AMD Demonstrates World's First Native Quad-Core X86 Server

[LordPope]

http://www.amd.com/us-en/Corporate/VirtualPressRoom/0,,51_104_543~114496,00.html

 

[evilr00t]

9-inch!!

Oh, wait...

 

[ElMoIsEviL]

http://www.amd.com/us-en/Corporate/VirtualPressRoom/0,,51_104_543~114496,00.html

Wow... Intel has been sampling native quad core 45nm processors for a while now (like a month). So this is hardly the world's first native Quad Core CPU. It's the second.

 

[evilr00t]

Really? I thought it was in the last two weeks or so. Are there ES's floating around yet?

 

[LordPope]

a first sample is a long way from being a ES

ES means its working some what,,,,,

 

[IamXero]

a first sample is a long way from being a ES

ES means its working some what,,,,,

More so, Intel doesn't typically give out Engineering Samples 6 months before a product launch..

Don't quote me on the 6 months. It could be tomorrow for all I know.

 

[evilr00t]

There were a whole lot of Conroes floating around months before Core 2's launch. But that was Intel pulling a guerrilla marketing stunt.

 

[IamXero]

There were a whole lot of Conroes floating around months before Core 2's launch. But that was Intel pulling a guerrilla marketing stunt.

Yes however not usually right after they announce the 45nm chips :?

 

[RandMcnally]

http://www.amd.com/us-en/Corporate/VirtualPressRoom/0,,51_104_543~114496,00.html

Wow... Intel has been sampling native quad core 45nm processors for a while now (like a month). So this is hardly the world's first native Quad Core CPU. It's the second.
Nehalem? I've not seen a single thing about it. That's supposed to be the first Intel native quad as of late.

 

[NMDante]

Where was the demonstration? Or did I miss it somewhere?

I checked out the Virtual IT site, and it was a bunch of slides and animations, but it did show a task manager with 16 CPU processing. That was the demo?

With something this big, I was looking for an actual program or four running, with the task manager showing the loads on each processor, or something. Not a bunch of slides and animations of how things work.

This wasn't a demonstration, more than it was a presentation.

 

[RandMcnally]

Well they didn't run any programs, just good ol' task manager.

 

[NMDante]

Well they didn't run any programs, just good ol' task manager.

That's what I mean. That's not a demo, that's just task manager. Whoo-hoo. If they meant to show a 4 socket platform with a quad-core CPU, they succeed. But that's no demo of what it can actually do.

The slides they showed, had it 40% and 70% faster at tasks. Show those tasks running. I mean, they have the data to claim it's that much faster, so why not show the system running the programs mentioned?

I know both Intel and AMD do these goofy "demonstrations" and they end up more like presentations, and they just annoy the hell outta me. When I see "DEMONSTRATES" or "DEMONSTRATIONS", I expect a system running programs. Hey, it's like when you download a demo of a game, you expect to play a level or two, not watch a movie about a level.

 

[LordPope]

well one thing thie DEMO shows is that DAMMIT is close to showing some scores

 

[shinigamiX]

Actually, it shows next to nothing.

 

[tchiwam]

http://www.tgdaily.com/2006/11/30/amd_native_quad_core_barcelona/

The interesting benefit I would think here is the number of pad on the die, they are quite big if you compare the chip estate versus transistors. Look at the picture the Intel dies are about the same total space than the AMD single die. Even if in theory AMD can use fewer connections because of the crossbar doing most of the job.

I can't wait to get the real stats and see why AMD is having such a big die. I know HT and memory controller needs more pins than a simpler FSB implementation.

I did mention before that putting the memory on a HT channel might be beneficial at the end of the day. This bring the extra benefit to shrink the die by taking away the controller and remove a lot of connections, keep the socket to a sane amount of pins. Intel proved that it is possible to have good performances with an external memory controller.

 

[weskurtz81]

well, you could try not be be the BM or 9 inch opposite. So, this guy makes a post, and you attack him. He didn't lie, he just made a post about something that is true. I don't think he personally made a claim about it being better, he just supplied a link. So chill out, stop being a fanboy, those comments make you just about as bad as BM and 9 inch.

wes

PS- what did you think about all those people calling Conroe crap before it was released? All I am saying, don't be them.

 

[gOJDO]

http://www.amd.com/us-en/Corporate/VirtualPressRoom/0,,51_104_543~114496,00.html

Noob, where do you live?
There are a lot of more than quadcore server CPUs demonstarated allready.
For example, native 9 core Cell processor!

 

[qcmadness]

http://www.amd.com/us-en/Corporate/VirtualPressRoom/0,,51_104_543~114496,00.html

Noob, where do you live?
There are a lot of more than quadcore server CPUs demonstarated allready.
For example, native 9 core Cell processor!

