Intel's “Nocona” Xeon Processor TEST

[priyajeet]

<A HREF="http://www.linuxhardware.org/article.pl?sid=04/10/27/204232&mode=thread" target="_new">http://www.linuxhardware.org/article.pl?sid=04/10/27/204232&mode=thread</A>

Lots of up/down graphs - maybe bcos of lack of nice 54bit testing apps ? Opteron does well in most pro benchies.

:tongue: <A HREF="http://www.geocities.com/priyajeet/fing.jpg" target="_new"><i><font color=red>Very funny, Scotty.</font color=red><font color=blue> Now beam down my clothes.</font color=blue></i></A> :tongue:

 

[raretech]

This is good news. This is what I've been waiting for. Nacona did A LOT better than I feared they would at 64bit. It seems competition is alive and well in X86-64. At least performance wise. Though that bit about a possible performance hit above 4gb is disapointing, but nothing for most workstation users to worry about.

I think that was a very good showing by Nacona. Worth noting, is that they compared dual procs with hyperthreading to the dual Opteron:
<i>This is one of only two benchmarks that we'll be showing here which takes advantage of multi-processor systems. In our Mozilla Firefox compile test we pass “-j#” to “make”, where “#” equals to the number of processors plus one. For the Opteron we used “-j3” and for the Xeon we used “-j5” in order to take advantage of the Hyper-Threading feature. As you can see, all three of the results are very close with the Opteron again taking a slight lead.</i>

This is far from conclusive, but it does substantiate my opinion that it won't make that much of a difference in general. Also worth noting is they compared Intel's fastest Xeon to AMD's second fastest, and soon to be third fastest, Opteron. And the current price of 3.6ghz Nacona's start at 950. Price watch lists it for 906 but if you click buy now, the site says 950. 248s start at 670.

Overall though, I'm impressed with Nacona's performance. Much better showing than I feared.




<i>Nemo me impune lacesset</i>

 

[Mephistopheles]

Agreed, it seems the nocona implementation (and therefore the prescott one) isn't so anemic as I was fearing it to be.

I was kind of expecting to find some scores for Opteron in 32-bit to compare the proportional performance increase when going to 64 bit. As is, however, I think this is still an incomplete view of things... :frown:

 

[raretech]

It is an incomplete view of things, true. But I'm less worried about x86 stagnation than I was. This is really good news for AMD64 owners who use windows. Now that Intel has a viable, and in some benchmarks faster 64bit implentation, there shouldn't be any Intel/Microsoft politics to hold back launch of Windows 64bit.

Though I am concerned about the incompatibility mentioned between the two implementations. I think that was a poor choice on Intels part, but we'll see how it plays out. My view is, we're going to see more x86-64 AMD chips shipping than Intel x86-64 chips, for a while at least. For no other reason than cost. Let's hope the incompatibilities will prove to be trivial for windows software developers to overcome, otherwise, I think it will be EMT64 that ends up marginalized, simply because of market penetration of AMD64.

All speculation of course, I'm sure Xeon will be along to tell me how wrong I am. :lol:

<i>Nemo me impune lacesset</i>

 

[Xeon]

Correct, this was a while back but I am sure you remember that great big thread with silver and p4man arguing over the implementation of Intel's take on x86-64 and AMD's.

What came of that was it appeared that Intel made no corrections to x86-64; they just followed AMD's notes on it. With that said it can still be a fair bet that the statement is correct, since in the end all they did was toss links at each other and I do believe p4man yielded further adding support to the "Intel did it the way AMD said it should be done, AMD didn’t do it the way they said it should be done."

Now I personally have an opinion on it but it really doesn’t matter, with the simple fact that I really don’t care about x86-64 or what directions it headed in. I was actually looking forward to seeing Intel’s implementation suck balls truthfully.

Just because "my god" supports the technology doesn’t mean I find it anymore useful or interesting, least dicko21 will agree with that.

Xeon

<font color=red>Post created with being a dickhead in mind.</font color=red>
<font color=white>For all emotional and slanderous statements contact THG for all law suits.</font color=white>

 

[trooper11]

well im glad it works, otherwise the only choice to the consumer would be itanium lol. i would htink youd be happy thre is a low cost alternative to put pressure on intel to move more decisively to push itanium's IA 64 instructions. competition is about the only thing to give them a kick in the pants so to speak.

 

[raretech]

If he wants Itanium on the desktop, this is really bad news for that in the short term. In the long term, this could prove to be a smoother transition to Itanium.

For example, from what I understand about the Linux kernel guys, they want X86-64 to succeed, to enable a smoother transition to 64bit.

<i>Nemo me impune lacesset</i>

 

[Xeon]

If he wants Itanium on the desktop

No IA64 I dont want a CPU with a billion transistors.

Xeon

<font color=red>Post created with being a dickhead in mind.</font color=red>
<font color=white>For all emotional and slanderous statements contact THG for all law suits.</font color=white>

 

[Mephistopheles]

I dont want a CPU with a billion transistors.

I don't care if it has 100 million, 1 billion, 100 billion or even like 50 transistors in it, as long as it does what I want it to: it runs reasonably quiet, runs fast and runs what I want it to run.

 

[AMD101]

That new Xeon actually isnt that bad but I wonder why they compared it to 2.2 opteron's rather than 2.4's which are widely available? Oh well hopefully someone will do a nice full review and compare on the 252's and the 3.6 xeon in the near future when the 252's are released.

