Six-Core Server Processor !!

[Deleted member 3]

and yes dan i know that makes no difference whatsoever

It's not a matter of making a difference, is a matter of it being completely *insert animal* poo. There is no such thing as a "true" 6 core and whatever the opposite is. If there are 6 cores it's a 6 core CPU, period. The way they talk to each other really isn't relevant, even if they do it via two cans with a piece of string in between.

 

[MilkyWay]

having a true core design should make it more efficient but thats not actually happening, if intel is faster it will still be faster

 

[Sihastru]

look at the pic you posted it has 3 banks of 45nm cores now go look at the AMD design

www.nordichardware.com/news,7725.html

istanbul is a true 6 core design while the intel version is a bunch of dual cores grouped together that talk accross the FSB. the size and number of cores possible in the system dont really matter to what i said so i do not know why you brought it up... and the 8 series istanbul chips will allow the same thing from AMD.

Well, if you look more closely you'll see the 12MB (on some models 16MB) L3 shared cache. Because you can clearely see the three dual core formations, you think the L3 cache is positioned after the system bus, but that's not true. The scheme is cores(w/L1) -> 3MB L2 cache (per dual core formation) -> 12/16 MB L3 cache -> system bus. Just because it is based on aging FSB architecture doesn't mean the design is not a true six-core.

Don't be fooled into thinking that AMD has a magical solution to avoid data transfers on the HT bus. There are times where this helps a lot, but most of the times, it doesn't help one bit. Not everything will fit into the cache. Actually, most of the time, data will not be "shared" internally by the cores and will actually be sent via the system bus (FSB, QPI or HT, whatever that is). I brought it up because in multi-processor servers the "true n-cores design" is very small because data will more times then you'd imagine be in transit on the system bus.

Dunnington may not be a very elegant design. But in commercial workload benchmarks will destroy AMD chips. Even if it's on FSB. AMD may have "design elegance" on their side, but they have a really hard time exploiting that elegance. In the server world, raw numbers are what matters, elegance is for artists.

 

[cdawall]

It's not a matter of making a difference, is a matter of it being completely *insert animal* poo. There is no such thing as a "true" 6 core and whatever the opposite is. If there are 6 cores it's a 6 core CPU, period. The way they talk to each other really isn't relevant, even if they do it via two cans with a piece of string in between.

is the 1066FSB going to hamper these though?

having a true core design should make it more efficient but thats not actually happening, if intel is faster it will still be faster

the AMD chip is more effecient all the way around its lower TDP higher clocked and unles i'm mistaken its higher performing clock for clock to with the new chips vs the old wolfendale based chips

Well, if you look more closely you'll see the 12MB (on some models 16MB) L3 shared cache. Because you can clearely see the three dual core formations, you think the L3 cache is positioned after the system bus, but that's not true. The scheme is cores(w/L1) -> 3MB L2 cache (per dual core formation) -> 12/16 MB L3 cache -> system bus. Just because it is based on aging FSB architecture doesn't mean the design is not a true six-core.

Don't be fooled into thinking that AMD has a magical solution to avoid data transfers on the HT bus. There are times where this helps a lot, but most of the times, it doesn't help one bit. Not everything will fit into the cache. Actually, most of the time, data will not be "shared" internally by the cores and will actually be sent via the system bus (FSB, QPI or HT, whatever that is). I brought it up because in multi-processor servers the "true n-cores design" is very small because data will more times then you'd imagine be in transit on the system bus.

Dunnington may not be a very elegant design. But in commercial workload benchmarks will destroy AMD chips. Even if it's on FSB. AMD may have "design elegance" on their side, but they have a really hard time exploiting that elegance. In the server world, raw numbers are what matters, elegance is for artists.

what benchmarks show that the intel chip destoys the AMD chip they should be very close in performance if not dead even just like intel quad vs new 45nm amd quads

the only benchmarks i can find in a server enviroment show the older 65nm quad core barcalona based AMD quads against the newest intel 6 core chips

 

[btarunr]

look at the pic you posted it has 3 banks of 45nm cores now go look at the AMD design

www.nordichardware.com/news,7725.html

istanbul is a true 6 core design while the intel version is a bunch of dual cores grouped together that talk accross the FSB. the size and number of cores possible in the system dont really matter to what i said so i do not know why you brought it up... and the 8 series istanbul chips will allow the same thing from AMD.