You are just behaving as a fanboy as him.

You may argue with the definition of "native quad-core".

But Cell processor surely is not a x86 processor.

 

[weskurtz81]

gojdo, come on man, Cell, is not 4 symetric cores, Cell is a different animal. We all know there are more powerful CPU's out there for certain things, that have more "cores" albeit not symetrical, but why flame this guy for simply starting this thread.... does he have a history of being a 9 inch or BM? Did he say something that wasn't true? It doesn't seem so to me.

wes

 

[gOJDO]

both, weskurtz81 and qcmadness are right. I missed the "x86" in the title.
Anyway, I'll skipp the nativ vs non-nativ debate, ebcause it is pointless.
What matters is performance scallability and Kentsfield/Clvoertown showed that "non-native quadcore" CPUs with the "slow FSB" can scale amazingly linear in perfromance.

 

[qcmadness]

both, weskurtz81 and qcmadness are right. I missed the "x86" in the title.
Anyway, I'll skipp the nativ vs non-nativ debate, ebcause it is pointless.
What matters is performance scallability and Kentsfield/Clvoertown showed that "non-native quadcore" CPUs with the "slow FSB" can scale amazingly linear in perfromance.

Thanks to the excellent pre-fetching alogarithm from the Prescott experience, a quad-core will not saturate the 8.5GB/s FSB by Tom's Hardware's test.

Intel is pushing multi-FSB northbridges, so bandwidth would not be a problem for them. As the design cost and the stability will be a concern from the builders, Intel will debut CSI in 2008.

 

[Ycon]

Who the heck cares?

 

[weskurtz81]

It's cool man, I am not debating anything with the performance between AMD and Intel quads, in all honesty(4x4 aside), we will not know the performance between the two until we have something to put ours hands on. The one thing I CAN say, if the wattage numbers are accurate, there is something to be said about AMD keeping the power envelope where it is. That in my opinion is impressive. But, if it gets it's but whooped by the Intel quads, I guess it shouldn't use as much power ..... who knows.... and native or non native, who cares, it is all about performance and power. The OP didn't make any false statements.... my only point.

wes

Edit: had to patch up my bad writing/typing form the best I could.... run ons.
Edit#2:Ycon, I could say the same about your posts.... however, you haven't been as bad as some of the "others" recently. It used to seem like your account was a finger puppet to give BM 9 inch and the others someone that would say things to make them seem intellegent. The world may never know.

 

[Threshold]

9-inch!!

Oh, wait...

Y'see, there's the problem right there. And I'd like to thank you guys who yelled at him for it, but I'm going to anyway because I have a bit of a general statement.

I'm about sick and tired of how any time someone posts a pro-AMD link...or even a link that isn't, GOD FORBID, pro-Intel...or even a link that isn't necessarily either, just informational, albeit not much, like this one...someone has to post before they think, and accuse the OP of being 9-inch. Now, I wasn't here when 9-inch was (and I thank God for that), but I'm going to guess he'd probably include some BS and/or FUD with the link, which wasn't the case here, just a simple link.

Then there's the fact that the guy has 182 posts, at this point in time. Jesus Christ man, get a life. I'm positive if it looks like 9-inch everyone will notice regardless of whether or not you make a single-word (or, in this case, three-word) post. Furthermore, I'm fairly positive 9-inch won't be coming back, although we do still have to deal with Baron.

A year ago, most of you would have had exactly the same attitude towards anyone who posted anything pro-Intel, non-pro-AMD, or anything Intel related that didn't even include AMD. And yes, they both do shitty marketing, for the most part. Shitty as in not 100% truthful...probably not even 70% truthful, I mean I know it's advertising, but COME ON.

I don't care if you flip-flop. I don't care if you "don't care" who makes the chip so long as it's fast. But for God's sake quit accusing people with no evidence whatsoever, and take a chill pill.

Now then. To get back on topic. Interesting, and was on TGDaily as well. http://www.tgdaily.com/2006/11/30/amd_native_quad_core_barcelona/ Slightly more info, IIRC. Looks like they have an actual chip in hand, although obviously it's not ready to sell yet. Topic might be misnamed, but it's close enough, and it's the same as TGDaily's title, except for the x86, so I'm not going to complain.

I wouldn't touch native or non-native with a 30 foot long DIMM. (I wouldn't touch 4x4, either.)

My apologies for the offensive tone, but I find it rather annoying.

 

[weskurtz81]

I obviously agree with you threshold, gODJO is normally fairly well informed.... then there are those that just start with the attacks.... on both sides. It sad that I even said that, "on both sides".... like this is some kind of religion. Anyway, I agree with you. Like I already said, I am extremely interested with the power numbers, two more cores, while touting a lower TDP. We will see if those claims hold water though.

wes