 

[Mephistopheles]

Hm, the 2.4Ghz opterons (250s) probably give the xeon quite a run for their money (mind you, 2.4Ghz is one hell of a clock speed for an A64 processor!)...

If AMD gets the 252 (2.6Ghz) opterons out now, they'll probably make nocona look like a very sad processor... even if they do manage a 3.8Ghz Nocona by then, it's not nearly enough to hold a candle to a fully-fledged 2.6Ghz opteron. Isn't the 3.8Ghz nocona due out soon too?...

This general picture is good, because it might just convince Intel to reach for better designs than netburst and so on...... Xeon will be dead in the water, unless they repackage yonah (!) and tweak it for performance rather than power efficiency. And yonah would make a good precursor to the quad-core Whitefield xeon, which is a departure from netburst altogether. (whitefield is also expected to be P-M based, but highly tweaked and at 65nm. 4 current dothan cores still don't exceed 100W right now, therefore this idea makes a lot of sense)

In any case, Intel would be very stupid not to repackage dothan and sell it to the world ASAP. Current dothan architecture is good for at least 2.4, 2.5Ghz... and there's always the added performance of desktop-class components. Add x86-64 to yonah, and you've got a pretty damned good processor...

(it has been demonstrated that an overclocked 2.4Ghz dothan is currently quite able to hold its own even against 3+Ghz-rated processors! It beats them at several benchmarks too.)

 

[raretech]

No IA64 I dont want a CPU with a billion transistors.

Then what do you want on the desktop? You don't seem to like X86-64...

<i>Nemo me impune lacesset</i>

 

[darko21]

<<Hm, the 2.4Ghz opterons (250s) probably give the xeon quite a run for their money (mind you, 2.4Ghz is one hell of a clock speed for an A64 processor!)...>>

Yes, I,m sure since the 248 won more test in that article over the rare xeon 3.6 we can draw the conclusion the widly available 250 would have done even better. The real question is why do it that way, why not use the widly available 250?.

To top it off they show the price of a 250 for comparison which they never used (why) and the price of a p4 3.4 which they never used (why). BTW 2.4 ghz might be a good clock speed for an a64 but they are now up to 2.6 Ghz with the fx on .13 so I'd guess 2.6 for opteron should be very soon, too bad they only used 2.2Ghz.

Maybe they can even the score soon, when the 2.6Ghz opteron is released. You know do an article using a 2.6 Ghz opteron to a 3.4 Ghz xeon then close the article by showing what a 3.6Ghz xeon costs and a 2.2 Ghz opteron costs.

Think that might happen? Personaly I don't think a website like amdzone would do somthing as biased as that. IMHO anyway. Keep in mind that 2.4 Ghz is a high a64 speed but 3.6 Ghz is a crazy prescott speed just think how many cpu's must be produced to cherry pick those samples.

If I glanced at a spilt box of tooth picks on the floor, could I tell you how many are in the pile. Not a chance, But then again I don't have to buy my underware at Kmart.

 

[Spitfire_x86]

He wants P4 with IA64

------------
<A HREF="http://www.foood.net" target="_new">FOOOD's Icons</A>

<A HREF="http://geocities.com/spitfire_x86/myrig.html" target="_new">My Rig</A> & <A HREF="http://geocities.com/spitfire_x86/benchmark.html" target="_new">3DMark score</A>

 

[raretech]

I get it. Well too bad, not going to happen any time soon now. Nacona did a good showing, X86-64 is here to stay for the foreseeable future.

Neener neener neener Xeon. :tongue:

<i>Nemo me impune lacesset</i>

 

[Xeon]

Please be reasonable when you try and speculate what I want in a processor. I want to see an OOO IA64 derived processor not a P4 with IA64.

Xeon

<font color=red>Post created with being a dickhead in mind.</font color=red>
<font color=white>For all emotional and slanderous statements contact THG for all law suits.</font color=white>

 

[priyajeet]

thats never gonna happen. P4 uses ASIC and IA64 uses EPIC, totally diff rivers. Will never join.

:tongue: <A HREF="http://www.geocities.com/priyajeet/fing.jpg" target="_new"><i><font color=red>Very funny, Scotty.</font color=red><font color=blue> Now beam down my clothes.</font color=blue></i></A> :tongue:

 

[raretech]

I took that to mean a redone IA64 chip for the desktop. I guess I didn't understand what Spitfire meant.

<i>Nemo me impune lacesset</i>

 

[Mephistopheles]

Isn't it conceivable that what is making prescott absurdly hot is the presence of IA64-code execution hardware? It has already been previously stated that tejas was to incorporate IA64 execution hardware, and Intel has the tradition to add disabled circuitry on their chips like HT and enabling them later (HT was actually even on willamette chips, just broken).

This would make the whole "staying with netburst for smithfield" debate much more understandable. I mean, if prescott was so hot only with necessary circuitry and not counting disabled circuitry, then smithfield being netburst would mean suicide. But if they suddenly chop a lot of the transistors and make a better design out of it, they might just manage to lower the 6xx' series heat (remember, they're tejas-like prescotts) and smithfield wouldn't be so far-fetched as it seems to be for us. (still not quite so good as yonah for LGA775, though, but whatever)