Deneb 2,2 GHz --> 4,0 GHz

I think you are confusing the order of magnitude of some things here.
Cache misses occur at the lowest level of code: instruction level.
It was just one sample of what could be a problem, threads starting upp, threads inititalized, threads waiting for other threads, spinlocks moved from one cache to another etc. On what core threads are placed, there are many things that could delay the game.
Have you checked memory consumption for crysis warhead?
 
Would this be a bad time to mention this Nehalem overclocked to 4.2 GHz?
http://my.ocworkbench.com/bbs/showthread.php?p=438583#post438583

Yeah but it required unlocked multiplier in CPU there were zero fsb gains in there.

While AMD got to 260 fsb which means even non BE deneb would be able to reach 2.8-2.9 Ghz on good mobo.

Now compare the prices of Intel extreme edition cpus and amd BE cpus and normal mortals won't be running 4 Ghz nehalems anytime soon.

If I can get Deneb which will run at 3.4-3.6 Ghz at prices comparable to Q6600 then I'm jumping in.
 
It was just one sample of what could be a problem, threads starting upp, threads inititalized, threads waiting for other threads, spinlocks moved from one cache to another etc. On what core threads are placed, there are many things that could delay the game.

What was your point again?

Have you checked memory consumption for crysis warhead?

Again, what is your point?
 
Yeah but it required unlocked multiplier in CPU there were zero fsb gains in there.

We don't know that.
We know that they USED a CPU with unlocked multiplier.
But we still have no information whatsoever on how well you could overclock with other models.

While AMD got to 260 fsb which means even non BE deneb would be able to reach 2.8-2.9 Ghz on good mobo.

Which wouldn't be very impressive if performance is indeed at around Kentsfield levels.
You still wouldn't be able to outperform an overclocked Q6600, and I doubt you could beat the value-for-money either, because I don't see how AMD could sell their new 45 nm CPUs at the super-low prices of Q6600 and still get a return on their 45 nm fab investments. Heck, AMD hasn't broken even in ages. I'm quite sure that their new investors will want AMD to stop bleeding cash.

Now compare the prices of Intel extreme edition cpus and amd BE cpus and normal mortals won't be running 4 Ghz nehalems anytime soon.

We don't know that.
Currently you don't need an unlocked Extreme Edition to overclock a Core2 Quad to 4+ GHz either.
It could well be that non-Extreme Nehalems can also be overclocked to ~4 GHz.
 
There are a lot of other things that we don't know yet but since you mentioned this 4.2GHz overclocked Nehalem in this Deneb thread, I'm also just mentioning that this 3.2GHz unlocked Nahalem needs 1.72V to reach the 4.2GHz OC speed. Imo it is not as impressive as this 2.2GHz Deneb OCed to 3.975GHz with just 1.6V on air. I'm not making any conclusion yet but I'm just pointing out the facts between these two CPUs.
 
Imo it is not as impressive as this 2.2GHz Deneb OCed to 3.975GHz with just 1.6V on air.

That's a conclusion you're drawing. It's not a fact, because of reasons that I already stated in my previous posts.
Put another way: we don't know enough about either CPU (or how the overclock was achieved exactly) to draw any conclusions on how impressive either OC is.
 
That's a conclusion you're drawing. It's not a fact, because of reasons that I already stated in my previous posts.
Put another way: we don't know enough about either CPU (or how the overclock was achieved exactly) to draw any conclusions on how impressive either OC is.

What is not a fact? 2.2GHz Deneb OCed to 3.975GHz with just 1.6V on air or 3.2GHz unlocked Nahalem needs 1.72V to reach the 4.2GHz OC speed? :confused:

Like I said before, based on the facts we already have, I think that the particular OC on that particular Nahalem is not as impressive as the particular OC on that particular Deneb.
 
What is not a fact? 2.2GHz Deneb OCed to 3.975GHz with just 1.6V on air or 3.2GHz unlocked Nahalem needs 1.72V to reach the 4.2GHz OC speed? :confused:

"I'm not making any conclusion yet but I'm just pointing out the facts between these two CPUs."
That reads as if you consider it a fact that the Deneb overclock is more impressive.
It's not a fact, it's an opinion.

Like I said before, based on the facts we already have, I think that the particular OC on that particular Nahalem is not as impressive as the particular OC on that particular Deneb.

