Does RAM have anything to do with the proccessor or MB?

[the Black Monarch]

3.0Ghz has an 800Mhz FSB, the 3.06Ghz has a 400Mhz FSB.

Actually, the 3.06 has a 533 Mhz FSB. Not 400

[dwchang]

3.0Ghz has an 800Mhz FSB, the 3.06Ghz has a 400Mhz FSB.

They do? Uhm...the 800/400 FSB is the same isn't it. They quadruple their FSB which is really 200 and say 800 due to the "double-data-rate" we do the same thing except we just double ours and hence the 333 and 400. The true FSB for either of us are either 133 (267 for us, 533 for them), 167 (333 and 668) and 200 (400 for us, 800 for them). As I stated, they quadruple the true FSB and we double ours...it's an interesting strategy from both of us...don't be at all mistaken...the FSB IS 200 Mhz for both the 400 and 800. That's why you have a 3 Ghz part...200 FSB x 15 Multiplier. If it were 800...that'd be insane. I don't know of any processor that has that as the real FSB.

As for if it's 400 vs. 800. I dont' believe that. I'd believe if it's 800 FSB vs. 668 (167 x 4). This makes sense since 3,006/167 = 18 multiplier. Obviously 200 does not divide evenly in it and thus, you are right on the FSB...well right in that it's not 200...wrong in that it's not 400. That doesn't go into 3,006 evenly.

[dwchang]

Whaa??? The website that I went to for such information gave 13.5 as the voltage for the mobiles and 15 as the voltage for the desktops. Are you going by Intel voltages or AMD voltages? There may be a reason why you don't see many laptops with AMDs in them

Oh shit, I made a HUGE mistake in my reply..I meant POWER not Voltage. The core voltage of most chips is around 1.65 V, so in essence my reply is still the same in that 15 is wrong. However, I'll answer the power issue since that is what I meant to discuss..voltage has always been around there for the last 2 - 3 years and hasn't been a real issue. Power has.

Actually the hottest (Power-wise) are Intel...not AMD. Ours average 65 W core, while the new Northwoods can go in the 80's (maybe it was high 70's). These are typical desktop powers, I can assure you. I work on this stuff and specifically did some of this testing in the past.

As for laptops, actually we have a 16 W uPGA part that is being used in our thin-and-light notebooks carried about Fujitsu. On average though, normal notebooks (both intel and ours) are between 25 W - 45 W. These are the normal sized ones and not thin-and-light. Stuff like Centrinos and our thin-and-lights are easily below 20 W which is why they operate longer..consume less power.

It's known that Intel is taking advantage of public stupidity. However, I've yet to see anything suggesting that Intel is in any way responsible for said stupidity. (unless one of those Blue Man Group commercials said something like "Gigahertz is everything. Do not pay attention to instructions per clock cycle, that's a bunch of crap" and I missed it...)

Almost all the commercials on TV ALWAYS specify fairly loudly "comes with the new Intel Pentium IV processor at 3 Ghz." They almost always stated the frequency in the Dell, HP/Compaq and Gateway commercials. I can easily see how one can miss it since one can just get used to hearing it.

Yeah, I thought it had something to do with being 64-bit, but I couldn't remember well enough to say it with any kind of certainty.

Was the Itanium II any better?

Itanium II was better, but they're still pretty slow and no one is buying since they are expensive as hell.

Whoa, I didn't know you'd upped your L2 cache. I need to visit those websites again.

A whole megabyte of L2 cache... oooh... drool...

Yeah our Barton cores (model 2800+ - 3200+) have increased L2's and some have higher FSBs.

No, it's fair enough after you slammed Intel so many times


Personally, it seems to me like you all too greatly enjoy saying things like "Yeah, our chips and their chips perform about the same, but our chips are better because they're more efficient" and "Yeah, the P4 goes way faster than the P3 ever could, but the P3 was more efficient so it's better." Something just doesn't add up there ... I mean, I like efficiency and all (my first car is going to be a Honda Insight), but come on... when you see a P4 overclocked to 4.44 Gigs and the P6 core (Pro/2/3/Celeron/Centrino/whatever) hasn't even been pushed past 2.0... maybe that horrible inefficiency isn't quite as much of a drawback as you thought

Well as I said in the earlier argument Performance = Speed * instructions/clock. Obviously speed DOES play a roll and thus if they operate say 800 Mhz faster than us, even if we have the more efficient core...they even out at around the same performance. There are two variables to the equation (although most consumers think only speed matters). Basically if you do things fast...how many things are you doing in that time? We do more things per clock, but our speed is slower. They do less, but do things more often. Make sense? This is how a 2.25 Ghz Athlon can have COMPRABLE performance to that of a 3.0 Ghz P4.

