amd quadfather 4X4 VS intel Kentsfield

[digitaldreamzz]

i aint going to buy either of them as at my place they will cost a fortune...but im very much interested to hear experts say their opinions about them...i have not seen any previews of the 4X4...so here are my qns plzz enlighten me on them...

1>How practial will a 4X4 be?I mean will it be directly competing with the Kentsifeld alone or is amd going to let any 2 athlon 64x2 am2 processors work as 4X4 there by competing with the core 2 duo's in terms of total cost

2>-will the FSB bottle neck the kentsfield enuf to let the 4X4 take the performance crown at the high end..is it even true that FSB will be a bottle neck to the kentsfiled

3>-can the 4X4 work as good as theoretic quad core in the sense it has total 4 cores...i.e. aint it similar to how intel has made their current first gen quadcores...

4>-will the 4X4 give the same performance of an imaginary amd quad core (supposedly the deerhound) with 2 dual cores glued together in same packet jus like intel quad cores

plz let me know ur views...thnx...looking fwd to a good discussion

 

[gOJDO]

1. 4x4 will be almost practicle as s940 Opteron 2xx. The 4x4 is capable to work only with FX series and they are expencive. The mainboard for the 4x4 will be much more expencive and at least 4 modules of RAM are required. 4x4 will perform competetive to Kentsfield system, but will not beat it and will cost more.

2. The FSB will bottleneck Kentsfield, but not that bad. Even with the bottlenecked FSB, it still outperforms same clocked quadcore AMD K8 system.

3. 4x4 will scale more linear in performance than Kentsfield for multithreaded apps with the NUMA model. But it is not quite good as native quadcore where all cores are connected to the same crossbar. The adventage of two dualcore K8 CPUs are the two ODMCs, improving the total RAM bandwidth.
Kentsfield is glued using "the old" Pentium-D technique of glueing singlecores.

4. No, Deerhound will perform better.

 

[mesarectifier]

There are no solid answers to any of your questions, only speculative.

 

[gOJDO]

There are no solid answers to any of your questions, only speculative.

Not quite. We've allready seen what he was wondering about.

 

[gOJDO]

There are no solid answers to any of your questions, only speculative.

Not quite. We've allready seen what he was wondering about.

What do you mean? Is there full benches for anything he was asking about? Kind of off...
yeap. there are benchmakrs on unreleased CPUs. We've allready seen how K8 scales on 2P(4P & 8P) and how 2P Core 2.

 

[digitaldreamzz]

well i jus wanted to know what to expect....unbiased speculations will do fine

 

[m0rk]

There are no solid answers to any of your questions, only speculative.

Not quite. We've allready seen what he was wondering about.

What do you mean? Is there full benches for anything he was asking about? Kind of off...
yeap. there are benchmakrs on unreleased CPUs. We've allready seen how K8 scales on 2P(4P & 8P) and how 2P Core 2.

Wow, can you post a link to the full benchmarks of the amd 4x4 and kentsfield? I would like to see it and if you wouldn't mind, clear up the confusion of everybody else that doesn't seem to know what's going on either, including myself. And isn't woodcrest a lot different from kentsfield? Your comparing dual duo cores to a quad core. Your making a lot of assumptions here with no solid information. Even if it is similar... We have yet to see how 4x4 will actually work also.

http://www.xtremesystems.org/forums/showthread.php?p=1527913

perhaps not the best, but im to lazy to search on sundays

 

[chunkymonster]

Kentsfield is a quad core processor. While there have been ES releases of Kentsfield and preliminary benchmarks of its potential, there is nothing definitive.

4x4 is not a processor, it is a high end enthusiasts platform to showcase AMD's scalability, overclockingability, and system tweakability...read this news release for more info. Contrary to what others are posting there have been no releases of 4x4 and it is impossible to get a direct comparison of a Kentsfield processor to the 4x4 platform.

Any answers at this time are pure speculation.

