Last message on previous page:

Come on...don't attribute seasonal shifts from the best quarter of the year (Q4) to Q1 as being due to AMD. Intel should definitely have negative growth in revenues QoQ - if they didn't, that would be incredible. Imagine being a company that has Christmas like sales twice during a year. 8O

Edit: Also, how do you "burn through cash" when after all expenses, you're still making money?

Actually I find that assertion highly dubious.

Take a look at any of the major price search engines, look up desktop, and sort by popularity.

What you are going to find is that the most popular desktops are dominated by Pentium 4's, Celerons, and Semprons / low end Athlon X2s.

We are long past the time when a single blockbuster app moved the market. The last time that really happened was Quake 1 / Quake 2. Quake 1 pushed the move to pentiums, and Quake 2 pushed the move to GPUs. That was a long long time ago.

The most APPARENT thing looking at these benchmarks of various processors over the past 5 years is not how much faster C2D is - it's how SLOW the progress is in general on the CPU side.

There was a recent article extolling the power of the new CPUs vs the old - comparing them to the likes of a P4 2.4Ghz from 2001. That was 6 years ago, and that chip held its own in *many* benchmarks. The inevitable conclusion is that for most tasks, it is still plenty of CPU.

Now roll back to sometime in the 1990s, pick any year , and see if anything around 6 years prior was even remotely comparable to what was available in the year you pick. You'll find that the older chips were maybe 1/4 the speed of the newer chips, at best. ie, a 400+% increase in power in that time period - and in many cases you'll find 10x increases.


Today, we have a bunch of yahoos hooting and hollering over a measly 50% increase in a similar time frame. If this kind of 'advancement' continues, AMD and Intel both have much bigger problems to worry about than each other.

Oh cmon now...do you really have to ask? You know the answer

Its simple, you burn through cash when its a negative adjective as applied to the company you simply dont like.

Baronomics 101. The company in the red is healthier than the company in the black when those companies are AMD and Intel respectively. QED as one of my profs used to say.

That is an excellent observation. Writing multithreaded code can be very hard and also not all jobs are able to be multithreaded very much, very well, or sometimes even at all. My guess is that with clock speeds pretty stagnant, game makers will eventually run into situations where things start to get CPU-bound when using only a thread or two. They will have to either make their code much more efficient or spawn more threads. Both will require a lot of work and I think that multithreading will be a better use of their time as it can pay off much better in the long run, so they'll do it then. I think most games are graphics-bound at normal resolutions like 1280x1024, 1440x900, 1400x1050, 1600x1200, or 1680x1050 so it might be a few years before we start to see really multithreaded games.

But what you do see here are a lot of non-game programs being multithreaded. Encoders and compilers just love multiple threads, as do many engineering and math applications. Of course everything but the encoders have been running on MP setups for decades, so it's nothing new.

Not to jump in between you and the captain, but I'd have to say I'm closer to the captain's side...
I've seen, first hand, the amount of pirated software in various asia countries - from large companies using $3US (converted - or less) copies of photoshop cs to individuals w/ matlab and other complex software - not that they necessarily use it all...

either way - think of this - when it only costs $3 for a photoshop type of app that can utilize multiple cores, why wouldn't the "average" user use that vs. something else (I myself got a reduced price version just to check it out - it's not that complicated to do basic editing w/ it even as an amateur... i.e. you can't just count the "experts" that pay for it...)

that's not to say I think pirated s/w will actually drive people to quad core en-mass... but I think you're under estimating how much people actually use it... and if they don't spend $$ on s/w, they may just spend some more on cpu's or other hardware...

So is there a consensus here? The issue is whether it is _rational_ to want more cores, not whether people will buy them anyway, remember. This doesn't mean you're an idiot if you do so, but just that it's not a rational decision. There's a big gap between rationality and idiocy.

As far as the server side of things go, loads of cores will always be handy. We can continue to expect market adoption of quad-cores and more in the near future. (In fact, have you seen Arstechnica's little piece on why Intel did the 80 cores thing: quite interesting. Link: http://arstechnica.com/articles/pa [...] scale.ars)

As far as the client side of things go, quad-core already has some benefits for photoshop users, and so a small yet significant proportion of users would benefit from desktop quad-core solutions.

