X2 and C2D Question

[kaen]

Wondering what are the specific differences between the X2's and C2D's that allow C2D to overclock much higher? Is it the 90nm vs 65nm, caches, etc?

 

[Pippero]

Nope guys, he was talking about clock, not general performance.
The pipelines of both CPUs are pretty similar in length, of course the number of pipeline stages give only a rough estimate about the clocking possibilities, because the architectures are so different.
Anyway, i'd say that the main difference is in terms of process.
It's not just that Intel is using a 65nm process, but it's using an industry leading 65nm process.

 

[enforcer22]

wish i could see a c2d blow away a x2 :-/ unless the relativly small difference to you is blowing something away.


ANYWAYS it also is going to depend on the chip you get. Not all the chips in the same series are going to OC the same since the manufacturing is never exactly the same as the chip before and after ect.

 

[gOJDO]

First of all the C2Ds blow away the X2s even before OCing.

Hmmm...I think a more accurate statement would be that "some", "C2D's" blow away some "X2's". Its also equally accurate to say some "X2's" blow away some "C2D's"

Even more acurate will be:
Some C2D are blowing all AMD CPUs. Its also equally accurate to say no AMD CPU can blow away some C2D's.

Its architecture, the same way A64 and X2 blew away pentium 4 and pentium D for three years straight.

Wasn't it more acurate to say "some A64 and X2 blew away some Pentium 4's and some Pentium D's"?

 

[purdueguy]

From the articles I've read on AMD's Athlon 64 chips that have 65nm technology, they still can't shake a stick at the Core 2 Duo.

Take a look at Anandtech's review, here.

Prozac26 is correct, the main difference is in the architecture.

 

[Pippero]

Hmmm...I think a more accurate statement would be that "some", "C2D's" blow away some "X2's". Its also equally accurate to say some "X2's" blow away some "C2D's"

I think a more accurate statement would be that "some" C2Ds blow away all "X2s"...

Its architecture, the same way A64 and X2 blew away pentium 4 and pentium D for three years straight.

The OP said:

Wondering what are the specific differences between the X2's and C2D's that allow C2D to overclock much higher?

Now, explain me how the architecture of A64 and X2 allowed them to blow away Pentium 4 and Pentium D for 3 years in terms of clock speed and overclocking... :roll:

 

[enforcer22]

I guess when people think blowing something away blowing something away now doesnt really mean much. I seen the benchmarks and i dont see anything blowing anything away looks more like a nice little lead.

However if we want to look at the p-4 side the amd cpus at that time (if we go clock per clock) didnt just blow them away they nuked them and blew thier dust into outter space. But since AMD isnt a retarded company they didnt waste thier time building cpus at the same clock speed simply to destroy intels chips. No point all that money spent to prove something everyone already knew.

Seriously if this is all it takes to blow something away the term blowing something away loses alot of meaning.

 

[Pippero]

Guys, here we should not confuse the performance which comes from IPC, from the performance which comes from clock speed.
The architecture of Core 2 gives it the advantage in terms of IPC (clock for clock), not in terms of overclocking.
It's the process which gives it also the advantage in terms of clock speed.
Concerning AMD's 65nm lackluster overclocking: Jack time ago provided some data which showed how even the best published results for AMD's 65nm process would be 10-15% worse than Intel's.
This means that, even when AMD's 65nm will be fine tuned and with all the performance enhancement in, they will still have a clock speed disadvantage of 10-15% (so for example, 3.5-3.6GHz Vs 4.0GHz).
But the first revision of AMD's 65nm is not as mature as Intel's current 65nm process, i.e. the difference can easily be 20% or more (for example, 3.2GHz Vs 4GHz.. which is more or less what we can observe right now).

 

[epsilon84]

I guess when people think blowing something away blowing something away now doesnt really mean much. I seen the benchmarks and i dont see anything blowing anything away looks more like a nice little lead.

29% is a 'nice little lead' indeed. :roll:

I don't know what 'schlagt' means in German but it sounds remotely like 'slaughter' and that is exactly what C2D does to AMD. They are not even in the same league.

 

[m25]

Wondering what are the specific differences between the X2's and C2D's that allow C2D to overclock much higher? Is it the 90nm vs 65nm, caches, etc?