With that said...here's my two cents...
1) 4x4 will not be practical for the average everyday gamer or web-surfer, just as SLI is not practical for he average everyday gamer/web surfer...Exactly how 4x4 will work and with what procs, aside from FX models as stated, has yet to be determined.
2) The Kentsfield FSB will not be a cause for any bottle necks
3) Can't answer that for sure as we haven't seen a 4x4 system, but 4 cores is 4 cores...I equate this to the old arguement of dual processor or dual core, both perform about the same
4) Again, can't answer that for sure as we haven't seen a 4x4 system, but 4 cores is 4 cores...I equate this to the old arguement of dual processor or dual core, both perform about the same

 

[BaronMatrix]

1. 4x4 will be almost practicle as s940 Opteron 2xx. The 4x4 is capable to work only with FX series and they are expencive. The mainboard for the 4x4 will be much more expencive and at least 4 modules of RAM are required. 4x4 will perform competetive to Kentsfield system, but will not beat it and will cost more.

2. The FSB will bottleneck Kentsfield, but not that bad. Even with the bottlenecked FSB, it still outperforms same clocked quadcore AMD K8 system.

3. 4x4 will scale more linear in performance than Kentsfield for multithreaded apps with the NUMA model. But it is not quite good as native quadcore where all cores are connected to the same crossbar. The adventage of two dualcore K8 CPUs are the two ODMCs, improving the total RAM bandwidth.
Kentsfield is glued using "the old" Pentium-D technique of glueing singlecores.

4. No, Deerhound will perform better.

4x4 is not restricted to FX. It is not possible to restrict it. FX has one HT link, X2 has one HT link. For any chip they just need to turn on an extra link.

There have been closed system demos where it was said to be 80% faster than one dual core.

Kentsfield will be fast but it is said it will use 125W+ so 4x4 can compete power wise. K8L will kill them both though, BUT FX is supposed to get the K8L core at the end of next year, maybe sooner.

 

[akscottbot]

Baron, whats the deal dude? What happened to you to be so staunchy pro-AMD? There a good company but, jeez do you owe em money? Did Intel run over your dog or something? core 2>k8>netburst thats just way it is right now. When K8L comes out itll be great im sure. Why hype it so much? Arent a64's good enough all ready? every time i read your posts its 'K8L STOMPS EVERYHTING!! ITS TEH ROXXORS IT PWNS YOUR MAMA AND ITS OUT NEXT WEEK!!! YAY!!!!'

Everytime i read one of your posts, its blatent lies. I dont think you believe the BS you spout on here. What gives man?

 

[evilr00t]

It's 178W, not 158W.

 

[exit2dos]

I think that the 4x4 is like SLI. If the mobo supporting it costs about the same as one that doesn't, and the CPUs with 2 HT links are also priced about the same as those with one - then the 4x4 platform would make a good choice if you're building an AMD system.

As long as you can run it with just one CPU, then I would base a AMD build on this platform since you could just pop in another CPU later to upgrade performance.

Again, this is supposing that the components are about the same price as the non-4x4 ones, and that you are going to build a AMD system anyway.

 

[wickedmonster]

Here's a 2P woodcrest vs 2P Opteron
http://www.anandtech.com/showdoc.aspx?i=2772&p=1

Not exactly kentsfield vs 4x4 but close. Results show woodcrest wins by a mile.

Here's a comparison of 4 core Woodcrest vs 2 core woodcrest on a Mac Pro

http://www.anandtech.com/showdoc.aspx?i=2816&p=4

4-core processors are about as useful as a Ferrari on a residential street. Software today barely take advantage of dual core much less quad core. It will be at least 2-3 years before we see any quad core advantage over dual core.

 

[tamalsmith]

are there any benchmark of revG quadcore AMD K8 ?

 

[turpit]

Jack,

I appreciate what you do, but some causes are lost even before a person begins the fight. You know even if Hector were to come out himself and admit Intel currently beat AMD in every category, The Horde would simply jump in with some new BS competitive/comparative" category.........


"................ In fair and unbiased comparison testing performed by Charlie Demerjian in cooperation with AMDZone, AMD retail box packaging has been shown to be %78 more visually attractive than Intel retail box packaging. This will create astronomical demand for AMD products driving the evil empire of Intel out of business"


Peace

 

[digitaldreamzz]

1. 4x4 will be almost practicle as s940 Opteron 2xx. The 4x4 is capable to work only with FX series and they are expencive. The mainboard for the 4x4 will be much more expencive and at least 4 modules of RAM are required. 4x4 will perform competetive to Kentsfield system, but will not beat it and will cost more.