How exactly do you arrive at that conclusion?
 
If (thread in section "AMD Processors" > 2 pages)
{Set Scali2 mode to "troll";}
Else
{exit;}
 
How exactly do you arrive at that conclusion?

Because it's AMD :p.

No, seriously, I EXPECT Nehalem to hit ~4ghz based upon Intel's Penryn performance (different architectures be damned!)... I was similarly expecting ~3.2ghz for Deneb based upon the Phenom lineup so far.

If (thread in section "AMD Processors" > 2 pages)
{Set Scali2 mode to "troll";}
Else
{exit;}

Let's get that in Machine code.
 
I'll tell you what Scali2, you really need a professional help. I just mentioned the facts from both of the OC. Then I said that I'm more impressed with facts from the Deneb OC than with facts from the Nahalem OC. If you are not impressed like me then fine. No one asks you to be impressed like me.

Would you have a problem too if a straight person is impressed with a hot naked woman but not with a tough looking guy or if a gay person is impressed with a tough looking guy but not with a hot naked woman?

Btw to be impressed by a hot naked woman, I don't need to know whether the boobs are real or not and I also don't need to know whether the woman is married or not . ;)
 
No, seriously, I EXPECT Nehalem to hit ~4ghz based upon Intel's Penryn performance (different architectures be damned!)... I was similarly expecting ~3.2ghz for Deneb based upon the Phenom lineup so far.

I don't know what to expect yet.
For AMD we have no indication of what its 45 nm process is capable of yet.
With Intel we know that they can produce CPUs that can do 4+ GHz on 45 nm... but Nehalem is a very different architecture, so it's too early to say how its clockspeed scaling relates to Penryn.

I mean, if it turns out that most Denebs run 4+ GHz with ease, and Nehalems can only hit 4+ GHz with cherry-picked CPUs and extreme cooling solutions, then Nehalem was the impressive one, right?
But we can't really draw any conclusions from just one CPU, and little info on the cooling solution or stability or anything.
 
I'll tell you what Scali2, you really need a professional help. I just mentioned the facts from both of the OC. Then I said that I'm more impressed with facts from the Deneb OC than with facts from the Nahalem OC. If you are not impressed like me then fine. No one asks you to be impressed like me.

I just asked you why you were impressed. Why are you avoiding the question? Why resort to personal insults?

Btw to be impressed by a hot naked woman, I don't need to know whether the boobs are real or not and I also don't need to know whether the woman is married or not . ;)

And would you answer if I asked what it is about her that impresses you?
 
Most people already know how I think about this whole forum section thing.
When one clicks the AMD section that person may want to read about AMD, he/she should go to the Intel section reading about Intel. VERY simple logics
 
When one clicks the AMD section that person may want to read about AMD, he/she should go to the Intel section reading about Intel. VERY simple logics

Really? Then why is it okay to mention Intel in the AMD section when you want to talk about competition, keeping prices low and all that... or claim that Deneb is a better gaming CPU than Nehalem?
And when you post an overclocked Nehalem it's suddenly all wrong to discuss Intel?

Seems like there's some hypocrisy going on. You yourself have compared AMD to Intel often enough, trying to claim that AMD has better gaming CPUs, or AMD is better value for money.
You can't make such claims in either forum section, if you follow your own rules.
 
I just asked you why you were impressed. Why are you avoiding the question? Why resort to personal insults?



And would you answer if I asked what it is about her that impresses you?

I'm impressed with the facts, the facts that the CPU only needs 1.6V to make the jump from 2.2GHz to 3.975GHz, which is around 80% of core speed increase and it was done on air.

As for the woman, she is hot and NAKED :eek: Do I need more reason to be impressed?
 
Really? Then why is it okay to mention Intel in the AMD section when you want to talk about competition, keeping prices low and all that... or claim that Deneb is a better gaming CPU than Nehalem?
And when you post an overclocked Nehalem it's suddenly all wrong to discuss Intel?

Seems like there's some hypocrisy going on. You yourself have compared AMD to Intel often enough, trying to claim that AMD has better gaming CPUs, or AMD is better value for money.
You can't make such claims in either forum section, if you follow your own rules.
To make it simple for you! start a new thread and set the title to "i7 vs deneb"
 
I'm impressed with the facts, the facts that the CPU only needs 1.6V to make the jump from 2.2GHz to 3.975GHz, which around 80% of core speed increase and it was done on air.