Responsible, more than the architecture, cache etc, is Intel's strong 65 nm process. The latest netburst pieces, the Celeron Ds also show this strength and by this time Intel could have launched the ultra-4GHz P4 (if they were interested in this and you course they're not)

 

[Pippero]

I don't know what 'schlagt' means in German but it sounds remotely like 'slaughter' and that is exactly what C2D does to AMD. They are not even in the same league.

Nah, 'schlagt' simply means "defeats", or "beats"..

 

[Pippero]

Finally someone who gets it. *applauds*
But in case of Netburst, it was also the architecture which allowed such insane clock speeds, since the pipeline had twice as many stages as Core 2 and K8.

 

[enforcer22]

your right in one area its not in the same league since the core2duo is next gen and the fx62 is not. But hell if i took intels top cpu and put it againced amd's mid top rang cpu i expect it to win by a considerable magin to.

so i share in your :roll:

 

[nategarst]

First of all the C2Ds blow away the X2s even before OCing.

Hmmm...I think a more accurate statement would be that "some", "C2D's" blow away some "X2's". Its also equally accurate to say some "X2's" blow away some "C2D's"

Even more acurate will be:
Some C2D are blowing all AMD CPUs. Its also equally accurate to say no AMD CPU can blow away some C2D's.

Its architecture, the same way A64 and X2 blew away pentium 4 and pentium D for three years straight.

Wasn't it more acurate to say "some A64 and X2 blew away some Pentium 4's and some Pentium D's"?

I think it is accurate that you said this.

 

[Barcelona_Xtreme]

Hmmm...I think a more accurate statement would be that "some", "C2D's" blow away some "X2's". Its also equally accurate to say some "X2's" blow away some "C2D's"

I think a more accurate statement would be that "some" C2Ds blow away all "X2s"...

Its architecture, the same way A64 and X2 blew away pentium 4 and pentium D for three years straight.

The OP said:

Wondering what are the specific differences between the X2's and C2D's that allow C2D to overclock much higher?

Now, explain me how the architecture of A64 and X2 allowed them to blow away Pentium 4 and Pentium D for 3 years in terms of clock speed and overclocking... :roll:

Yeah, right!

I can easily setup a benchmark that it'll make a quad core Kentsfield look like crap compared to a dual core X2 machine. Does it mean anything for anyone else? I doubt it.

Intel has lower power consumption, significantly better performance on CPU intensive consumer applications, more overclocking headroom and more bragging rights for the buyer, but remember that Non-SSE applications run faster on AMD.

X2s excel in applications that require high bandwidth. They also are suitable for encryption algorithms thanks to it's strong FPU and complex decoders. There are many apps which an X2 can beat hands down a C2D (as long as it's not SSE) and viceversa.

 

[1Tanker]

Hmmm...I think a more accurate statement would be that "some", "C2D's" blow away some "X2's". Its also equally accurate to say some "X2's" blow away some "C2D's"

I think a more accurate statement would be that "some" C2Ds blow away all "X2s"...

Its architecture, the same way A64 and X2 blew away pentium 4 and pentium D for three years straight.

The OP said:

Wondering what are the specific differences between the X2's and C2D's that allow C2D to overclock much higher?

Now, explain me how the architecture of A64 and X2 allowed them to blow away Pentium 4 and Pentium D for 3 years in terms of clock speed and overclocking... :roll:

Yeah, right!

I can easily setup a benchmark that it'll make a quad core Kentsfield look like crap compared to a dual core X2 machine. Does it mean anything for anyone else? I doubt it.

Intel has lower power consumption, significantly better performance on CPU intensive consumer applications, more overclocking headroom and more bragging rights for the buyer, but remember that Non-SSE applications run faster on AMD.

X2s excel in applications that require high bandwidth. They also are suitable for encryption algorithms thanks to it's strong FPU and complex decoders. There are many apps which an X2 can beat hands down a C2D (as long as it's not SSE) and viceversa.Provide some links...please.

 

[gOJDO]

Yeah, right!

I can easily setup a benchmark that it'll make a quad core Kentsfield look like crap compared to a dual core X2 machine. Does it mean anything for anyone else? I doubt it.