2. The FSB will bottleneck Kentsfield, but not that bad. Even with the bottlenecked FSB, it still outperforms same clocked quadcore AMD K8 system.

3. 4x4 will scale more linear in performance than Kentsfield for multithreaded apps with the NUMA model. But it is not quite good as native quadcore where all cores are connected to the same crossbar. The adventage of two dualcore K8 CPUs are the two ODMCs, improving the total RAM bandwidth.
Kentsfield is glued using "the old" Pentium-D technique of glueing singlecores.

4. No, Deerhound will perform better.

4x4 is not restricted to FX. It is not possible to restrict it. FX has one HT link, X2 has one HT link. For any chip they just need to turn on an extra link.

There have been closed system demos where it was said to be 80% faster than one dual core.

Kentsfield will be fast but it is said it will use 125W+ so 4x4 can compete power wise. K8L will kill them both though, BUT FX is supposed to get the K8L core at the end of next year, maybe sooner.

well i remember reading that 4X4 demo...but that system used registered dimms...so it was obviously opty's....and not the actual 4X4 they were testing....and abt that 80% was only while what amd calls as "MEGATASKING".....

i have also read that each processor will have its own dimm's...so that means if its total 4DIMMS each of them gets only 2? is it true....is it gonna affect is performance eh?

 

[exit2dos]

Honestly, I completely expect K8L to rock the world, it just how Baron puts it that crawls under my skin --- and he eats up the Inq 'reported 80% better' bullcrap like it's a Wolfgang Puck gormet steak. It's fun poking fun at him as it is sooooo easy.

To be fair to the Baron, it is AMD reporting the 80% increase.

http://www.amdzone.com/modules.php?op=modload&name=News&file=article&sid=5937

 

[thematrixhazuneo]

This was on toms page, so maybe quad core wont consume that much?

"While Intel's Kentsfield and Clovertown processors are expected to bring a substantial increase in power consumption to a thermal designer power rating of at least 110 watts, AMD claims that its quad-cores will not consume more than today's processors."

http://www.tgdaily.com/2006/09/01/amd_quad_core_update/

 

[exit2dos]

That's referring to the upcoming "K8L" CPU, not the 4x4 - which is going to use 2 seperate dualcore CPUs.

 

[gOJDO]

Wow, can you post a link to the full benchmarks of the amd 4x4 and kentsfield? I would like to see it and if you wouldn't mind, clear up the confusion of everybody else that doesn't seem to know what's going on either, including myself. And isn't woodcrest a lot different from kentsfield? Your comparing dual duo cores to a quad core. Your making a lot of assumptions here with no solid information. Even if it is similar... We have yet to see how 4x4 will actually work also.

I can post links, but I will not. Others allready done that.
I think that 4x4 will perform like s940 2xx becouse their architectures are same, they have same L2, same HTT links at same freqfency, both NUMA. There is only one difference in the external architecture of the CPUs, their ODMCs and system RAM type. There is no real advantage of DDR2 on the K8. K8 is much more dependend of the RAM realtime latency(freqfency / latency) and less dependend on the RAM bandwidth. My expectations about the 4x4 is to perform just like 2P s940 2xx and in some rare cases with 0-5% advantage.
Benchmarks on Kentsfield 2.66GHz FSB1066 ES showed that there is no "great" bottleneck as AMD fanboys yield, but it is getting bottlenecked with FSB1066 as the freqfency raises. At the end of this year there will be more apropriate chipsets supporting FSB1333 and there should be no discussion about quadcore Core2 3GHz 2x4MB L2 FSB1333. If you compare the known numbers today and scale them to the apropriate freqfency, the two overpriced 3GHz K8(the best available Q4 2006 and until H2 2007) will be defeated by 3GHz Kentsfield.
But not everybody needs the best and not everybody will buy the most expencive. The most important factor when buying DT for most people around the world is performance/price. Kentsfield system will wipe the floor with same priced 4x4 system.

 

[BaronMatrix]

1. 4x4 will be almost practicle as s940 Opteron 2xx. The 4x4 is capable to work only with FX series and they are expencive. The mainboard for the 4x4 will be much more expencive and at least 4 modules of RAM are required. 4x4 will perform competetive to Kentsfield system, but will not beat it and will cost more.

