P4 vs. Athlon CoNtiNuEd

G

Guest

Guest
"Intel's New Pentium 4 Processor"
This has been driving me crazy for some time now. P4 vs. Athlon. Lets get some things straight! Athlon is a RISC processor, and I think all Pentiums are CISC. That should tell you why AMD has it better in the long run. Mac processors have been RISC, and therefore have been able to handle extreme mulimedia (decoding, rendering, etc.) very fast. Secondly, why the hell is the bus @ 100mhz and Rambus set @400mhz in one of your articles? So far, I have understood some things but this is freakin me out. How is it possible to have an bus @100mhz, yet the RAM is @ 400mhz and call it PC800? I think you are talking about the qaud fueled hyper pipeline, but this is just the same corner DDR memory took, since the new AMD boards are @ 266mhz DDR bus, or should I say, 133mhz(2x) (theoretically, but only benchmarks can tell right?), just so the processor can be left @ 133 FSB. Is there any reality that this freaky-ness will stop, and have real memory come out. When PC66 went up to PC100, there was no double quad super ultra mega data rate memory to deal with. This ideal makes me think that nothing has really changed and P4 is still @ 100mhz FSB pumping 4 times the info through its sorry pipe. And its probably not only that Athlon is a RISC processor, but that cramming all that you can in the pipeline really does not do anything. All that has really been done is increase the CPU speed from p3 1ghz to p4 1.5 ghz, and saying, oh my gosh its faster; like a half ghz increase would mean something in the benchmarks...nahhhhhhh. There is also the issue of latency, which puts DDR RAM at an advantage. But the bottom line is that PC2100 is actually @133mhz(2x) and that PC800 is actually @ 100mhz(4x). This really changes nothing and deserves very little merrit from me. Hey, its the clock that counts, and the lower the multiplier, the better speed you will get. Intel should of came out with Itanium and put AMD back in its place! But, hey, its not too late. As far as the channel goes, AMD will make the most profit, cus its a hell of a lot cheaper to put together.
 

FUGGER

Distinguished
Dec 31, 2007
2,490
0
19,780
RISC vs CISC (pentium), Intel adds instructions to the CPU so RISC is completely stupid point to base that lame decision on.
RISC is reduced instruction set, what part of reduced instructions don't you understand. AMD is adding instructions too, SSE2 to be more precise. imagine that!

Double pumped vs quad pumped. I first heard the term "double pumped" when AMD tried to fool people in thinking that the FSB was 200Mhz when in reality it was only 100Mhz. Intel used same lame ass terminology to coin "quad pumped". I agree they should have coind a new term to show off 3.2GB sec bus, and AMD CANNOT TOUCH that FSB bandwidth.

Maybe you dont understand the potential of the P4 yet, granted its a new CPU with new instructions that are not written into all software yet. or maybe you just cannot afford with the job you have or the type of work you do does not need the power of the P4, but in time most everything will be SSE2 optimized including AMD processors.

If your doing CFM, rendering, analasys, RIP'ing, CADD, CAM, CAE or anything bandwidth intensive the P4 is for you.

If your on a strict budget, or just cheap than stick with the AMD.



<P ID="edit"><FONT SIZE=-1><EM>Edited by FUGGER on 01/17/01 04:02 PM.</EM></FONT></P>
 

Kodiak

Distinguished
Dec 31, 2007
632
0
18,980
>> If your poor, cheap or stupid than, stick with the AMD

I've never called Intel owners/proponents stupid;
I would appreciate it if you'd extend the same courtesy to me:)

regardless, I was under the impression that any x86-based processor is by definition CISC.
Sparc, PowerPC etc are RISC, but neither Athlon nor PentiumX are...
 
G

Guest

Guest
I am not sure of this, it has been a very long time since I read the article. So if someone knows this stuff better than speak up.

I think MR. meal is referring to the K6 line from AMD. The core is Risc, outer layer is cisc for interpreting incoming data/instructions, and there is a intermediate layer that crunches Complex instructions into Reduced instructions. The Risc portion of the processor runs faster than the cisc portion(by how much who knows-maybe x2).

As far as the K7 processors go, I don't know that they used this method. If they did they have been able to stream line the process substantially. Maybe thats the reason AMD cpus are able to be clocked @ higher frequencies than their competition @ .18 micron. It could also be the reason that the cpu runs a bit hotter than an equivelent Intel processor.

Again (for those with short memories), it has been a long time since I have read the article that went in depth on processors, so read with caution. If someone can elaborate please do.