Woah: 3GHz on 4800 X2...on air!

[Jcollins82]

First, I'd like to get this out of the way. Obviously, I respect the things that Fishy and Sen say. Not because we even know each other, cause we don't, but because I've learned alot from their threads which is a pretty big deal to me as I am making a pretty big push at the moment to try and learn about computers.

So, that out of the way, I'd like to say that I think comparing any DC CPU to any single core CPU makes little sense, at least from the way I understand things.

First, it's obvious that both intel and amd are having troubles getting these things to clock at single core speeds, a simple look at their product offerings would tell you that. Do you think that they would purposely sell you a cpu that's going to run slower than they could? That they would pass up the opportunity to "one up" the other guy? That they would give up the opportunity to legitimately charge more money?

Second, I really think that trying to compare the far more matured manufacturing process of 130nm to the relatively newer 90nm offerings is a bit more difficult than at first glance. As I understand it, as a process matures, their yields improve, as far as speeds go as well as functional chips. And while, obviously, 90nm has several advantages over 130nm chips it's also alot younger, which I believe is why the Winchesters refused to overclock as well as their predecessors, despite being 90nm, and it's also the reason the new revision E chips the gap has been closed.

That might make you make you think that 90nm has come far enough so that original difference in 90nm to 130nm doesn't matter anymore, and I would agree, if we were still talking about the same physical chips but these things are entirely different. They essentially have to relearn the manufacturing process, for lack of a better term.

Also as I understand it, 90nm has problems of it's own to overcome, in that while they require less voltage to run and thus generate less heat, the interconnects and transistors are closer together thus regaining some of that dissipated heat which is why CPU clock speed growth has been slowing down as compared to previous generations of processes. From what I've read, it's one of the main reasons amd and intel have chose to begin dc production. Most people, myself included, would logically expect to see that these things are creating more heat and sucking up more voltage. And combined with the already established fact that both amd and intel are not able to match their single core yields speed-wise consistently enough, I'd say that the fact they can even do a 3.0GHz suicide shot is pretty impressive at this point in the game.

I think it will be pretty interesting to see where the next few years take us because on the one hand, you have amd who has the obvious lead as far as dc architecture goes with their HT link between processors and so on but on the other hand you have intel who with the P-M have made leaps in heat dissipation, power usage and so on which I think will play a major role in the ability to make DC a more feasible solution for years to come.

Now, keep in mind, I'm a newb. Didn't really start reading about computers till about March of this year. So, if I'm wrong go easy. But I would like to be corrected if I'm wrong.

 

[Otter]

2.8GHz on air is impressive because of the heat generated. That OC is in the same range as the rev E single cores, and many of us thought the dual cores wouldn't overclock as well without phase. It's too soon to tell if this will be a typical OC, but it's an ecouraging start.

I think there will be some good 4400+ OCs. These chips are divided into only two bins, and I don't imagine AMD is going to sell very many 4800s at $1000 each. Hence, some of the 4400s will as good. Of course, the worst of the chips go into the 4400 bin too, so it's a crapshoot as usual.

As for all this AMD vs. Intel fanboy stuff, both companies have their strengths and weaknesses, and I really don't see the point of squabbling over it. If you don't like Intel or you don't like AMD, just don't buy their chips. We don't all want the same thing from our machines, and hence we make different choices. Most of us are using what we're happiest with, and that is all that really matters, eh?

 

[Quailane]

I just think that AMD is doing much better than Intel with their dual core processors. AMD's 90nm chips put out very little heat compared to Intel's. The dual core athlon runs the same or less thermal load as single core pentium 4's (well, at least at stock). I don't think it is possible for Intel to compete at the same level, at least with most of the enthusiast crowd. Could you imagine the heat output of a dual core pentium when overclocked as high as the single cores are averaging? It would give a prommie a run for its money, I am serious. The dual core Athlons though, as we can plainly see and can't deny, CAN get as high as their single core counterparts. With the intels it is just NOT POSSIBLE with any reasonable means untill we see a dual cored pentium M roll onto the scene with all the full support that the pentium 4 has now.