2. The FSB will bottleneck Kentsfield, but not that bad. Even with the bottlenecked FSB, it still outperforms same clocked quadcore AMD K8 system.

3. 4x4 will scale more linear in performance than Kentsfield for multithreaded apps with the NUMA model. But it is not quite good as native quadcore where all cores are connected to the same crossbar. The adventage of two dualcore K8 CPUs are the two ODMCs, improving the total RAM bandwidth.
Kentsfield is glued using "the old" Pentium-D technique of glueing singlecores.

4. No, Deerhound will perform better.

4x4 is not restricted to FX. It is not possible to restrict it. FX has one HT link, X2 has one HT link. For any chip they just need to turn on an extra link.

There have been closed system demos where it was said to be 80% faster than one dual core.

Kentsfield will be fast but it is said it will use 125W+ so 4x4 can compete power wise. K8L will kill them both though, BUT FX is supposed to get the K8L core at the end of next year, maybe sooner.

Ummm, nope.... you are spreading mis-information again.

Kentsfield is positioned at 110 watts, and the exact specs are not known:
http://www.tgdaily.com/2006/08/17/core_2_extreme_quad_confirmed/

4x4 will require special CPUs in special packagaging for 2 hyperlinks true, those parts have not been disclosed be it FX-62, 5000+ or what have you. As AMD is shooting for the performance crown I doubt you will see low end, but that is a wait and see. Unless AMD uses energy efficient CPUs then the 4x4 solution is a true multiply by 2 power envelope, thus the 89 watt parts will come in at 158 (correction, can't multiply by 2 it is 178) watts total, or, heaven forbid they use an FX-62 and it comes it at a whopping 250 watts total.

Even the enthusiast sites are raising the question mark:

Moving on, there were some very high powered (as in power draw) announcements. First off, AMD is pushing a new high end enthusiast platform consisting of dual socket motherboards for dual core processors combined with quad GPU solutions. In an incredibly unoriginal moment of indiscretion, this platform has been dubbed 4x4. Uninspired, yet very appropriate: the platform will very likely be large, loud, and so power hungry we will need a gas powered generator to run it. That doesn't mean we wouldn't want to own a system. We just aren't sure we'd want to pay for it.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2768&p=2

Also, please provide the link to the benchmark data that shows K8L will kill them both.... by my recollection it has just taped out, even if that happens to even be the K8L and not just a run of the mill quad core Rev G K8 core -- which is more likely.

AMD is very ambiguous on the K8L naming, part positioning, etc. It is not at all clear that the tapeout announcement was for K8L because AMD doesn't even call it K8L.

@ DiGiTaLDreamZZ -- watch out for BaronMatrix, he is the forum AMD fanatic second only to 9-inch, and often attempts to prop up AMD with conclusionary statements that cannot be backed up by fact. He does not understand how computers work, how they are made, or how the market responds to yield/supply/demand issues. He is metacongatively challened to recognize his own limitations.


Jack


Because I am very busy I won't quote on the fact that it's almost electrically impossible for Kentsfield to have 20W less than two Core 2s (65W). The last I heard the 2.67 Kentsfield would come in around 125W. You are the one who attempts to refute facts with crap links that as you said CAN'T BE EXACT BECAUSE THE THING HASN'T BEEN RELEASED YET.

Also, AMD already has 2.6 Opterons running at 55W, so it's not amazing that the slowest iterations of 4x4 will run at the same or less than Kentsfield.
And because people who would buy that may also buy 2xxx Opteron, how could the power consumption be so much greater?

WHy don't you stick your head back in the chromatograph and see if you left something?

Just like Core 2s specs said how it would perform, so do K8Ls. Ars did a real in depth look at it and just that it may be better in most things that 51xx, so that means Kentsfield and 4x4 will HAVE TO TAKE A BACKSEAT.

Again, comments liek these are why I don't respond to you.

A-HOLE.