Well, my point was that it depends on the circumstances, which we don't know yet.
For example, I consider a Phenom at 3.5 GHz more impressive than a Core2 Duo at 4 GHz.
The Core2 Duo might have a larger overclock in terms of core speed increase, but it means less because so many Core2 Duo's can be overclocked to such speeds that easily. With the Phenom it's much harder to do.
 
And should that be posted in the AMD or in the Intel section? You're confusing me now.
That depends on what type of talk you want

2.2 to 4.0 is a VERY good clock if you want to clock the processor.
 
That depends on what type of talk you want

What do you mean, "what type of talk you want"?
There is only one way to compare CPUs, is there not?

I don't see HardOCP writing two CPU reviews, one for the AMD section, one for the Intel section. They just do their comparisons in one review.
 
Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll. Troll.
What? I'm sorry, I can't hear what you're saying. :confused:
 
Yeah but it required unlocked multiplier in CPU there were zero fsb gains in there.

Actually, the multiplier was lowered, not raised. Stock settings for the 965 are 24*133=3192MHz, and that CPU was running at 21*200=4200MHz. So you're completely wrong.
 
The early indications are that AMD made up some ground with Deneb, but in all honesty I think that Nehalem will still be the faster CPU in most tasks.

The realistic and honest question is, will that speed difference (which I'm betting will be about 10-20 percent on average) translate to a recognizable difference to the average user? I predict that at similar clock speeds, you could blindfold a user and put them down on either machine and not be able to tell which machine they were running in a side-by-side comparison.

I'm glad AMD made up some ground.. They need to keep it up though.
 
omg, what trolling. lets get back on topic shall we?

i consider deneb overclocks in general to be better than expected, and if this particular overclock is true, its good news for all of us since it would prove that deneb has more headroom than expected

intel fanboys jsut wont accept that amd can do any good, and they dont seem to want it even though it would benefit them if AMD did good.
 
Kyle himself said that just about any multi-core chip over 3GHz is good enough for gaming, and at that point you're hitting the limits of the video card. (Kyle, please correct me if I am incorrect)

So, gaming benchmarks=trash. The only real benchmark that will matter with these chips is non-gaming applications, like encoding.

I myself am now looking at AMD systems after a long period spent considering Intel. Currently I run a 5000BE@3GHz with an 8800GT pushing 1280*1024, 4GB memory. This system plays everything I have just fine. The only place I find it lacking is encoding and other non-gaming tasks. Even then, it's not horrible. So, how can I justify a full Intel build at around $300/400 new, when I could go AMD for about $100 less? And honestly, gaming won't change much, and I'll gain performance in non-gaming applications.

Look at this with logic. The benchmarks, the "leaked" info-it's all marketing.
 
Most people already know how I think about this whole forum section thing.

Well, the forum division is there for a reason. Might I suggest following it? And what you think about "this whole forum section thing" is not entirely relevant to this thread ;)
 
Well, the forum division is there for a reason. Might I suggest following it?

Well, perhaps you can clear it up for me then.
How come people can come here talking about how Deneb is a better gaming CPU than Nehalem? That's discussing Intel in the AMD section.
Then when I post some overclocking results of Nehalem (which might indicate that Nehalem could be a better gaming CPU than some people were thinking, so relating to the discussion earlier in the thread), suddenly you can't mention Intel.

The way people interpret it here seems to be the following: people are only allowed to mention Intel in the AMD section if it's in a negative relation to AMD, and vice versa.
Is that the reason why there's a forum division?
Because I don't see how that could possibly work. If you wanted to compare two CPUs, such as Deneb and Nehalem (like the thread people were suggesting), you could only mention Deneb's strong points and Nehalem's weak poitns in the AMD section, but only mention Nehalem's strong points and Deneb's weak points in the Intel section.
So in the end neither thread would provide a reasonable overall comparison.
And as I pointed out, why do you have two forum sections then, if you don't write two articles when you review a CPU? How exactly is such a review to be discussed on your own forum?

Please clear it up for me, as I have no idea anymore what I can post, where, how etc...
 
Well, perhaps you can clear it up for me then.
Please clear it up for me, as I have no idea anymore what I can post, where, how etc...

