- Joined
- Mar 7, 2005
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.
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.