AMD suited for selling 4 core chips as 8 core chips.

If that is true then the chips are up for AMD. They should have abandoned t he PC market 6 to 10 years ago and focused on mobile. They also mess up th e transition to 64 bit computing by introducing a 64 bit PC chip with legac y instruction set support first, where Intel had wanted to remove all that junk from the instruction set. Actually the ARM 64 bit design is messed up too, with tons of unjustifiable SIMD instruction that will never be used ( by a c compiler or human), and just waste silicon. The ordinary 64 bit ins truction set is actually great and would have made a very small and effecti ve core on its own. AMD are due to pay back a ton of money in the next couple of years, VW-ing isn't a good idea.

So, what, exactly, is a core? How much hardware can be shared between two cores before they cease to be two cores? Who decides?

If consumers get stung by latching onto marketing words that have no widely agreed meaning, then they deserve what they get.

Sylvia.

Really? What should they use, clock speed? Benchmarks? No one has ever come up with a good way to measure performance. I think "cores" is as good as any mostly because they are easy to count... at least they used to.

Intel doesn't count hyper-threading as cores do they? I think the AMD thing is similar to that in that you are using less functional units than a full core.

You ask how to define a core? Let AMD define it by their other products. If they have four full cores in one chip and they count 8 cores in another, but the hardware is only four units with a core and a half each, which is really a core? Obviously one is wrong.

Or "core" was never well defined, was always just marketing speak, and no one should have assumed that it corresponded to performance.

Sylvia.

I have never considered "core" to be marketing speak. Is there anything that is *not* marketing speak in your world?

Well, I definitely prefer that manufacurers give clear and precise information. However, I think that AMD was right to say that thay have 8 core chip. Consider simple thougt experiments. First, disable FPU-s. IIUC then you get machine with 8 core that are as independent as it gets. Second, split FPU-s into half and assign each half to a sigle core. Again this is clearly 8 core machine, with possible not so fast FPU. Now, AMD claims that shared FPU is in practice faster than splitting it into two halfs and assigning each half to single core. Now, if this claim is true, then there is no deception. Anyway, shared FPU was clearly mentioned in all texts that I saw.

You ask how to judge speed: simple measure is maximal number of instructions that can be done in a single cycle. Now, this is something that could be improved: modern processors claim to support FMA and claim capability to perform two floating point operations per cycle. But a[5~t least from marketing speak it is not clear if one can do 2 full-width FMA-s per cycle (which for appropriate workload give 4 arithmetic operations per cycle) or if 2 instructons case is limited to one being multiply (maybe FMA) and the other addition.

Fixed that for you (subject).

I'm not. Not long ago AMD was a rather competent competitor to Intel. Currently they're losing tons of money already, and these lawsuits, if taken seriously by the courts it might be the end. Then Intel has no more serious competition in the PC area. I don't like the idea.

O.t.o.h. I think this lawsuit is frivolous and should be thrown out. For the same reasons as Waldek has mentioned in his reply.

joe

}snip{

I haven't looked at AMD recently, but some years ago they fell behind in the process technology by more than a year. At that point, I realized they could never compete again with Intel. In the old days, they were around six months behind in process technology which could be overcome by good architecture design. More than a year is too much with both speed advantages as well as cost advantages. So now AMD has to have a significantly higher ASP than Intel just to break even and their disadvantage in processing technology makes it harder to even reach the same ASP.

The only reason they are even around is because Intel lost an anti-competitive lawsuit and paid AMD about $2 billion if I recall. $2 billion doesn't last long in the semi business. It's all gone, as are many of the AMD fabs and now they have no lever to even catch up with Intel.

This lawsuit is many dB below the financial noise for AMD. BTW, the article says the suit claims the chips "cannot perform eight instructions simultaneously and independently as claimed". That sounds like it can be verified.

Years ago I advised friends to invest in AMD at times I could see they were going to pull out of an apparent nose dive and become profitable. Three times I was right. I no longer see this happening again short of winning more lawsuits.

There is another pending lawsuit against AMD on APU:

AMD Sued for Overestimating APU Success to Investors

I wonder how these 2 billions were "laundered"... :)

I doubt that they are frivolous, but I am not sure what is behind them and who would gain from AMD going bankrupt - and, no, it's not the money as such but the loss of 'confidence' that would bring AMD down. Cui bono? And, no, it's not likely to be Intel.