As long as it's on the same die, it is "true". Dunnington is. Besides, true or not, the design approach doesn't make any difference, as proven. Intel Core 2 Quad smoked Phenom at SANDRA inter-core bandwidth tests.

Also, Nordic's picture is an artist's impression.

 

[FordGT90Concept]

the AMD chip is more effecient all the way around its lower TDP higher clocked and unles i'm mistaken its higher performing clock for clock to with the new chips vs the old wolfendale based chips

Intel uses inclusive caching and that usually means more cache space to do the same job. Intel, therefore, has a larger cache and large cache means bigger die and bigger die means more power.

 

[Disparia]

What a touchy subject. Some people are just curious about CPU architecture

As for that 1066Mhz bus, there's a ton of L3 on the chip as well as a cache on the north bridge (64MB). This is what allowed a 4P/16C setup to scale as well as it did to 24C, I assume. Considering the mem/nb/cpu setup.

4P Dunnington vs 4P Opteron? - Yes, the difference is pretty vast, but is what I expect from a 24C vs 16C match-up, unless someone can site a differing review?

 

[cdawall]

What a touchy subject. Some people are just curious about CPU architecture

As for that 1066Mhz bus, there's a ton of L3 on the chip as well as a cache on the north bridge (64MB). This is what allowed a 4P/16C setup to scale as well as it did to 24C, I assume. Considering the mem/nb/cpu setup.

4P Dunnington vs 4P Opteron? - Yes, the difference is pretty vast, but is what I expect from a 24C vs 16C match-up, unless someone can site a differing review?

thats the only review i found and not only was it 24C vs 16C the 24C were 45nm and based off of wolfendale while the 16C were the crappy AMD barcelona cores that i have sitting in my server

 

[Darren]

so now when is AMD going to respond to this and look at it is is a 45NM CPU as well . Seems to me like the core wars are on and Intel is taking a huge leap forward ! I never though it would cost so much though think they will ever bring that price down ? $2,399.00 !! That is sure a ton of cash !

Flame bait?

I bet by page 4 everyone is arguing, trickson and/or R9 will be at the center of the conflict justifying spending "$2,399.00" on Intel's processor opposed to AMD's alternative.


Edit:

you could have read thru the whole thread before posting that



you did fine nothing you did deserved that one....

Indeed I did read the entire thread.


I'll wait till page 4 or 5 before taking back my statements. These types of threads always end in one way.


Edit 2:

it hasn't so far there is a person trying to push them in that direction but i dont think he can find a single benchmark to back up his point

Indeed, there is no flaming going on, but I thought I'd give you guys the heads up before it starts

Edit:

I actually saw that 6 core Intel CPU on newegg a few months ago.

 

[trickson]

Flame bait?

I bet by page 4 everyone is arguing, trickson and/or R9 will be at the center of the conflict justifying spending "$2,399.00" on Intel's processor opposed to AMD's alternative.

WOW I think the discussion is going great until you come on and post up this . Flamebait ? NO I am asking a valid question . I got my answer I did not even KNOW that AMD had the first 6 cored CPU so you are the one trying to flaimbait me . I have not tried to justify the price at all to any one . so just what is this post of yours intended to do ?

 

[cdawall]

Flame bait?

I bet by page 4 everyone is arguing, trickson and/or R9 will be at the center of the conflict justifying spending "$2,399.00" on Intel's processor opposed to AMD's alternative.

you could have read thru the whole thread before posting that

WOW I think the discussion is going great until you come on and post up this . Flamebait ? NO I am asking a valid question . I got my answer I did not even KNOW that AMD had the first 6 cored CPU so you are the one trying to flaimbait me . I have not tried to justify the price at all to any one . so just what is this post of yours intended to do ?

you did fine nothing you did deserved that one....

