Laptop PC processor speeds Vs Mac's

I

Incubated Funk

Registered
Hi Guys

I have a general query regarding the processing power of a PC Laptop Vs a Macs.

I wondered why Mac powerbooks processors only go upto 1.67Ghz whereas a PC processor goes up above 3Ghz. My friend said that you cant compare the numbers directly for various reasons, but he didnt know anything about the mac processors to say why.

Does anyone know how they compare and if the powerbook processor can compete at only 1.67Ghz (if not why didnt they stick a 2.5Ghz in there?)???

Thanks in advance
 
Only the G5 can go beyond 2GHz for now, and the G5 is way too big/hot/power-hungry to fit in a laptop. Maybe next year. The G4 just hasn't reached beyond 1.67.

That said, the G4 is faster MHz-for-MHz than PC processors. On top of that, the processors in PC laptops are a lot slower than desktop processors, because they're scaled down a lot to meet the size/heat/power requirements of laptops. Intel's processors have always been un-laptop-friendly monsters. I think AMDs are better, but I'm not really sure. On the other hand, the G4 chips used in laptops are the same ones used in desktops, because the G4 is lean enough to begin with.

How exactly does a G4 stack up? Well, that's a loaded question open to much debate. There are many factors in performance. I think the general consensus is that a 1.67GHz G4 would be equivalent to a 2~2.5GHz Pentium 4 (depending on the task), but I'm not sure how THAT translates to a mobile Pentium 4 (or whatever they're called).
 
amd processors, although they may say 2700+, are actually much closer apple's processor speeds - they've learned the "bang-for-buck" theory now, whereas intel are just adding speed. intel, btw, are currently lagging behind AMD in nearly every arena in terms of raw power, so the theory's being proven. (simple terms, i realise).

i think, just have confidence that the comparable price ranges between mac and pc are very similar in terms of actual real-world speed. mac's used to be faster, but i think it's about level now, as all the chip companies are hitting barriers in the speed stakes, and are finding it harder to improve clock speeds
 
The architecture of the chip plays also an important role..
An important thing to keep in mind is that Apple has chosen to use a dual core set up for its top-line machines. This is important since there's a theoretical limit in the resolution of the chip fabrication, let say 10 nm. The problem is that it's really hard to define a technique to produce at an industrial level circuits with that resolution. It could be done at research level with x-ray lithography, or nanoimprint lithography (you should be a guru of NIL to reach 10 nm, BTW), but it's really hard to scale these techinques to industrial processes. A practical limit is nowadays 50 nm, and in fact I've hearded of just one chip that reached this resolution. And in fact, the number of dual core machines is now growing...
 
An important thing to keep in mind is that Apple has chosen to use a dual core set up for its top-line machines.
Click to expand...

Show me a CURRENT shipping mac with dual cores.

Current PowerMacs use dual CPUs.
 
You'd be surprised, some PC laptop makers do in fact stick a desktop CPU into their laptops. Yep, those are the ones that don't see the light of day. By that I mean they're forever tethered to a power adapter plugged into the wall, can't go portable with them for more than an hour or 2.

3 GHz vs 1.67 GHz, lets see. MHz is bollocks if the CPU and software don't make good use of them. I read white papers on Intel and AMD docs for learning Assembly, what I discovered is that

1) Every CPU always misses the first branch instruction. It's just nature of the beast. But it'll get it the second time around.

2) The penalty for missing a branch is somewhere in the neighborhood of 14 cycles.

To clarify, when I say a branch instruction I'm referring to something like:

if (some condition)
then do this
else
do that

So with 3 GHz, or 3 billion cycles per second, that number starts to dwindle down if the code you are executing involves a very big loop and/or many "IF" statements. So the more the code jumps around the more painful it will be for the CPU to process.

The penalty of a branch miss for Intel's P4 CPUs is highest of them all. And yes, that's partially due to the longer pipelines but there's more to it than that. But anyways, maybe now you can see how even though Intel chips have high MHz clock speeds, the majority of those cycles will get wasted, kinda like getting a speeding ticket.
 
I know full well the "weight" of a desktop CPU in a laptop. I have a HP Compaq nx9110. It was listed on their website as "desktop replacement". they aren't kidding.

It has a normal "desktop" P4 at 3ghz. it weighs a ton and has no battery life.

note: i have no choice over what we buy, but it's always gonna be Windows..

i'm chaning to a nc8230 soon (half the weight, and thickness :) )
 
You must log in or register to reply here.
Back
Top