<< I never said Intel was ripping off people by selling the P4. They are (IMO) by using the mhz rating for the performance of their chips. Alot of people think that mhz = performance. About the amd rating, I'm sure their may be people that think the model rating is mhz. Amd doesn't claim it to be mhz. They even put a + in the model rating. Not a mhz. Intel sells their chips as a mhz rating. Which is not accurate of the performance. Of course this is all IMO, so if you don't like it. Don't cry about it. >>
See, the difference is, the P4 is ACTUALLY at that MHz. Intel just doesn't mention that MHz isn't everything. AMD's "rating" system is just pure symantecs. They're hoping when people see the "2000+", they think 2000MHz. And yes, this has happened. Guys on TechTV are already refering to it as 2000MHz. Say what you want about Athlons or P4's but the point of the matter is, they're trying to mislead the consumer. I find it quite amusing that one would deem Intel to be dishonest but credit AMD for doing practically the same thing. I'd call that biased. BTW. While I do think the P4 was marketed for the higher MHz, it was not designed so. The Engineers who designed the P4 recognized that a longer the P6 core will eventually reach its maximum scalability.
Think about it, if the P4 was truely just to pump up MHz, it would take a 2.0 GHz P4 to perform the same as a 1.0 GHz P3 or Athlon. But this is not the case now is it? Compare the P6 to P7 core, the P6's core is maxed off at 1 GHz for the .18 micron die while the P4 at the same die is (not maxed but it shifted to a new manufacturing process) at 2.0 GHz. That's a 200% of the clock speed. But what about IPC you say? Well, if the IPC was 50% of the original, then yes, it would've been a pure attempt to confuse the consumers because the drop in IPC and the increase in clockspeed would cancel eachother out. But that's not the case now is it? Although the increase from a 1.0GHz P3 to a 1.5 GHz P4 is not linear (not as if it were a 1.5 GHz P3) there is an increase and when you go to 2.0 GHz, there is a pretty significant increase. Let's be generous and say that the 1.0 GHz P3 is at the same average performance as the 1.5 GHz P4. That would mean that the P7 core had an IPC of 66% of the P6 core. But consider that the increase in scalable clockspeed is well beyond 200%, but for the sake of being generous, let's say it is 200%, that's still an overall theoretical performance of 132% of the P6 core as far as maximum scalable performance is.
So while there is a drop in IPC along with the increase in clockspeed, the drop in IPC is less in its effect on overall performance than the increase in clockspeed. I'm of course, being generous to the old P6 core as most of the figures I used were overstated. Considering how the P4 2.0 GHz can perform on par with a 1.5 GHz AthlonXP (which has significant performance per clock benefits over the P3), I'd say the overall gain when combining gain in clockspeed and loss of IPC is significantly higher than some people seem to be under the impression.
I find it interesting that some people seem to think that the fact that the Athlon is able to keep up while being 400Mhz behind the P4 is some kind of achievement. If you consider the design of the core, it should take a 3.2 GHz P4 to equal the performance of a 1.6 GHz Athlon, but this is not the case. Nobody seems to ever mention this. So kudos to the engineers who found so many ways to minimize the decrease in IPC of the P4 design.
. i love how apple always forgets that last little number when they do their benchmarks. aceshardware has included dual g4's in workstation reviews, and it is never pretty for the mac.
Facebook
Twitter