 

[Sentential]

I think we've all grown tired of Sent's general pessamistic attitude in the AMD forums. Do you see us going to the Intel forum and Ranting "Intel Sucks" every chance we get?

See thats something I dont get. How am I being pessimestic? I am saying 3ghz as a whole isnt too impressive. That clearly says "AMD should be clocking higher because the design is better".

it's because of threads like this. everyone knows that 3ghz screenies aren't impressive, especially with a $1k cpu. threads like this are misleading.

Hits on my point as well.

2.8GHz on air is impressive because of the heat generated.

What heat? Thats the point people dont understand. A64 is such an impressive design that they do not produce alot of heat. What we are seeing are pure transister limits. Something we have *never* seen prior to this point. Before it was simply a limit of heat. In this case no. I hardly consider 50C* load a "heat" limit.

You want to see a heat limit? Come take a look at my ****ing Prescott. Friggin 67C* on load at the volts required for 4.3ghz with an XP120 and Delta on full blast. THAT is a heat limit. Heat = Poor design. Something the word "prescott" has been associated with for almost 2 years now.

I honestally dont understand how I am defending Intel at all with these posts.

 

[Silver]

Me very happy. I would/will be very extatic to get a 4400+ folding two units in FAH at 2.8ghz. Yes, very happy.

 

[Jcollins82]

So much for AMD being the cheaper way out.

One is just as expensive as the other. It's just being distributed differently.
Whereas with the AMD, you're spending your arm and leg on just the processor. And with the Intel, you're spending it on the CPU and motherboard. The difference in cost is completely negligible.

And as I said before, personally, I'm just impressed that we're seeing speeds comparable to single core offerings this early in production even if this chip turns out to be one in a thousand or more. I think anybody expecting 2.8GHz consistently with these chips is setting their standards a bit too high.

 

[TekeTorvo]

One could easily say that x2 is running at 5.6ghz.

2.8 times 2 = 5.6

 

[Jcollins82]

Actually, not really. Because the two cores cannot operate on the same thread at the same time.

 

[Sentential]

Actually, not really. Because the two cores cannot operate on the same thread at the same time.

Not yet atleast. There are talks by both AMD and Intel to do just that.

 

[Otter]

What heat?

Lost Circuits measures 89W (for the CPU alone) for 4400+ compared to 30W for a Venice, and that's at default speed and Vcore. Push the frequency and voltage to the max, and you could easily hit 150W when both cores are fully loaded.
http://www.lostcircuits.com/cpu/amd_x2/12.shtml

Thats the point people dont understand. A64 is such an impressive design that they do not produce alot of heat.

I agree. The point isn't that it's a toaster. And it isn't that these 2.8GHz air OC's indicate the overclocker has god-like powers either. The point is that even with two cores, the X2s are poking their noses into Venice and SD territory, and yes, we can hope for more.

 

[Quailane]

What heat? Thats the point people dont understand. A64 is such an impressive design that they do not produce alot of heat. What we are seeing are pure transister limits. Something we have *never* seen prior to this point. Before it was simply a limit of heat. In this case no. I hardly consider 50C* load a "heat" limit.

Ok, everyone who has ever used extreme cooling has met the transistor limits of their cpu, but that is besides the point.

People should just be happy that the dual cores clock the same as the single cores. If the single cores clock at 2.8Ghz, and the dual cores only clock at 2.2Ghz, that is something to be dissappointed about. You are acting like the dual cores suck, or at least are not worth talking positively about because they aren't clocking higher than the single cores go. But they are overclocking the same (or at least this limited sample so far) as the single cores, which is the best that can possibly be hoped for since they are the exact same design.