Regards, Nick Maclaren.

Now THAT one I considered to be a frivolous lawsuit. You pays your money, you takes your chances :-) Doesn't everyone who is looking for investors puff up their product to be better than what it really is? They are all like used-car salesmen; they'll tell you what you want to hear in order to take your money, and if you don't get a lemon, you've done all right. And if you do get a lemon, well, caveat emptor :-)

I, for one, and getting sick and tired of these companies lying and cheatin g, the banks among them. Wolksvagon was bad enough. Enron was bad enough. N ow my beloved AMD ? Are you kidding me ? I used AMD in new builds because o f the perception that you get more bang for the buck. Now it is just bullsh it ?

What is next, politicians lying ? (LOL)

There are no speed advantages. Note that the fastest 14nm Broadwell is quite a bit slower at 3.3/3.7GHz than it's 22nm Haswell brother (4.0/4.4GHz), and the next-generation 14nm Skylake is supposed to be available at 4.0/4.2GHz, but apparently Intel's 14nm process is barely up to the task, because even now, 3 months after the chip was officially released, it is available only in very small numbers.

Comparing the 32nm and the 22nm process, the 32nm Sandy Bridge was available at 3.5/3.9 GHZ (if we ignore the 130W and 150W variants), and its 22nm Ivy Bridge brother at the same 3.5/3.9GHz. The Haswell (refresh) is a little faster (see above), but has a revised design.

So no, there is no performance advantage (have not really been for about 10 years). There are probably cost advantages to using a smaller process, though.

- anton

If they use clock speed, they deserve to get burned, too. The bottom line is Caveat Emptor.

They have. There is no reason not to. It's a separate processor.

No, it's not obvious. The only obvious part of it is that number of "cores" means no more than the clock "speed".

I, for one, am getting pretty sick of people whining about companies trying to show their products in the best possible light. ...particularly when the consumer really ought to be doing his own research.

More whining.

" So, what, exactly, is a core? How much hardware can be shared between two cores before they cease to be two cores? Who decides?

If consumers get stung by latching onto marketing words that have no widely agreed meaning, then they deserve what they get.

Sylvia. "

Fully capable of executing 386 instruction set, or 486 instruction set or AMD64 instruction set.

Preferably without reduced performance characteristics.

Meaning market-conform l1 data chaches, l1 instruction caches, l2 caches.

First ones seem to be anywhere from 32 KB to 64 KB to 128 KB to 256 KB :)

Bye, Skybuck.

Fixed that for you (subject).

" I'm not. Not long ago AMD was a rather competent competitor to Intel. Currently they're losing tons of money already, and these lawsuits, if taken seriously by the courts it might be the end. Then Intel has no more serious competition in the PC area. I don't like the idea.

O.t.o.h. I think this lawsuit is frivolous and should be thrown out. For the same reasons as Waldek has mentioned in his reply.

joe " Then you are a big as a moron as he is.

Let's suppose intel sells you an 8 core chip, but it doesn't have any caches at all ?!

Then who will you bitch at ?! HAHA ! =D

GET REAL BRO !

Bye, Skybuck.

" That sounds like it can be verified. "

Very easy:

only 8 instructions will be executed:

thread 1. mov thread 2. and thread 3. or thread 4. xor thread 5. not thread 6. sub thread 7. add thread 8. mul

Since there are only 4 decoders.

This chip cannot execute these instructions simultaneously and independenlty as claimed.

Since the program is only 8 threads, each with 1 instruction the chip does not meet what was claimed.

That this chip may execute these 8 instructions later on because of "already decoded instructions" and/or "caching effects of instructions" is irrelevant.

Since these instructions will never be executed again... the chip can't do it... it's just 8 threads, 1 instruction. Done.

If some of these instructions happen to be paired, replace with more expensive, non-pairing instructions, perhaps a div, mod, or something nice like that.

Time to add the assembler newsgroup.

Your mission is to write a computer program to demonstrate AMD's 8 core chip false advertising ! ;) :) =D

I think I have already done a pretty good job at it.

Bye, Skybuck.