Intel has lower power consumption, significantly better performance on CPU intensive consumer applications, more overclocking headroom and more bragging rights for the buyer, but remember that Non-SSE applications run faster on AMD.

X2s excel in applications that require high bandwidth. They also are suitable for encryption algorithms thanks to it's strong FPU and complex decoders. There are many apps which an X2 can beat hands down a C2D (as long as it's not SSE) and viceversa.

LameNoobMike?

 

[Barcelona_Xtreme]

Hmmm...I think a more accurate statement would be that "some", "C2D's" blow away some "X2's". Its also equally accurate to say some "X2's" blow away some "C2D's"

I think a more accurate statement would be that "some" C2Ds blow away all "X2s"...

Its architecture, the same way A64 and X2 blew away pentium 4 and pentium D for three years straight.

The OP said:

Wondering what are the specific differences between the X2's and C2D's that allow C2D to overclock much higher?

Now, explain me how the architecture of A64 and X2 allowed them to blow away Pentium 4 and Pentium D for 3 years in terms of clock speed and overclocking... :roll:

Yeah, right!

I can easily setup a benchmark that it'll make a quad core Kentsfield look like crap compared to a dual core X2 machine. Does it mean anything for anyone else? I doubt it.

Intel has lower power consumption, significantly better performance on CPU intensive consumer applications, more overclocking headroom and more bragging rights for the buyer, but remember that Non-SSE applications run faster on AMD.

X2s excel in applications that require high bandwidth. They also are suitable for encryption algorithms thanks to it's strong FPU and complex decoders. There are many apps which an X2 can beat hands down a C2D (as long as it's not SSE) and viceversa.Provide some links...please.

Intel Versus AMD: The Truth Unfolds

 

[dasickninja]

You cite an AMD fanboy as a source?

 

[Barcelona_Xtreme]

Just to add insult to injury, An X2 3800 @2GHz can be 20% faster than a E6600 @2.4GHz on (single-threaded) Reed-Solomon codes, or roughly equal on most openssl libraries. Why? Probably because in these cases, L2 cache size/speed doesn't matter as much as L1 cache size/speed; nor do these programs improve much by reordering loads and stores. By using a large shared L2 cache and including the memory-reordering logics, Core 2 actually introduces some overheads that may have negative impact on workloads that do not benefit from them.

BTW, a (255, 128) Reed-Solomon code can correct up to 64B/512bit errors in a 128B/1024bit message. It takes a tremendous amount of CPU power. RS codes are used in CD/DVD/HDTV/DSL etc. Talking about a media server - this is what you'll definitely use.

Not to mention cryptography is getting more importance by the day. When is the last time a website compared Core 2 and K8 with a standard cryptography, anyway? (I know a few cryptos can be optimized for SSE, but pretty much none of the standard ones can).

 

[Barcelona_Xtreme]

You cite an AMD fanboy as a source?

Have you read his entire blog to conclude that he's a fanboy?

He's far away of being like Sharikou which I admit that he's a die-hard AMD fanatic and which I don't even read his crap.

Scientia has hard facts and until this day, no one has contradict him.
Anyone can enter his blog to offer a rebuttal to his articles.

 

[evilr00t]

You don't know what the hell you're talking about, fanboi.

BTW, a (255, 128) Reed-Solomon code can correct up to 64B/512bit errors in a 128B/1024bit message. It takes a tremendous amount of CPU power. RS codes are used in CD/DVD/HDTV/DSL etc. Talking about a media server - this is what you'll definitely use.

Maybe these Reed-Solomon codes are... *gasp* implemented in HARDWARE, not SOFTWARE?

Not to mention cryptography is getting more importance by the day. When is the last time a website compared Core 2 and K8 with a standard cryptography, anyway? (I know a few cryptos can be optimized for SSE, but pretty much none of the standard ones can).

Don't forget VIA!! The Via pwnz all at encryption; K8, C2D can't compete here.

 

[Barcelona_Xtreme]

If you say so... :roll:

 

[evilr00t]

If you say so... :roll:

Lemme guess, you're now going to claim that hard drives use the CPU to calculate their ECC?

 

[cb62fcni]

Oh crap, and I heard Crysis is gonna use Reed-Solomon algorithims in its code! It probably won't even run on a C2D!!!