 

[n-ster]

I was also surprised when you brought the flame bait up... cuz I see none and you can clearly see that there is nothing even approaching a flame war here lol...

Now I am kinda a noob in CPU architecture etc... so this means that a (not dual cores sticked together) six-core has barely any benefit over the (sticked dual cores together) six-cores? I thought the Qxxxx were better than Phenom etc, because they just had a much superior architecture except for the part that it is 2 dual cores...

 

[cdawall]

Indeed I did read the entire thread.


I'll wait till page 4 or 5 before taking back my statements. These types of threads always end in one way.

it hasn't so far there is a person trying to push them in that direction but i dont think he can find a single benchmark to back up his point

 

[zAAm]

This is pretty cool! Can't wait for Intel and AMD to begin waging their 6-core DESKTOP wars! Don't care much for the server processors. Competition is always good because it encourages growth and innovation so I'm all for AMD vs Intel I think some people take it a teeny tiny bit too far though... It's almost turning into a religion for some people. They stand by their argument and nothing is going to convince them otherwise

Mainly, I just want to see some desktop CPU's with an increasing amount of cores soon

 

[n-ster]

I'm just waiting for games to take advantage of all these cores lol Seriously, hardware is going past software core wise etc, but games need more pure MHZ and bigger GPUs (Crysis )?

no flame war is going to start since we have no solid info (performance wise etc) about the six-core CPUs... and we are a civilized bunch it TPU most of the time

 

[trickson]

One thing is for sure the core wars have started at least for the server market . This is exciting as we will ( I Hope ) see them starting to go to the desktop soon . how fun is that going to be ?

 

[Deleted member 3]

One thing is for sure the core wars have started at least for the server market . This is exciting as we will ( I Hope ) see them starting to go to the desktop soon . how fun is that going to be ?

Until more programs utilize that many cores it doesn't matter much.

 

[trickson]

Until more programs utilize that many cores it doesn't matter much.

I agree . But with the advent of this new tech this should happen fairly quickly . More games are coming out with multy core support some are even taken the Quads for a ride .

 

[DrPepper]

Ah dunnington. I remember reading about that on anandtech when it came out. Quite a monster, A bit faster than the AMD equivelant if I remember correctly yet the opteron wasn't designed to beat it performance wise, more as a faster replacement for older opteron's without having to change the entire server all thats needed is update bios and drop the opty in.

 

[erocker]

I agree . But with the advent of this new tech this should happen fairly quickly . More games are coming out with multy core support some are even taken the Quads for a ride .

Lol, name some games that will utilize six cores! This processor has nothing to do with "gaming". I wish more games could utilize my two and three core processors!

 

[Disparia]

Seems relevant to our interests:

Can see old vs new Xeon cores, vs Opteron (8356), and the scaling before, at, and beyond core count. They're all 16C except for the 24C E7460 setup.

 

[Disparia]

Lol, name some games that will utilize six cores! This processor has nothing to do with "gaming". I wish more games could utilize my two and three core processors!

Hellgate London, was even demostrated on a SkullTrail box. Unfortunately for all the time they spent designing a multi-threaded engine, they forget that even perfect design isn't worth much if you lack engaging game play.

 

[cdawall]

Seems relevant to our interests:

http://www.theburnerishot.com/photo/totalthrput.jpg

Can see old vs new Xeon cores, vs Opteron (8356), and the scaling before, at, and beyond core count. They're all 16C except for the 24C E7460 setup.

those opterons scale damn good

 

[trickson]

Lol, name some games that will utilize six cores! This processor has nothing to do with "gaming". I wish more games could utilize my two and three core processors!

I mean that in the future they will . Some day soon I hope we will see games taking a huge turn for the good . Putting that 6 cores to use might mean faster better games and graphics .

 

[t77snapshot]

Seems relevant to our interests:

http://www.theburnerishot.com/photo/totalthrput.jpg

Can see old vs new Xeon cores, vs Opteron (8356), and the scaling before, at, and beyond core count. They're all 16C except for the 24C E7460 setup.

What about Opteron's 8435 2.6GHz?