Woodcrest 6/25/06 product availability, Cloverton Q1'07

pxc

Extremely [H]
Joined
Oct 22, 2000
Messages
33,063
Intel is set to announce Core and Dual-Core WoodCrest on the 19th of June 2006, with the goods being delivered no later than 25th of June 2006.
http://www.vr-zone.com/print.php?i=3647

The prices of the Woodcrest Xeons is pretty reasonable for server processors, especially if you're planning on a 2 socket board:
$209 Xeon 5110 (1.6GHz/4MB L2/1066MHz FSB)
$256 Xeon 5120 (1.86GHz/4MB L2/1066MHz FSB)
$316 Xeon 5130 (2GHz/4MB L2/1333MHz FSB)
$455 Xeon 5140 (2.33GHz/4MB L2/1333MHz FSB)
$690 Xeon 5150 (2.66GHz/4MB L2/1333MHz FSB)
$851 Xeon 5160 (3GHz/4MB L2/1333MHz FSB)
 
Do we think the Conroe EE will accually be a Woodcrest?

Just wondering what they are going to put inside the new intel Powermacs....
 
Its the same die/core for both I'm sure, just like Pentium 4 & Xeons were usually the same exact thing.

Yes, there have been versions with extra cache added like the Gallatin core 2mb L3 & early P4EE. But these have the same 4Mb cache as the Conroe, so I'm pretty sure they're going be identical. Later on they might make a fancy 8Mb shared cache Xeon... or something... that would be REALLY kick ass if they turned that in to the Conroe EE!
 
Well, if the Conroe EE will be a Woodcrest in disguise, what's the point of getting it? The most expensive Xeon is 851 and all EE Pentiums always retail for at least 1000. So, what would be the advanatge of this Extre Edition Proc over a Woodcrest?
 
tdx said:
So, what would be the advanatge of this Extre Edition Proc over a Woodcrest?

Not having to buy a $400+ dual xeon motherboard and $$$$ FBDIMMs. Woodcrest is for the server market.
 
Since the next gen Intel chips are all based on essentially the same architecture, can we use Woodcrest's performance as an indicator as to whether or not Conroe is going to live up to it's hype?
 
gwai lo said:
Since the next gen Intel chips are all based on essentially the same architecture, can we use Woodcrest's performance as an indicator as to whether or not Conroe is going to live up to it's hype?


Not really.... they are 2 entirely different implementations of the same processor. FBDIMMs, Dual FSBs, etc... these will all sku the results. There are people benchmarking conroe on the web.... look for it.
 
Poncho said:
Not having to buy a $400+ dual xeon motherboard and $$$$ FBDIMMs. Woodcrest is for the server market.
So are all dual socket Woodcrest boards going to require FB DIMMs? I was wondering where AMD's dig in the K8L announcement ("FBDIMM where appropriate") came from. ;)

(edit) and FB DIMMs are still cheaper than RDRAM. :D
 
Okay, yeah I kind of guessed. Just wasn't entirely sure on how greatly the differences in the platforms would impact performance.

Yeah, I've seen benchmarks of Conroe...just wondering if people doubting would be possibly convinced with the release of Woodcrest (as it's a month ahead). Oh well.
 
pxc said:
So are all dual socket Woodcrest boards going to require FB DIMMs? I was wondering where AMD's dig in the K8L announcement ("FBDIMM where appropriate") came from. ;)

(edit) and FB DIMMs are still cheaper than RDRAM. :D


Yes, the Bensley Chipset is all FBDIMM. FBDIMM is, hopefully, going to become the standard for the server market and I would imagine that AMD's "FBDIMM where appropriate" is coming from the fact that they market their "server" hardware to the enthusist as well.... where the added cost of FBDIMMs would hurt their sales. Last part is just opinion FYI.
 