 

[clairvoyant129]

1. 4x4 will be almost practicle as s940 Opteron 2xx. The 4x4 is capable to work only with FX series and they are expencive. The mainboard for the 4x4 will be much more expencive and at least 4 modules of RAM are required. 4x4 will perform competetive to Kentsfield system, but will not beat it and will cost more.

2. The FSB will bottleneck Kentsfield, but not that bad. Even with the bottlenecked FSB, it still outperforms same clocked quadcore AMD K8 system.

3. 4x4 will scale more linear in performance than Kentsfield for multithreaded apps with the NUMA model. But it is not quite good as native quadcore where all cores are connected to the same crossbar. The adventage of two dualcore K8 CPUs are the two ODMCs, improving the total RAM bandwidth.
Kentsfield is glued using "the old" Pentium-D technique of glueing singlecores.

4. No, Deerhound will perform better.

4x4 is not restricted to FX. It is not possible to restrict it. FX has one HT link, X2 has one HT link. For any chip they just need to turn on an extra link.

There have been closed system demos where it was said to be 80% faster than one dual core.

Kentsfield will be fast but it is said it will use 125W+ so 4x4 can compete power wise. K8L will kill them both though, BUT FX is supposed to get the K8L core at the end of next year, maybe sooner.

Ummm, nope.... you are spreading mis-information again.

Kentsfield is positioned at 110 watts, and the exact specs are not known:
http://www.tgdaily.com/2006/08/17/core_2_extreme_quad_confirmed/

4x4 will require special CPUs in special packagaging for 2 hyperlinks true, those parts have not been disclosed be it FX-62, 5000+ or what have you. As AMD is shooting for the performance crown I doubt you will see low end, but that is a wait and see. Unless AMD uses energy efficient CPUs then the 4x4 solution is a true multiply by 2 power envelope, thus the 89 watt parts will come in at 158 (correction, can't multiply by 2 it is 178) watts total, or, heaven forbid they use an FX-62 and it comes it at a whopping 250 watts total.

Even the enthusiast sites are raising the question mark:

Moving on, there were some very high powered (as in power draw) announcements. First off, AMD is pushing a new high end enthusiast platform consisting of dual socket motherboards for dual core processors combined with quad GPU solutions. In an incredibly unoriginal moment of indiscretion, this platform has been dubbed 4x4. Uninspired, yet very appropriate: the platform will very likely be large, loud, and so power hungry we will need a gas powered generator to run it. That doesn't mean we wouldn't want to own a system. We just aren't sure we'd want to pay for it.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2768&p=2

Also, please provide the link to the benchmark data that shows K8L will kill them both.... by my recollection it has just taped out, even if that happens to even be the K8L and not just a run of the mill quad core Rev G K8 core -- which is more likely.

AMD is very ambiguous on the K8L naming, part positioning, etc. It is not at all clear that the tapeout announcement was for K8L because AMD doesn't even call it K8L.

@ DiGiTaLDreamZZ -- watch out for BaronMatrix, he is the forum AMD fanatic second only to 9-inch, and often attempts to prop up AMD with conclusionary statements that cannot be backed up by fact. He does not understand how computers work, how they are made, or how the market responds to yield/supply/demand issues. He is metacongatively challened to recognize his own limitations.


Jack


Because I am very busy I won't quote on the fact that it's almost electrically impossible for Kentsfield to have 20W less than two Core 2s (65W). The last I heard the 2.67 Kentsfield would come in around 125W. You are the one who attempts to refute facts with crap links that as you said CAN'T BE EXACT BECAUSE THE THING HASN'T BEEN RELEASED YET.

Also, AMD already has 2.6 Opterons running at 55W, so it's not amazing that the slowest iterations of 4x4 will run at the same or less than Kentsfield.
And because people who would buy that may also buy 2xxx Opteron, how could the power consumption be so much greater?

WHy don't you stick your head back in the chromatograph and see if you left something?

Just like Core 2s specs said how it would perform, so do K8Ls. Ars did a real in depth look at it and just that it may be better in most things that 51xx, so that means Kentsfield and 4x4 will HAVE TO TAKE A BACKSEAT.

Again, comments liek these are why I don't respond to you.

A-HOLE.


Atleast he is providing links to back up his claim, even if you think they are "crap."

