Razz,
No offense, but you should learn how to read before you post.
I mentioned the Pentium 4 was a bad design.
But point by point, since you aren't really too knowledgeable about processors, I'll try to help you out.
The K6-III was a good processor, as I mentioned, so I'm not sure why you're bringing it up. But, it was not as fast even as the Katmai processors. You're correct in one regard, but like most simpletons, miss the overall picture. The K6-III was faster than the Katmai, clock normalized, in integer performance running partial 16-bit code, i.e. Windows 95/98. For Windows NT, the Pentium II and Pentium III were faster clock normalized, and, most importantly, they could run at faster clock speeds. Even running Windows 98, the fastest Pentium II/III was always faster than the fastest K6-III, on integer. More than that, floating point performance was not even close, since the K6-III did not have a pipelined floating point unit.
The K7 was actually slower than the Pentium III Coppermine clock normalized, and for a while, they were available at close to the same clock speeds. This despite the fact it was a four year later design, was much bigger, and used much more power. This is not a strong design. Luckily, as I mentioned, Intel moved to the Pentium 4, which was a "marketing" design, focusing on perceived performance in the form of clock speed, rather than real performance. Nonetheless, Intel kept the P6 (aka Pentium III) design alive on the mobile devices, and this better than either the K7 or Pentium 4.
In fact, the K7 did not wipe the floor with the Pentium 4. It was very competitive, but the damning thing was the size of the Pentium 4 and power use by comparison. The K8 also did not beat the Pentium 4 hands down. It won most benchmarks, but not even close to all. It's a bad design, but better than the Pentium 4.
Here's the difference - Intel knows the Pentium 4 was a bad design, and went back to the P6, which was a good design. AMD keeps dragging the lousy K7 design around, with relatively minor modifications, and tweaks here and there, and somehow think it's going to work. You can put lipstick on a pig, but it's still a pig! That's why they can't match Intel processors, they have fundamental flaws that they have not addressed.
And if you're stupid enough to think the Phenom is nothing but a K7 with some enhancements, look at the internal design of it. They did improve scheduling, so it's now where the Pentium Pro was (1995), and there are a couple of extra steps to improve IPC, not clock speed, added to the K8, but internally, it's essentially the same. Look at where the AGUs are. Look at the micro-ops the executable engine runs. Look at the way too powerful x87 processor, which is nearly useless now. It's the same processor! If you don't know it, spend less time posting and more time reading.
I'll leave you with one more thought. Did you know the Phenom II is the same size as the Core i7? Look at the performance difference, it's shocking. How can there be any other verdict besides it's a flawed design? Did you know the Core 2 is much smaller than the Phenom II? Did you know it easily beats it in the vast majority of benchmarks (only when Intel processors are running at lower clock speeds, which most reviewers show because of price do they even look close)? What conclusion are you coming to?
Try not to be patronizing until you have some knowledge of the subject. The good news is, outside of fools like you, most people know the K7 isn't going to cut it. Certainly AMD has to know it. They can look at Intel's processors and see what makes them better, and why their's suck so bad. I'm confident the Bulldozer will be a good design, there's just no reason it shouldn't be. If it's another K7, I'll scream. So will their stockholders.
I still think they should resurrect the K6 and use it for mobile. The AMD policy of pigs and lipstick applies to their mobile chips too. They are using K8s for them now, which was not what it was designed for, and they can't fool anyone with it. A K6-III is a very power efficient chip. I don't know why they don't start with that for their mobile line instead of some K7 iteration that is poorly designed for it.
No offense, but you should learn how to read before you post.
I mentioned the Pentium 4 was a bad design.
But point by point, since you aren't really too knowledgeable about processors, I'll try to help you out.
The K6-III was a good processor, as I mentioned, so I'm not sure why you're bringing it up. But, it was not as fast even as the Katmai processors. You're correct in one regard, but like most simpletons, miss the overall picture. The K6-III was faster than the Katmai, clock normalized, in integer performance running partial 16-bit code, i.e. Windows 95/98. For Windows NT, the Pentium II and Pentium III were faster clock normalized, and, most importantly, they could run at faster clock speeds. Even running Windows 98, the fastest Pentium II/III was always faster than the fastest K6-III, on integer. More than that, floating point performance was not even close, since the K6-III did not have a pipelined floating point unit.
The K7 was actually slower than the Pentium III Coppermine clock normalized, and for a while, they were available at close to the same clock speeds. This despite the fact it was a four year later design, was much bigger, and used much more power. This is not a strong design. Luckily, as I mentioned, Intel moved to the Pentium 4, which was a "marketing" design, focusing on perceived performance in the form of clock speed, rather than real performance. Nonetheless, Intel kept the P6 (aka Pentium III) design alive on the mobile devices, and this better than either the K7 or Pentium 4.
In fact, the K7 did not wipe the floor with the Pentium 4. It was very competitive, but the damning thing was the size of the Pentium 4 and power use by comparison. The K8 also did not beat the Pentium 4 hands down. It won most benchmarks, but not even close to all. It's a bad design, but better than the Pentium 4.
Here's the difference - Intel knows the Pentium 4 was a bad design, and went back to the P6, which was a good design. AMD keeps dragging the lousy K7 design around, with relatively minor modifications, and tweaks here and there, and somehow think it's going to work. You can put lipstick on a pig, but it's still a pig! That's why they can't match Intel processors, they have fundamental flaws that they have not addressed.
And if you're stupid enough to think the Phenom is nothing but a K7 with some enhancements, look at the internal design of it. They did improve scheduling, so it's now where the Pentium Pro was (1995), and there are a couple of extra steps to improve IPC, not clock speed, added to the K8, but internally, it's essentially the same. Look at where the AGUs are. Look at the micro-ops the executable engine runs. Look at the way too powerful x87 processor, which is nearly useless now. It's the same processor! If you don't know it, spend less time posting and more time reading.
I'll leave you with one more thought. Did you know the Phenom II is the same size as the Core i7? Look at the performance difference, it's shocking. How can there be any other verdict besides it's a flawed design? Did you know the Core 2 is much smaller than the Phenom II? Did you know it easily beats it in the vast majority of benchmarks (only when Intel processors are running at lower clock speeds, which most reviewers show because of price do they even look close)? What conclusion are you coming to?
Try not to be patronizing until you have some knowledge of the subject. The good news is, outside of fools like you, most people know the K7 isn't going to cut it. Certainly AMD has to know it. They can look at Intel's processors and see what makes them better, and why their's suck so bad. I'm confident the Bulldozer will be a good design, there's just no reason it shouldn't be. If it's another K7, I'll scream. So will their stockholders.
I still think they should resurrect the K6 and use it for mobile. The AMD policy of pigs and lipstick applies to their mobile chips too. They are using K8s for them now, which was not what it was designed for, and they can't fool anyone with it. A K6-III is a very power efficient chip. I don't know why they don't start with that for their mobile line instead of some K7 iteration that is poorly designed for it.
While Intel makes some awesome products, they should also learn to compete fairly... It's downright dirty to pay manufacturers to postpone or cancel a direct competitor's launch... I use Intel stuff, but I feel it served them right to be fined like that... It's respectful competition that leads to better and better inventions and products being made, not a major monopoly..
Facebook
Twitter
Instagram