Poncho said:
Yes, the Bensley Chipset is all FBDIMM. FBDIMM is, hopefully, going to become the standard for the server market and I would imagine that AMD's "FBDIMM where appropriate" is coming from the fact that they market their "server" hardware to the enthusist as well.... where the added cost of FBDIMMs would hurt their sales. Last part is just opinion FYI.

Aren't FBDimms supposed to be much much faster than regular DDR2 though? Enough to make it worth the price premium?
 
NulloModo said:
Aren't FBDimms supposed to be much much faster than regular DDR2 though? Enough to make it worth the price premium?

Yes they are faster, as for if it's worth it... that's ENTIRELY up to the market and AMD targets a lot of "budget" builders which I would imagine is a reason that they'll release a platform that doesn't use FBDIMMs and one that does. With Intels server/workstation market this is less of an issue since performance is often more important than cost. Again... this is opinion.
 
If Intel really did cut the power consumption by 23%, i'm impressed. If i was Intel i would start doing another revision and hopefull cut the power down a bit more. And i would start boosting clock speeds. I was very supprised by the news of the new stepping of Woodcrest and it's positive effects.
 
Poncho said:
That's pretty accurate actually. :D Make no mistakes... Woodcrest is a BEAST. Whether or not it's successful is completely dependent on FBDIMMs. If it's another Rambus... yea, that wouldn't be good. :D

thats cool, literally :D
 
Poncho said:
That's pretty accurate actually. :D Make no mistakes... Woodcrest is a BEAST. Whether or not it's successful is completely dependent on FBDIMMs. If it's another Rambus... yea, that wouldn't be good. :D

Sweeeeeettttt!
 
Poncho said:
That's pretty accurate actually. :D Make no mistakes... Woodcrest is a BEAST. Whether or not it's successful is completely dependent on FBDIMMs. If it's another Rambus... yea, that wouldn't be good. :D
Are the numbers the exact numbers? And is this a cut across the board? I mean will Conroe also get a 23% reduction in TDP? And i assume b/c of this lower then expected TDP, Intel is going to release faster then expected CPUs?
 
Duke3d87 said:
Are the numbers the exact numbers? And is this a cut across the board? I mean will Conroe also get a 23% reduction in TDP? And i assume b/c of this lower then expected TDP, Intel is going to release faster then expected CPUs?

LOL... you always get a bit ahead of yourself man. :D All I can say is don't make assumptions. :D
 
Poncho said:
LOL... you always get a bit ahead of yourself man. :D All I can say is don't make assumptions. :D
yeah, I know. I'm majoring in CE so this stuff is like a strange passion. It's nice to know that Intel is coming out with an excellent product even though it will never enjoy the lead that it did in the early 2000s. I wonder if the release of such a powerful product will help Intel's third and fourth quater and help them continue their double digit growth.
 
Duke3d87 said:
Are the numbers the exact numbers? And is this a cut across the board? I mean will Conroe also get a 23% reduction in TDP? And i assume b/c of this lower then expected TDP, Intel is going to release faster then expected CPUs?
This would be way cool if Woodcrest came in at a 65W TDP, much better then expect, it may spell good things for Conroe as well, I don't expect a drop in TDP there maybe use it to create high binned chips. :) At least that is what I am hoping for.

To me Intel has used the better stepping to decrease power consumption rather then go for performance, if these rumors pan out.
 
Its great that woodcrest will consume only 65w. For a 3ghz cpu with 1.33ghz bus its fantastic.

Is there a reason why intel is not producing MP version of woodcrest. I hear stories that Tigerton will not release till 2008 (maybe wishful thinking of AMD !!!!!!s). But whatever the reason Dell has gone with opteron for 4way and greater servers.
 
shawmanus said:
Its great that woodcrest will consume only 65w. For a 3ghz cpu with 1.33ghz bus its fantastic.

Is there a reason why intel is not producing MP version of woodcrest. I hear stories that Tigerton will not release till 2008 (maybe wishful thinking of AMD !!!!!!s). But whatever the reason Dell has gone with opteron for 4way and greater servers.