Writing in a thread means that you need to think about what others want to read and what could be interesting for others. When you write I think that you have one goal and that is to bash AMD. You have said it before that you think Intel is better on everything, you have said that you think processors might be better if there wasn't any competition (Intel was alone). Saying these things and some others, you just make people stop listening to you because most understand that your opinions are not backed up with facts.
That the reality. If you have one opinion that differ from others you need to have A LOT of proof that you are right.
Most of us in the AMD section thinks that competition is good, maybe 99,99% of the science in this field is saying the same. Most of us know that AMD is cheaper compared to Intel, most of us know Nehalem is a server processor and optimized for that and most of us understand that Nehalem is expensive when we look at the prices for hardware needed to that processor.
 
I think I've given ample of arguments, fact or proof, as other people have attested to in a previous thread where you accused me of the same. In many threads it seems that the other people cannot counter any arguments, and fall silent (I can recall one instance with you involved, where you claimed that AMD had superior core-to-core communication, but you failed to back up your claims with facts. I then pointed you to the opensource cache2cache benchmark, which proved that the Core2 had the best core-to-core communication, so if I were you, I wouldn't go down that road really).

You also have to understand that if I follow your logic, and post what *I* think others find interesting to read, that does not necessarily mean I arrive at the same things that *you* find interesting to read.
Namely, I think most people here would like an unbiased, balanced view of things, with background info from people with a lot of knowledge and experience on the subject. I think they come here because they want to learn how a CPU works, what makes it perform well, what makes it good or bad at certain tasks, and how software is developed, optimized etc.
When I write software, I don't limit myself to a single CPU manufacturer either. I either try to write code that is balanced so it runs well on CPUs of all (recent) types from all manufacturers, or I actually write multiple code paths, each optimized for different classes of CPUs.
 
SCALI:
There is another BIG CLUE that you can use in order to know what others might think is interesting and that is the title of the thread. I think that you use threads as some sort of chat area and you have a big problem with anyone that says that AMD have strong points compared to Intel
 
SCALI:
There is another BIG CLUE that you can use in order to know what others might think is interesting and that is the title of the thread. I think that you use threads as some sort of chat area and you have a big problem with anyone that says that AMD have strong points compared to Intel

I only have problem with people saying AMD has strong points compared to Intel (or vice versa) when it's patently false, such as your core-to-core communication claims.
That just seems to happen a lot in the AMD forums. It seems like people are hell-bent on finding any strong points of AMD, no matter how nonsensical or convoluted. You are one of the biggest proponents of that actually.
I just come in to interject some facts and common sense, and put things in a realistic perspective.

I have no problem if people claim that AMD systems have more memory bandwidth than Core2, or that AMD's hyperlinks make it scale better than Intel's FSB for multiple CPU systems. That's simply true, and easy to verify.
 
I only have problem with people saying AMD has strong points compared to Intel (or vice versa) when it's patently false, such as your core-to-core communication claims.

I can't run vmware workstatioin on my new Dell Latitude E6400 (P8400 @ 2,2 Ghz, 4 GB ram, 320 GB 7200RPM disk) because it is to slow. Exactly the same vmware machine (you just move files from one computer to another) runs very good on all AMD computers. Have the same setup for some computers here. X2 processors run this well too. On the dell I had to install applications etc running on the main operative system.

There is a reallity out there that people need to adapt too even if you think that Intel is best.
 
I can't run vmware workstatioin on my new Dell Latitude E6400 (P8400 @ 2,2 Ghz, 4 GB ram, 320 GB 7200RPM disk) because it is to slow. Exactly the same vmware machine (you just move files from one computer to another) runs very good on all AMD computers. Have the same setup for some computers here. X2 processors run this well too. On the dell I had to install applications etc running on the main operative system.

Even if that were true (I wouldn't know about that, and never said anything on that matter, but pxc already commented on that matter), that still has nothing to do with your unfounded claims about core-to-core communication or your 'analysis' on cache usage in game engines etc.
 
Even if that were true (I wouldn't know about that, and never said anything on that matter, but pxc already commented on that matter), that still has nothing to do with your unfounded claims about core-to-core communication or your 'analysis' on cache usage in game engines etc.
That is another discussion from this thread, this thread was about clocking the deneb.
moving data on from one core to another core C2Q isn't fast if core are located on the other C2D (C2Q = 2xC2D)
 
Back
Top