM , Rating: 2
quote:
I'm your worst nightmare. I'll be sticking around this pathetic "tech site" for recreational purposes. It sharpens my slamming


I'd be happy if you'd just stop shouting...


RE: Not so fast
By RandomFool on 8/15/2006 10:41:15 AM , Rating: 2
Well aren't you just the coolest kid around.


RE: Not so fast
By Ringold on 8/15/2006 6:04:04 PM , Rating: 2
it's a shame one can't be voted below even -1. maybe that should tell you something, dallas, that each one of your posts was voted to oblivion. Perhaps its suggestive of, oh, I don't know, blatant arrogance?

I really should stop changing the threshold to -1. Really isn't worth it. you've been a intel fanboy troll in every thread I've seen you post in with the same blind enthusiasm as those nuts swearing science is full of lies and Adam and Eve ran with dinosaurs. Can't you intel boys just be happy Conroe is a nice improvement and currently the desktop champ?!


RE: Not so fast
By retrospooty on 8/15/2006 10:30:32 AM , Rating: 2
I dunno about that... Most industry indsiders agree that when it comes to the server side with 4 way or higher processers AMD will own Intel's Conroe/woodcrest based CPU's. This is due to a superior quad implementation and hypertransport link. Intel supposedly doesnt have anything to match it and will have bottlenecks.

In the end though, benchmarks of final shipping products will tell all.


RE: Not so fast
By coldpower27 on 8/17/2006 12:08:52 AM , Rating: 2
There is nothing to own in the 4P Segement as Intels' first product Tigerton for 4P+ servers isn't due for some time.

We will have to see how Tulsa holds up against these processors I don't expect them to win not against 2.8GHZ K8's. Though we will have to see won't we.


RE: Not so fast
By psychobriggsy on 8/15/2006 12:09:30 PM , Rating: 2
Isn't Tulsa based on Netburst? It's hard to see that as a threat.

Don't you mean Intel's Core 2 based quad-core processors, whatever their current codename is? Kentsfield for single processor, what was it for dual processor and multi processor systems?

HyperTransport is pretty good overall, and it, along with the on-die memory controllers, does allow AMD systems to scale very well for the price point. Currently this does make up for some of AMD's core weaknesses compared to Core 2.


RE: Not so fast
By coldpower27 on 8/17/2006 12:11:27 AM , Rating: 2
Clovertown is the 2P Quad Core variant based on 2 Woodcrests, while Tigerton is the 4P+ variant, though I don't have a timeline on that particular SKU.

The 2P and Desktop version are slated for Q4 2006.


RE: Hm..
By Blazin Trav on 8/15/2006 2:38:59 PM , Rating: 2
Don't forget that Woodcrest requires FBDIMM RAM which is far more expensive than DDR2.


RE: Hm..
By coldpower27 on 8/17/2006 12:47:01 AM , Rating: 2
Care to provide any proof that FB-DIMM's are far more expensive then ECC DDR2?


All other things aside...
By Regs on 8/15/2006 9:39:29 AM , Rating: 2
Seeing how the transition to DDR2 for desk tops was just a mere change in ram usage, how does this effect performance on a server? If the desk top saw a 1% increase why should the server edition be any better? All they did was add more RAM bandwidth to a core that doesn't use it.

AMD I'm sure will say this was not released to compete with Core, however they'll keep the prices at a premium. Talk about double talk.

F for Failure.




RE: All other things aside...
By Glavinsolo on 8/15/2006 10:05:27 AM , Rating: 2
Servers switched to DDR2 for pricing reasons. The price of DDR2 has crossed the threshold of where price for DDR was higher. Also Intel will be using DDR2 with a special bit in each DIMM that will allow you to run as many sticks of ram that you want without performance penalty.

A lot of you are comparing benchmarks which is fine. However, businesses hardly ever use AMD for file/print/email/application servers. They use them for VMWare boxes which with AMD a VMWare solution will perform nearly 20-30% better on average over Intel.



RE: All other things aside...
By psychobriggsy on 8/15/2006 12:15:30 PM , Rating: 2
DDR2 is only part of the platform change.

Socket F also has 4 HyperTransport links, although these won't be used until K8L which has the prerequisite 4 HyperTranport controllers (with link-splitting, for 8 links of half the width). I guess this is one reason why is has a vast 1207 lands.


Fix the Article
By Ard on 8/15/2006 1:26:08 PM , Rating: 2
There's an error in the article. The x220 models run at 3GHz, not 2.8GHz.

http://www.amd.com/us-en/Processors/ProductInforma...




RE: Fix the Article
By psychobriggsy on 8/15/2006 1:49:48 PM , Rating: 2
Why aren't they x222s then, instead of x220s?

I'd really like AMD to have 3GHz dual-core processors on 90nm as retail products, it would show that 90nm for AMD was pretty good actually. However I think they're meant to be 2.8GHz.


RE: Fix the Article
By Anh Huynh on 8/15/2006 4:59:59 PM , Rating: 2
Opterons for AM2?
By peternelson on 8/15/2006 4:30:20 PM , Rating: 2

I think I remember hearing that AMD planned to launch some opterons (1xx type) on the AM2 socket?

Is that still true? And if so, would these AM2 opterons work in "4x4" platform motherboards designed for FX-62?

Also hope we can see some reviews of dual socket F motherboards soon.

Just to throw in another point, AMD's HT based architecture in 4+ processor configurations has more capacity for I/O expandability since you just hand more chipsets and controllers off HT links on more than one processor. Intel still can't do that, irrespective of differences in processor speed.




RE: Opterons for AM2?
By PT2006 on 8/15/2006 5:22:38 PM , Rating: 2
Yes ad Yes.


What's telling...
By Viditor on 8/15/2006 10:54:45 AM , Rating: 2
Woodcrest was released a month ago, and the Socket F Opteron just today...and already the Opteron has far more design wins with OEMs!
This is what I was trying to explain in many previous threads...it will probably take until next year for Woodcrest to be qualified for mission critical servers.
2006 will still belong to AMD...




What a hell?
By bozilla on 8/15/2006 8:11:41 PM , Rating: 2
Where the hell is my initial post. Why was it deleted?




"If you look at the last five years, if you look at what major innovations have occurred in computing technology, every single one of them came from AMD. Not a single innovation came from Intel." -- AMD CEO Hector Ruiz in 2007

Related Articles













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