Tigerton I thought was the desktop quad-core, with Clovertown being the server version (someone correct me if I am wrong). Either way, Q1 or Q2 '07 is what I had heard.
 
NulloModo said:
Tigerton I thought was the desktop quad-core, with Clovertown being the server version (someone correct me if I am wrong). Either way, Q1 or Q2 '07 is what I had heard.
Yeah, that isn't correct, Tigerton/Clovertown MP, is the MP server equivalent of Clovertown which is basically 2 Woodcrests in 1 package, with Kentsfield being 2 Conroe on 1 Package.

H1 2007 is the timeframe I heard as well for Intel's Quad Cores.
 
just a quick question: what is the difference between cloverton and kentsfield?
 
wizzackr said:
just a quick question: what is the difference between cloverton and kentsfield?

Cloverton is for XeonDP and Kentfield is for desktop. And Tigerton is for XeonMP
 
wizzackr said:
just a quick question: what is the difference between cloverton and kentsfield?

I think Clovertown is a glue job of two Conroes while Kentfield is native quad core.
 
perplex said:
I think Clovertown is a glue job of two Conroes while Kentfield is native quad core.
Kentsfield is still dual die. Logically, if Intel could make a single die quad core and dual die quad core they would not give the desktop one the single die quad core being that the server market is more lucrative.
 
Thanks a lot guys! So if cloverton gives you 8 cores on one dual socket mainboard, I wonder how well that dual cloverton configuration will scale in comparison to two DP woddcrest systems - isn't the FSB going to bottleneck such a system?

...Gotta love how quickly technology evolves - if I look at my 'high-end' system from two years ago... shhh :D
 
perplex said:
I think Clovertown is a glue job of two Conroes while Kentfield is native quad core.

Yup and that's why the motherboards for these has TWO NORTHBridges or One WiderNorthbridge and two speparate indedpendant but linked FSBs. At least that's the way I read it. SO it will be more like running a close-nit Cluster. Effective FSB/s is 2133MHz for one model and 2666 for the other. and about 16.8 and 21GB of Bandwidth repectively. Not the Opteron's 26GB but clearly nothing to turn your nose up at. So that's old news.

http://www.serverpipeline.com/showArticle.jhtml?articleID=173403192

It's after two-way that will cause problems until Processor links are improve. Intel should have done like AMD and copied Alpha's model.
 
Donnie27 said:
Yup and that's why the motherboards for these has TWO NORTHBridges


There is only one northbridge on Bensley. Keep an eye out for the product briefs on Intel.com. They should be out soon.
 
wizzackr said:
...Gotta love how quickly technology evolves - if I look at my 'high-end' system from two years ago... shhh :D

Or mine from 4 years ago (2.4 P4 Northwood, Radeon 9500pro, Still struggling with only 512MB of RAM), things change fast, but Conroe or Woodcrest look like hard to resist upgrade options.

I have to say though, it has taken me a while to get over the Mhz myth (or is it mith? I tend to forget which one was the game and which one is the spelling). I went from a pentium 100Mhz to a Pentium 133 to a K6-2 450 to a P4 2.4 Ghz, running a 2.4 now upgrading to a Conroe 2.66 has to be the smallest percentage upgrade I have ever done.
 
Poncho, why is Intel so reluctant to have a dedicated northbridge/CPU or dual CPU? I mean the motherboard costs enough for the MP line, the performance justifies the added cost (i mean why didn't they do it earlier) and what are the potential problems of doing that?
 
Duke3d87 said:
Poncho, why is Intel so reluctant to have a dedicated northbridge/CPU or dual CPU? I mean the motherboard costs enough for the MP line, the performance justifies the added cost (i mean why didn't they do it earlier) and what are the potential problems of doing that?

Complexity , complexity and cost for little perfomance increase if any.
 
Back
Top