Centrino vs Athlon Mobile 3200

The Pentium-M does throttle back to 600 MHz

Yes, the Pentium M 400FSB throttles back to 600MHz but the Pentium M 533FSB throttles back to 800MHz just like the Mobile Athlon 64.

If you want the most bang for your buck then I recommend that you get the Athlon.

An Athlon Notebook will get your more raw power, a Centrino notebook will get you a lighter notebook (depending on which one) and better batterylife.

I can tell Snorre is damn happy with his lappy, owning demons in Doom III, but you may want a lighter notebook. It's one of those coin-fliping moments

For now you're right, but when the Turion 64 notebooks eventually come available they will have the same battery life as the new Centrino "Sonoma" notebooks. Then it will come down to features and price/performance like it should be, and I have no doubt what's going to be the only logical choice then.

The weight difference is mainly due to the battery size, a CPU that drains more power requires a bigger battery. The battery life varies from model to model, so I suggest that you see what the vendor specifies. Mobile Athlon 64 models usually offer between 2-3 hours battery life, while the battery life of Centrino models varies from 2-4 hours. I suggest that you ask yourself this question: How often and how long do I really need to use the battery? If you will only use it on battery a few times each week and mostly on AC power then forget about the battery life, as it's really not important to you.

How much longer is it going to be until Turion is available?

I think a better question is "How much longer is it going to be until Turion notebooks are available?" Intel is doing their best to continue the delay of Turion notebooks by...presuading...various manufacters not to sell Turion notebooks for a while. But that being said, I think there will be several out by this summer. I'm sure Snorre will pop in and give you some more concrete numbers.

Depend on what type of M Athlon 64 you are using. I have the 62W version in my Gateway 7405, and it gets about 3:10 battery life on a full charge, and the 35W version should get somewhat more, probably 3.5 - 4 hours.

The PM should get anywhere from 4-5 hours on regular battery.

Performancewise:
Athlon64 >> PM,
even on the same clock speeds, most bench marks show that A64 performs around 115% that of a same clock rate Pentium M.
And the A64 is usually clocked higher on top of that. The 3400+ at 2.2 GHz, and the 3700 at 2.4 GHz. And no PM is clocked higher than 2.1GHz. So performancewise, it's really no contest.

How interchangeable are CPU?

I thought CPU's were one of the components that could be upgraded. They are just a chip that attaches to the motherboard, no?

It's only interchangeable when the interface (socket, or slots on really old CPUs) and the chipsets support the particular CPU, and voltage has to match as well, be it 0.9, 1.1, 1.5, or 3.3.

So then bascially, they are only interchangeable with CPU's of the same type, brand, and of the same generation. Ie. only P-M to maybe a higher clocked P-M, never P-M to Athlon64 or P-M Sonoma.

That's basically correct, with some exceptions, such as S754 Sempron can be replaced by an S754 Athlon64.

Basically. There are exceptions, such as the Socket 754 one. Also, for quite a while, Intel and AMD used Socket A and Socket 8, and you could pick whichever company's CPU you wanted.

I do not think this is correct . A64 outperforms P4 at same clock speeds but not the P-M. In fact P-M at 2.1 Ghz performs (in most apps) like a P4 at approx 1.5 times the clock speed (3.4Ghz).

There are limitations inherent in a laptop CPU like P-M. But it is definitely comparable in performance to the higher clocked A64 or P4.
Check out
http://www.anandtech.com/linux/showdoc.aspx?i=2308&p=2
for some interesting benchmark data.

I don't think you are correct, check out these benchmarks:
http://www.hothardware.com/printarti...?articleid=620

in most of these tests, the A64 4000 clocked at 2.4 GHz as normal goes up aginst overclocked 2.5 GHz Dothan (which will be beyond impossible in a laptop), and wins most of the benchmarks. And the currently second highest clocked PM on the market at 2.0 GHz, is not even in radar range, on most of these tests.

PM simply does not match up against A64 in terms of real life performance, clock by clock, for example:

I own the Gateway 7405, with A64 3200+ clocked at 2.0 GHz, with MR 9600 Graphics 64MB VRAM, and 512MB RAM; my friend has the HP nc6000, with PM 755, clocked at 2.0 GHz, also with MR 9600 64MB VRAM, and 512MB system memory. He had a faster HDD, with a 80GB 5400rpm, I had only an 80GB 4200rpm.

One of the newer games, Immortal Cities: Children of the Nile, is a new genre of fully 3D city building/simulation type of game, and is very CPU intensive, involving AI for upto a couple thousand characters on screen at the same time, also pretty demanding on video. I loaded the game no problem, and played it at medium resolution due to graphics memory, and had no problems, and got above 30fps even during the most intensive part of the game.

So I lent this game to my friend, with the same clock PM, on the other hand, had trouble keeping up with the 3D engine and the characters at the same time. After the population (each character has its own AI in this game) grew to a few hundred, slowed down tremendously, had only about 15fps at that time. Not only the graphics was bad, the game play, mouse, keyboard became so unresponsive that he had to stop the game at that point, coming to a conclusion that the game was not playable on his computer, at its most intensive portions, when the greatest amount of characters are being rendered and AI calculated by the CPU.

Let me review, we had the exact same graphics, system memory, and he had a slight advantage on HDD. Why was this game playable (barely on my computer) but not on his? Because the 2.0 GHz A64 3200 performed much better in floating point calcuations than his 2.0 GHz Dothan 755. A64 is simply a much superior gaming platform, even clock for clock, anyone claiming otherwise should test it out for themselves.

If anything these benchmarks more or less prove than despite some limitations P-M hangs right there with the A64 and higher speed P4s.

Please recall the earlier benchmark (anandtech) showed the superiority of the new 533Mhz FSB Dothans. And P-M's shortcomings in the benchmarks (which used the old 855 chipset) referred by you, mainly result from memory bandwidth. Even the reviewer seemed to suggest that with the statement

"In the first part of '05 it is rumored that Intel will be introducing a 533MHz FSB Pentium M processor, which should really help further extend Dothan's overall performance profile. Now, what we would really like to see is a Pentium M motherboard from DFI or others that is built on Intel's upcoming "Alviso" chipset which is essentially an i915P mobile solution. A 533MHz system bus Pentium M, coupled with 533MHz DDR2 DRAM and a PCI Express graphics slot, would smoke up the benchmarks in a big way for sure. "

A64 is undoubtedly a faster processor. I am not arguing that. But it is nowhere neer 115% faster. In fact in many apps P-M does equally well and P-M does have a lesser clock speed.

Of course in some apps which will involve hyperthreading etc ... P-M will suffer. But we have to keep in mind it is a laptop processor first and foremost. For a vast majority of business applications it is right there performancewise with the big guys.


I cannot explain this of course. Maybe guys who own the new Sonoma laptops can chime in. It can be laptop specific of course. I would think that maybe the HP's GPU was probably underclocked. I will agree however that for gaming a64 is superior. But again the P-M is not just for gaming. If someone wants the lappy primarily for intense gaming .. then definitely a64 or a high end P4 will be a better choice. But for most real world applications (esp business applicatuions), P-M does equally well - at the same time giving other advantages like battery life, mobility etc.