We all know you're a pro-AMD fanatic, so why should anyone believe you? :lol:

 

[BaronMatrix]

1. 4x4 will be almost practicle as s940 Opteron 2xx. The 4x4 is capable to work only with FX series and they are expencive. The mainboard for the 4x4 will be much more expencive and at least 4 modules of RAM are required. 4x4 will perform competetive to Kentsfield system, but will not beat it and will cost more.

2. The FSB will bottleneck Kentsfield, but not that bad. Even with the bottlenecked FSB, it still outperforms same clocked quadcore AMD K8 system.

3. 4x4 will scale more linear in performance than Kentsfield for multithreaded apps with the NUMA model. But it is not quite good as native quadcore where all cores are connected to the same crossbar. The adventage of two dualcore K8 CPUs are the two ODMCs, improving the total RAM bandwidth.
Kentsfield is glued using "the old" Pentium-D technique of glueing singlecores.

4. No, Deerhound will perform better.

4x4 is not restricted to FX. It is not possible to restrict it. FX has one HT link, X2 has one HT link. For any chip they just need to turn on an extra link.

There have been closed system demos where it was said to be 80% faster than one dual core.

Kentsfield will be fast but it is said it will use 125W+ so 4x4 can compete power wise. K8L will kill them both though, BUT FX is supposed to get the K8L core at the end of next year, maybe sooner.

well i remember reading that 4X4 demo...but that system used registered dimms...so it was obviously opty's....and not the actual 4X4 they were testing....and abt that 80% was only while what amd calls as "MEGATASKING".....

i have also read that each processor will have its own dimm's...so that means if its total 4DIMMS each of them gets only 2? is it true....is it gonna affect is performance eh?

Ummm, goofy, if you really read the review you would know that they REPORTED THAT THEY WERE ENGINEERING SAMPLES @ 2.8 AND THAT THE MOBO WAS USING ECC. That would be because there weren't any two socket boards that didn't use ECC. All you have to do is ADD A 940th PIN. And with AM2 you don't even have to do that. You just change the socket and temporarily enable the ECC pin.

And the CEO of VoodooPC was there along with other major companies that wouldn't take well to lies about the chip used. They do want them to support the platform after all.

Face it, around XMas you'll be drooling over my 4x4 system with Vista Business.


That maybe why they said it was an engineering sample. DUMMY.

 

[0zzy]

i aint going to buy either of them as at my place they will cost a fortune...but im very much interested to hear experts say their opinions about them...i have not seen any previews of the 4X4...so here are my qns plzz enlighten me on them...

1>How practial will a 4X4 be?I mean will it be directly competing with the Kentsifeld alone or is amd going to let any 2 athlon 64x2 am2 processors work as 4X4 there by competing with the core 2 duo's in terms of total cost

2>-will the FSB bottle neck the kentsfield enuf to let the 4X4 take the performance crown at the high end..is it even true that FSB will be a bottle neck to the kentsfiled

3>-can the 4X4 work as good as theoretic quad core in the sense it has total 4 cores...i.e. aint it similar to how intel has made their current first gen quadcores...

4>-will the 4X4 give the same performance of an imaginary amd quad core (supposedly the deerhound) with 2 dual cores glued together in same packet jus like intel quad cores

plz let me know ur views...thnx...looking fwd to a good discussion

1> 4x4 and Kentsfield will be two completly different products. One is a quadcore and other is 2 dual cores put together. Any benefits in terms of cost could only be calculated when we have the 2 systems infront of us. We can then do benchmarks, get power readings and heat produced. Please do not give any attention to current benchmarks on 4x4 or kent.

2> Intel has always competed with AMD, with this bottleneck. It will still remain competitive.

3> I am sure 4x4 will work just as good.

4> I am sure there will not be much difference in performance. What i would be looking at is the difference in heat produced by the two and therfore which needs more cooling.

 

[BaronMatrix]

Atleast he is providing links to back up his claim, even if you think they are "crap."

If you were smart rather than clairvoyant a simple search would find the same info your slightly smarter cousin JagOff finds; like this


4x4


A quote:

Running the company's Modo application on the 4x4 systems, Peebler saw an extreme boost in rendering capabilities. "The end result is we get 90 percent boost as we add multiple cores to handle the rendering tasks.

Get a life.