Please define the "mhz myth"

[Kitros]

I keep seeing it mentioned, along with a "PR" rating.

SInce I anm outta the loop, someone pleaase explain it to me.

Thanks.

 

[Hanpan]

THe mhz myth is in it's simpilist terms the idea that two processors that operate at the same frequency (mhz) are equally fast.

Please note by fast i mean in computer power not mhz speed.

 

[gogeeta13]

Lately, it has been attached to the PenitumIV

This is mainly because people see benched of a 1.4GHz athlon whipping the bejeezus out of a 1.8socket423 PIV.

Your average Joe Shmoe Intel buff who walks into Best Buy will see a 2GHz system and assume it is faster than the 1.4GHz next to it. This is what we techies call the MHz myth. And whether we like it or not, we are a minority of CPU sales, the average customer wants MHz not anythign else(cept price). This is why the p4 is doing so well. This is also why the cheapo sdram p4 will do good. Cheap price, seemingly amazing speed(MHz)

 

[joshg]

... which is why AMD has decided to name their new CPU's in model numbers, like the Athlon XP 1800 Model, actually 1.53 GHz clock speed, not 1.8GHz

A lot more goes into actual performance/power of the CPU rather than clock speed, such as instructions per clock, as with AMD, etc.

 

[SaturnX]

In relation to the PR rating, the Athlon XP 1800 is actually running at 1.53Ghz but is equivalent to that of a Pentium 4 1.8

--Mark

 

[dullard]

I think the best example is when you watch a very young child run (2 years old is a good example).

The child puts his/her feet down at a stunningly high frequency. The childs parents move their feet at an unusually slow frequency. Yet both make it to their destination at the same time.

The reason is due to the amount of work done with each step. A small child may move his/her feet 3 inches (7.6 cm) with each step. While the parent may move 3 feet (.91 m) with each step.

The key is that neither the frequency nor the work done with each step is sufficient to say which is faster (an adult walking is slower than an adult running even if the step sizes are the same).

The MHz myth states: regardless of the amount of work done per step, the processors of equal frequency are of equal total speed. Someone 'in the loop' realizes you must multiply frequency with work per step to get the true computer speed.

 

[sandorski]

*ahem*

No, the AthlonXP 1800 is *not* the equivalent of a P4 at 1800mhz. It is the equivalent of a Tbird a 1800mhz. IOW, an AthlonXP 1800 is the fastest x86 processor on the planet, that's available for sale that is.

 

[NelsonMuntz]

<< *ahem*

No, the AthlonXP 1800 is *not* the equivalent of a P4 at 1800mhz. It is the equivalent of a Tbird a 1800mhz. IOW, an AthlonXP 1800 is the fastest x86 processor on the planet, that's available for sale that is. >>


I don't think that's quite what AMD was going after when they decided to start using the "PR" thing. AMD doesn't consider the AthlonXP 1800 to be equal to a T-bird clocked at 1.8 GHz. There isn't an equivalent 266 MHz speed gain going from a T-bird to a chip based on the Palomino core. If that were the case, the Athlon MPs that came out not too long ago based on the Palomino core at 1.2 GHz would be easily beating the 1.4 GHz T-bird, but that isn't the case. The 1.2 MP beats the T-bird 1.2, but is closer in comparison to the 1.33 GHz T-bird than the 1.4 GHz T-bird. My understanding is that AMD really is trying to make a comparison to the Pentium 4 clock rates with the PR rating. For reference, I read about it for the first time here.

 

[Guilty]

Sandorski is correct,

http://www.vanshardware.com/articles/2001/october/011006_AMD_Model/011006_AMD_Model.htm
<FONT face=Arial>
"Instead of marketing their new XP processors according to clock speed, AMD will label their chips by a "Model Number" that is normalized to the clock speed performance of the older Athlon "Thunderbird" architecture (despite rumors to the contrary, the rating system is not based on the Pentium 4)."</FONT>
<FONT face=Arial>"Certainly some are going to interpret AMD's Model Number nomenclature as megahertz, but is that any worse that Intel's deliberately attempt to try to confuse the Pentium 4's megahertz with the Athlon's megahertz? Certainly some are going to confuse -- and already have -- that AMD's Model rating is based on the Pentium 4 instead of the Athlon Thunderbird, but since this error actually favors Intel it is not possible to argue deception on AMD's part."

1800 would be lowballing the 1.53GHz XP against the P4, we saw the 1.4GHz Tbird match or beat the 1.8GHz P4 quite easily. With the hardware prefetching, extended TLB entries, and SSE, the XP is easily 10-15% than a Thunderbird.</FONT>

 

[FuManStan]

Also, Apple has been battling the megahertz myth for a long time, since their G4 processors were only around 500mhz for the high end computers when AMD and Intel were racing to 1 GHZ.

 

[glenn1]

<< Also, Apple has been battling the megahertz myth for a long time, since their G4 processors were only around 500mhz for the high end computers when AMD and Intel were racing to 1 GHZ. >>



Yes, Apples are way faster than their rated clockspeed would seem to indicate they are, at least in cherry-picked applications that is. Here in the real world, however, most of us consider performance in tasks other than a handful of Altivec-optimized Photoshop filters. I know, i know... wanting to do more with my computer than a Gaussian Blur is a remarkable concept 😉

 

[Pabster]

Guilty wrote:

"1800 would be lowballing the 1.53GHz XP against the P4, we saw the 1.4GHz Tbird match or beat the 1.8GHz P4 quite easily. With the hardware prefetching, extended TLB entries, and SSE, the XP is easily 10-15% than a Thunderbird."

Agreed. The AthlonXP 1800+ (1.53GHz) could easily have a "PR Rating" of 2000+.

 

[RSMemphis]

The schweet thing though is that you can have PIVs and A_XPs run head to head - and therefore show that MHz are just myth in some applications and not in some others. You can't really compare Macs to x86 processors, the platforms are too different, as are the compilers, programs, etc.
Companies how specifially program for Macs can get a lot of power out of those, but something ported may not do the same. But with the x86, tests can show what is faster - yay!

The sad thing is that Intel is just milking customers right now. I am sure they could *EASILY* bring out a 2.2 Ghz PIV soon, but it is not necessary right now, because the 2GHz is still king of the hill. So they charge mucho dinero for the 2GHz, and as soon as the XP 1800+ is out, they will release the faster one, drop some prices and so on and so forth.

I can already see the ad at the back of a magazine: "2.2 GHz - my friends are jealous, my enemies are scared" PIV with 128 MB RD-Ram (benchmarks show that a PIV with less than 256 is castrated) and a GeForce MX or so. 🙂

Edit: Just a small typo (the != they)

 

[Odoacer]

MHZ Myth = Higher MHz is better. When you see the 1.4 GHz Athlon giving the 2.0 GHZ Pentium 4 a run for its money, you know MHz doesn't matter.