As far as gamers go, quad-core gaming is an intriguing possibility, but we would be well-advised not to take the example of Valve as paradigmatic, and thus it's a case of wait-and-see. For the foreseeable future (i.e. until 2009), gamers will receive little benefit from quad-core as opposed to dual-core, and the expenditure would not be justified on this point alone.

As far as the majority of users go, the same thing is true now that has been true for ages: they don't need to upgrade anything! So long as iTunes, Word and Internet Explorer work, their needs are fulfilled.



Oh, and just to link it back into the original post (here's where I'll just get flamed): Henri Richard is saying a lot of things that are in his interest, but that does not mean he is wrong. A lot of the quad-core adoption _has_ come purely from marketing pressure. You don't need to look far to see on these forums the straightforward assumption that 4 cores is automatically better.

Oh, whoops. I forgot. *sigh*...I must have skipped that class in college

Yes, I was just about to get back to the topic... whether or not it produces major "benefits" is moot - AMD (per Henri) are vastly underestimating the future demand for desktop Quad core. Reading between the lines, it's sounds more like wishful thinking on their part.

Seriously, what midrange to upper-range PC buyer is going to skip on a $200-$300 Quad CPU???

Frosties are ggggggggggggreat!!!!

Actually, I don't think my statement was far-fetched at all. There are 166 million PCs in the USA and over a billion worldwide (extrapolating the figures to population numbers), so I continue to maintain that this "secondary market" for quads is much bigger than anyone thinks.

Can anyone come up with a better statistical source for international PC numbers? It would be really handy!

TwistedSister: I agree that the point is moot as regards the way things will go. On this, I get the feeling Richard would be the sort of poker player who spends all night bitching about the cards they get dealt, rather than simply playing the hand he's got.

But it surely isn't moot if you're an enthusiast thinking about what sort of PC you'll need in the next couple of years. For instance, it tells me that my current setup, with perhaps a new GFX card once DX10 games start actually coming out, is not going to struggle for quite a while yet.

CRA: this site was good. 2004 figures, but they won't have changed that much. Note that it lists PCs, not computers in general.

http://www.nationmaster.com/graph/ [...] -computers

I'm very wrong on my previous 'fact'. Sorry! I'm eating a large piece of humble pie right now!

The US has 161m _personal computers_ out of a total of 452m. So the US only has 36% of the world's personal computers.

BUT: loads of the that 64% remaining is comprised of advanced industrial nations, who have as much problems with software piracy as the US, and where one would not expect the sort of practice you describe. You're not going to get very many pro users of Photoshop pirating copies and buying quads in industrialised nations.

That leaves a big chunk, I agree. I still can't really see Photoshop per se being that significant though. I mean, how many people use Photoshop to that extent?

I've got the eye of the tiger, it's the feel of the fight, rising up to the challenge of our rivals. With Tony beside me...

Come on...don't attribute seasonal shifts from the best quarter of the year (Q4) to Q1 as being due to AMD. Intel should definitely have negative growth in revenues QoQ - if they didn't, that would be incredible. Imagine being a company that has Christmas like sales twice during a year. 8O

Edit: Also, how do you "burn through cash" when after all expenses, you're still making money?

I'm not attributing that to slow seasons, it will be apparent if any drops in QoQ profits were caused from losing share.

Burning through cash is where you throw away profits and margins to hurt a competitor.

Oh cmon now...do you really have to ask? You know the answer

Its simple, you burn through cash when its a negative adjective as applied to the company you simply dont like.

Baronomics 101. The company in the red is healthier than the company in the black when those companies are AMD and Intel respectively. QED as one of my profs used to say.

God, another semantic stalker twisting words to make AMD buyers look bad. Intel is of course "healthier" in terms of cash flow and market, but that has nothing to do with Intel recording 2004 level revenues in Q4 2006. while AMD has increased revenue by 90%.

Go back to your hole, cave boy.

Really? Can you please link that?