As I stated earlier, the INITIAL reasoning for this design change was to get the speed since consumers want it and the P3 architecture (or any architecture that efficient) can't scale that high up. People would THINK the chips were slower, but in fact they are very efficient and thus performance. That's why they added pipeline stages and made it easy to get Frequency (they average 200 Mhz a quarter...amazing). At the same time, the design is inefficient. Again...INITIAL reasoning was this..obviously it's paying off since A) consumers like speed...continue to buy and B) speed is one part of the equation. I am just stating I dislike inefficient designs since I am an engineer and well...I couldn't see myself on PURPOSE designing something inefficient. However, I won't deny that it was an excellent move in terms of marketing and consumer-outlook.

If things go REALLY well in the next few weeks, I might be getting a quad Opteron desktop/server.

I hope you got a lot of money....

[alternatefutures]

Actually, the 3.06 has a 533 Mhz FSB. Not 400

Dah! You're right! 533 Mhz wasn't around for all that long and I just kinda forgot about it.

[dwchang]

Actually, the 3.06 has a 533 Mhz FSB. Not 400

Dah! You're right! 533 Mhz wasn't around for all that long and I just kinda forgot about it.

Yeah so in this case, it's a 133 FSB multiplied by 4 to make 533. Funny how the 167 and 133 both went into 3,006...sorry about that...I just assumed to much from them (having a higher FSB)

[alternatefutures]

Hey, maybe you can answer this question. How does 133X4 = 533? Are all the comp sci engineers bad at basic math or something?

[dwchang]

Hey, maybe you can answer this question. How does 133X4 = 533? Are all the comp sci engineers bad at basic math or something?

Well considering it's truly 133.33333333....and so on.

That multiplied by 4 = 533.3333333.....

Look at specs...it doesn't ever say 532 Mhz FSB. It says 533 Mhz.

As for the math, although it's not wrong...Computer Engineers think in 1's and 0's...there is no inbetween or anything past that .

[the Black Monarch]

Almost all the commercials on TV ALWAYS specify fairly loudly "comes with the new Intel Pentium IV processor at 3 Ghz." They almost always stated the frequency in the Dell, HP/Compaq and Gateway commercials. I can easily see how one can miss it since one can just get used to hearing it.

I have personally seen a grand total of ONE (1) PC company commercial in which the clock speed was brought up. Fry's brings it up more often, but they don't make the PCs. Aside from that, all the IBM/Gateway/Dell/Whatever commercials that I've seen just said the model name and not the clock speed.

I suggest replacing hertz with instructions/second as the standard by which CPUs are judged.

I am just stating I dislike inefficient designs since I am an engineer and well...I couldn't see myself on PURPOSE designing something inefficient. However, I won't deny that it was an excellent move in terms of marketing and consumer-outlook.

It wasn't just marketing. The higher clock speeds get the job done just as well as your much-lauded efficiency (maybe slightly less, as 2.25 Ghz x 9 inst/cycle > 3.06 Ghz x 6 inst/cycle, but you get my point). If designing something to be inefficient on purpose indirectly allows for superior performance, then why not do it? In the world of computing, you want miles per hour, not miles per gallon.

I hope you got a lot of money....

I have my eye on a $250,000 paycheck in the next few weeks. It might not happen, though. I'll keep you posted.


Miscellaneous notes:

What's the functional difference between the K6 and the original (non-XP) Athlon?

If the P6 core only executes 3 instructions per second, and the P7 core (Pentium 4) executes 6, and the Centrino is based on the P6 core, then how could a 1.6 gig Centrino perform like a 2.4 gig Pentium 4? The math just doesn't add up. George now says that the Centrino is based not on the P6 core, but on the upcoming Pentium 5! I think you should ask your AMD buddies if this is the case and they didn't correctly hear which core the Centrino team was modifying.

[klinky]

I think one could point out that there is a problem when your *NEW* 1.8Ghz CPU runs like your *OLD* 1Ghz CPU.

But one could also say that there is a problem if your old CPU could have never gotten above 1Ghz. So say your new cpu can go all the way up to 4Ghz, but it performs about as good as your old processor did @ 2.5Ghz. Well still that's a improvement because the old one would have never been able to reach 2.5Ghz.

Intel is currently sticking with it, their parts in the high-end while semingly brute-force have worked and the Pentium4 continues to scale well. The P4 3.0C 800Mhz cpu, depending on which benchmarks you look at meets or beats the Athlon XP 3200 400Mhz cpu. Of course the Athlon is running @ 2.2Ghz. Still one could say the Athlon is flawed for not being able to go higher. If it could run @ 3Ghz like the Pentium4 could, then it would be the undoubted performance leader. But currently it can't .

Also it should be noted that the Athlon XP 3200 is $445 compared to the $415 Pentium4 3.0C. Though if you look at the sub-$100 market. I'd much rather have a Athlon 2600(2.133Ghz) over a Pentium4 1.4Ghz.


It's a matter of opinion.


~klinky

[the Black Monarch]

Still one could say the Athlon is flawed for not being able to go higher.

Or for usually having less cache.

Also it should be noted that the Athlon XP 3200 is $445 compared to the $415 Pentium4 3.0C.