I'm listening - but at some point, I will just start to shrug it off if you aren't willing to accept objective facts and if you aren't willing to admit you jumped into this discussion without the proper knowledge

No matter how many times I repeat it you will never listen

And yes, it is kind of funny, but the Pentium Pro was a wonder board!

No matter how often you repeat it, no matter how hard you're trying to ignore my points, no matter how you hard to try to 'misunderstand' what I said, it doesn't change facts.

By the way: the argument 'they lost market share, you can see that on myself as a consumer' is really hilarious

Actually, I didn't, I miss a big chunk out about the first Pentium computers.

As I said, there was the Pentium class computer, which was bested by the AMD AM5x86, and Intel put out the Pentium Pro, and yes, they did lose market share, I too bought one because it allowed me to put 4 CPU's on the same board, it really made my day!

I've attached the data sheet of an AMD-K5. Just in case you're interested in working with source material instead of only subjective 'text books'.

Every AMD CPU from the 5x86 (the K5 is the next model) has been developed by AMD themselves. Their 486 were based on licensed Intel architecture, but all the 'generations' after were/are not. So much for your initial statement about "they use the same processor basically".

hehehehehe! The chip you're talking about was acually slower than its Intel aequivalents. That is also said indirectly in your oh-so-great 'text book' if you read carefully: they compare the AMD chip with 133 MHz to an Intel chip with 75 MHz and say AMD is faster. Notice something? They're comparing it just for people who wanted to keep using their 486 mainboards!

In the paragraph about the Pentium Pro (which wasn't the 'main model' at that time anyway), you missed the important part, too. This chip had a lot more cache than any other CPU because the L2 cache was included directly on the chip - in contrast to being on the mainboard (as it is still today). 1 MB more cache on Pentium Pros than on AMDs. So with this quote, you proved exactly the opposite of what you wanted to say.


In the time you are referring to, AMD actually lost market share, seen worldwide. As mentioned before, they managed to get quite a nice chunk in the days on the 386 and 486. Their 486DX4 was also remarkable. But Intel's move to call their next CPU 'Pentium' instead of 586 worked, because it confused the customers! AMD was not allowed to use the same name, thus their names like 'K5'.

It wasn't until late the late 90's (1999?) that AMD even sold computer chips here in New Zealand. The single thing that allowed them some market share was the cache size allowing for more performance for something that seemed by by Intel standards to be slower.

I had another look in my text book, and it states that even though AMD as been around for over 30 years now (Circa. 1969), it was this (below) that made the big difference in gaining market share over Intel.

Again, wrong. What put them (and Cyrix back then) into the spotlight was ironically enough the clock speed of their CPUs! Intel's 80386 only went up to 33MHz, AMD and Cyrix managed to push theirs to 40 MHz. That is what made them 'famous' in the first place - not cache size.

Mr Creosote: I never said that was the only way to speed up the CPU, but consistantly since the beginning this has been the main way AMD has beaten Intel. It was this method that put them into the spot light and made them so much money in the very beginning.

Well the kernel of an OS is very close to the CPU, making it a good candidate for slowing down on a faster processor.

Well... no. It sounds right enough, but there might always be programs which have been written so closely to the P3's architecture that they run extremely well on them, but not nearly as fast on any other CPU. Such basic programming has been 'out-of-style' for decades now, yes But I'm talking about theory.

I'm aware that there are many factors that influence the actual processor speed, and it can even be true that AMD processors running at slower speeds consistently beat their Intel counterparts. What I said was, if you have a P3 and P4 running at the same clock speed in identical systems, the P4 should always outperform the P3, precisely because of those improvements in the architecture. If not, then Intel is definetly doing something wrong and deserves to be sued.

However, I agree with you. This lawsuit is going nowhere. It might, however, generate some bad press for Intel, and I bet AMD won't mind that a bit.

Wrong. You obviously don't know much about the Von-Neumann architecture of CPUs (and I don't blame you for it - it's an incredibly boring subject), but I can tell you cache is just one of many factors. The whole pipelining technique, the splitting of the ALU, the addressing of MMU - it's all done differently in AMD processors than in Intel's.

It's completely obvious Intel will win this case in my opinion, because there is simply no objective way to determine the speed of a CPU. If it was like Delos said, it would be easy: you could calculate the 'speed index' with a simple combination of cycle speed and cache. That is not the case though (as hinted at - I don't really want to go into detail about that).
The benchmarked speed always depends on the other components in the computer, and since some parts are working together better with this CPU and others with the other one, there can't be an objective system for testing. Next quirk: what software to use? Some programs are optimized for the way an Intel CPU works, some like the AMD style better.

So what the suing people propose (an objective speed indicator) is not possible. And that demasks their obvious attempt to earn some easy money...

Dizzy: The speed of the processor is one thing, but the size of the cache of the processor (memory in the processor) is what turns a Snail into a Cheetah.

This is what AMD does to beat Intel everytime, they use the same processor basically, and then they put in a bigger cache so that more data is actually in the CPU to be worked on.

This works on the same principle of Memory (RAM) vs. Hard Drives, where Memory is approximately 100 times faster at retreiving data then from a Hard Drive.

What the people are sueing for is the fact that Intel tells people that it is the processor clock speed that people should look at when buying a chip.
AMD on the other hand have never claimed it is clock speed, and for marketing reasons they don't tell people what the clock speed is because it doesn't give a fair representation of it's abilities.

We can look at the P4 1.7GHz, this was beaten hands down by a AMD Thunderbird 1.4GHz, which was only beaten by the P4 2.1GHz.

But yet the average person saw the AMD '1.4GHz' and then saw the Intel '1.7Ghz', and they bought the 1.7 because they were told be Intel they were "faster processors" which is not true at all except for clock speed at which they run at.

*takes a deep breath*

Personally, I don't think Intel will be beaten, when they said "faster" they were of course refering to clock speed, and it was the average user that got the wrong impression because it was a misleading word ( "faster" ) in this case.

Edited by Delos at 03:13 on September, 13th 2002

It's stuff like this that makes me wish I lived in the States. I'd be a pretty good candidate for this one, too, since I was quite convinced a P4 would always be faster than a P3 running at the same clock speed. Come to think of it, I still am. Either these people have their facts wrong, or Intel has a very weird notion of what "techonological development" means.

I just heard today that the next generation of Intel processors would have special support for Microsoft's Palladium... Guess they'll be going out of business soon then

Intel has been sued in the USA (where else?) because their P4 CPUs are too slow. The suing customers claim Intel used unfair methods in their advertisement, using the MHz/GHz speed as a quality mark whereas P3 and AMD processors are in some cases faster with less MHz.

In the case of success, each suing customer can be 'awarded' up to $75000 (the case takes place in California). With approximately 50 million P4s being sold up until now, it would add up to quite a nice sum for Intel to pay on the whole