My point is just that there is absolutely no trade-off in performance if you get a dual core athlon, which at least I am happy about. How can you expect a dual core cpu to overclock higher than the exact same single cored counterpart? Also, we shouldn't argue because the jury is still out on this one. I assume that after a while, overclocking will improve slightly, but won't hit what the fx-55 does because it isn't using two of those cores. Now THOSE are cherry picked and can be from the zillions of chips that are made.

Well, whatever the deal is, this is showing a lot of progress. I remember reading at Tom's about the dual core athlons and how AMD just could clock them very high at all. They were able to run them at 400Mhz to start, but as time went on, they have been getting better and better (ie, closer and closer to single cores) untill what we have today which is ready for the market. Because as Sent said how heat is not the limiting factor for the athlons, they will get to the point not too far in the future where they will have overclocking parity with single cores. AMD has come a long way and is pretty close, at least by this sample. Even if it is a cherry-picked super-chip, it shows how far they have come, and we can expect them to get better.

 

[yo4444]

Not yet atleast. There are talks by both AMD and Intel to do just that.

Really? How far out is is that projection?

 

[TekeTorvo]

Actually, not really. Because the two cores cannot operate on the same thread at the same time.

That's what I have been saying about Intel's HT for awhile now. To Windows, it may look like 2 cores, but it can still only process one bit of information at a time, so what good is it? Now that isn't saying Intel sucks, so don't anybody think I said that. I said the HT doesn't do what they claim it to do. Which is process with 2 cores like it is doing twice the work.

 

[Jcollins82]

The easiest way to explain hyperthreading is that if someone was running two programs, one logical processor would handle program a, the other would handle program b.

Though if a program had multiple threads of operations, the cpu would divide those threads amongst the logial processors.

 

[Quailane]

The easiest way to explain hyperthreading is that if someone was running two programs, one logical processor would handle program a, the other would handle program b.

But that is VERY EXTREMELY SUPREMELY misleading. You make it sound like it can execute 2 programs at once.

The athlon can only have 1 thread active at one time and only process one thread at a time. The pentium 4 with HT can have 2 threads active at one time, but only process one at a time. It allows you to take advantage of the processing power better because when there is a cache miss, another thread that presumably has its instructions in the cache is waiting in the wings to take over immediately to use processing cycles that would otherwise be wasted while the processor is waiting for info from the other thread to process.

 

[Jcollins82]

I see, thanks for clarifying.

 

[Jakalwarrior]

Why do I hear people trying to make it sound like the intels dont clock well? If you want to push it youll need water but with the price difference...... I would still take the AMD though unless the dothan dual core comes out for desktop. When multithreaded games come out it could get really fun.

 

[Silver]

So much for AMD being the cheaper way out.

One is just as expensive as the other. It's just being distributed differently.
Whereas with the AMD, you're spending your arm and leg on just the processor. And with the Intel, you're spending it on the CPU and motherboard. The difference in cost is completely negligible.

On the nose. Really creates a windfall for AMD. At any rate if DFI releases the 939 agp in the next week or two then one can discount the vid card as well making the AMD solution very viable. Better performance and of course to all folders, less power.

On the AMD side, mobo and processor making use of all other components.

Intel side, processor, mobo, ram and vid.

At least this is the scenario I am looking at.

 

[Quailane]

Why do I hear people trying to make it sound like the intels dont clock well? If you want to push it youll need water but with the price difference...... I would still take the AMD though unless the dothan dual core comes out for desktop. When multithreaded games come out it could get really fun.

You need good cooling to push a single core. I can't imagine 2 cores hitting much over 4Ghz. That would be like a nuclear meltdown. It would be really fun to see how people cope with it and how inventive they get. Air is definately out. The stock intel cooler can barely handle a single 3.8Ghz chip. The thing I like about the AMD chip is it is on air. Plus, if you are poor like me, it won't make a big mark on your electric bill. I still might go with the pentium D though.

 

[Jakalwarrior]

Jump on over to the intel forum and read the guys post who got an 830. Hes using watercooling but its weak. One person isnt enough to judge all of their overclocking by but we can judge the heat output.