The math behind processors

[MielkeHBP]

So my friend won't come to grips, even tho benchmarks prove it, that my Athlon64 3200+ smokes his P4 3.2 W/HT in in everything and I was wondering if there was math behind it. I know im prolly wrong but this is what I kind of figured as a way our processors compare.

Athlon64 3200+
2200x64=140,800
P4 3.2
3200x32=102,400

Larger the number, the more data processed?

 

[Canai]

Soooo I take it then you are running a 64bit os?

edit: an no, that math is being pulled out of your ass I assume ?

 

[Megatomic]

If your friend is so deluded that he refuses to acknowledge the facts presented to him in reviews found on reputable hardware websites then leave him to his ignorance. He's obviously not worth the time and effort you are putting into this.

ps - your math is pretty funny

 

[Noubourne]

If he doesn't get it now, more than 3 years after AMD started spanking the motherfrack out of the P4 architecture on a clock-per-clock comparison, then he'll never get it.

He must think Conroe's are doubly crap, since they're "slower" than even Athlon 64s, which are "slower" than P4!! Why would anyone buy a Conroe? Right? lol!!!

This isn't really accurate, but it might help. Athlons have a wider and shorter pipeline than P4s, which is longer and skinnier. The data doesn't have to go as "fast" through the Athlon 64 pipeline, because more of it will fit in there at once, and it has less distance to travel - because the pipeline is shorter. That's (sort of) why an Athlon 64 gets stuff figured out before a P4 does - you can send more data through it at once. Hyperthreading was one approach to doing this - the Athlon 64 architecture was a MUCH more efficient approach. Conroe has (among other things), even a "wider" pipe than the Athlon 64, which helps to explain why it is faster than an A64, even though the data goes through it at a "slower" speed (Ghz).

If he doesn't get that, just kick him in the nuts, and then say ROSHAMBO. Tell him that if he'd been running an Athlon 64, he would have heard ROSHAMBO before you kicked him, but since he's on the "faster" P4 architecture, he didn't get to hear it till after. It doesn't really make any sense, but it will probably make sense to him.

 

[SexyK]

A 3.2C (Northwood) is not really "smoked" by a 3200+. In fact, even a 3.2E is faster in a lot of cases. Either way, love the math.

 

[Fullmetal Chocobo]

wow... Tell him the 0.9999(repeating) = 1 thing. That will keep him busy forever.

 

[Mr Fox]

Originally posted by: MielkeHBP
So my friend won't come to grips, even tho benchmarks prove it, that my Athlon64 3200+ smokes his P4 3.2 W/HT in in everything and I was wondering if there was math behind it. I know im prolly wrong but this is what I kind of figured as a way our processors compare.

Athlon64 3200+
2200x64=140,800
P4 3.2
3200x32=102,400

Larger the number, the more data processed?

Very Much an MIT Candidate....

 

[LittleNemoNES]

Originally posted by: Mr Fox

Originally posted by: MielkeHBP
So my friend won't come to grips, even tho benchmarks prove it, that my Athlon64 3200+ smokes his P4 3.2 W/HT in in everything and I was wondering if there was math behind it. I know im prolly wrong but this is what I kind of figured as a way our processors compare.

Athlon64 3200+
2200x64=140,800
P4 3.2
3200x32=102,400

Larger the number, the more data processed?

lol

Very Much an MIT Candidate....

 

[VooDooAddict]

I can hardly belive this is still a topic now with Core 2 Duo out ... but here's how I used to describe it to people... Yes it's not technicly accurate but it managed to get the point across to non-tech users. I preface to pepole that this isn't exact numbers but mearly trying to get the point across.

"The Ghz rating is how fast the chip can perform a single step in a larger task. The big differance between an Intel Pentium4 and an AMD Athlon64 is how many steps it takes to complete a given task. Lets take for example the simple task of multiplying two numbers. On a Pentuim 4 it takes 20 steps to multiply two numbers. On the Athlon64 it would take 9 steps to complete the same task. The Pentium completes each individual step faster then the Athlon. But the Athlon has less work to do before it achives the final result. This is why you can't compare the two types of chips on Ghz rating alone."

Most of the time from that they understood why something that with the smaller number could actually be "Faster." I also stressed that Ghz ratings could still be useful for comparing the relative speeds between products from the same line. ... Do you want to pay 35% more for a CPU that's only clocked 10% higher?


Coarse I should have just done the following as Noubourne suggested ... it would have saved me lots of time.

If he doesn't get that, just kick him in the nuts, and then say ROSHAMBO. Tell him that if he'd been running an Athlon 64, he would have heard ROSHAMBO before you kicked him, but since he's on the "faster" P4 architecture, he didn't get to hear it till after. It doesn't really make any sense, but it will probably make sense to him.

- Briliant