Is the P4 as big a blunder for Intel as IBM's MS DOS miscalculation?

[Degenerate]

<< P4 requires SSE2 instructions to obtain reasonable benchmarks. Who is going to buy all new programs just to be competitive >>



About all this SSE2, would not the software makers want their programs to run faster? natually, they adapt and catch on. They make use of technology and try and beat other software companies.

It wiil come.

 

[potz]

Isn't the new P4 supposed to have a bigger cache and such? The performance should be better, and it'll be on .13 micron process i believe, so it should be cheaper. If intel is able to ramp up clockspeeds quickly (i mean very quickly, quicker than now) then it has a chance. It still has a stranglehold on the main corporate market, thanks to Dell.

 

[Degenerate]

when the Athlon 4 1.5 comes out in when? june, july? P4 will reach 2.0. every step, they gain 100mhz in clock. Perhaps that is a worry for AMD.

 

[fkloster]

<< Err, no it doesn't. For 3D games a P3/Duron/TBird (733) spanks a G4 (733) on a clock-for-clock basis (performance)... >>



What crack pipe you been hitten'. I'll take that bet...

 

[shathal]

Isn't the new P4 supposed to have a bigger cache and such? The performance should be better, and it'll be on .13 micron process i believe, so it should be cheaper. If intel is able to ramp up clockspeeds quickly (i mean very quickly, quicker than now) then it has a chance. It still has a stranglehold on the main corporate market, thanks to Dell.

Oh, Northwood certainly has some interesting changes lined up - I guess we'll see the definite versions only when they launch it.

Re: Programmers incorporate new technologies
Actually - no. That is the big &quot;toothache&quot; - they tend to stick to &quot;standard&quot; compilers, not ones optimised by AMD or Intel. So, everybody has to wait for Borland/Microsoft &amp; so on to catch up. We're NOW seeing full SSE implementation - SSE2 usage will not happen tomorrow (or next month) - but it WILL come gradually . I can understand why programmers do it though - it's for reliability reasons.

AFAI am concerned, I would say that they should deliver encrypted source-code on CD's and then ask: &quot;Do you run (A)MD or (I)ntel CPUs?&quot; and then let the optimised compilers compile the source-code for the proper processor. As such, I doubt it'll ever happen, as nice as the thought is ...

Ah well - one can dream .

 

[Packet]

From a pure buisiness point of view

does it matter which company produces the best performing chip?

Nope, not really...

Does it matter which company sells the most units and makes the most money?

Thats the goal right there

 

[Micronaut]

And that's what got the big tobacco companies in trouble.

Just sell - who cares what you're selling.

Not gonna work - for long. And as long as computers don't contain nicotine, I think we'll be okay.

 

[gunf1ghter]

Well you know it's funny you say that.

We are all upgrading like mad, but my buddy who is getting his MBA at ASU just did a study with other grad students and he claims that the demand for higher end processors will start to taper off later this year.

Basically us game monkeys only make up 5% of the overall computer market and Intel and AMD can no longer convince joe six pack that he needs the latest greatest computer on the market to do exciting things like run quicken and surf the web.

You can see some of the fear of this in Intel's marketing engine when they initially tried to position the PIII and PIV as &quot;speeding up your internet experience&quot;.

I personally don't think that the race for more speed will ease up anytime soon, but what we really need now is a superior architecture, not more CPU cycles.

 

[grant2]

<< what we really need now is a superior architecture, not more CPU cycles. >>



What does matter how 2 cpus work if they can finish the same task in the same time? Why would we &quot;need&quot; one instead of another?

do you plan to crack open your die and examine your superior architecture CPU with an electron microscope for aesthetic pleasure?

 

[Alphacowboy]

net adds baby, net adds... LOL

 

[gunf1ghter]

grant2,

It is evident from your comments that you are not very knowledgeable about CPU architecture. Every time they have made an architecture improvement (such as the quad pumped bus, deeper pipeline, more cache, SSE, etc) it has had a MUCH greater impact on the overall performance than adding 100 mhz to the clock speed.

 

[Hulk]

midnight rambler -

<< Intel IS scrambling. Note the rushed release of P4 1.7GHz at the lowest ever prices for the top of the line chip as well as price reductions across the entire chip line. >>

Now I've seen everything ... people complaining about Intel lowering its prices to previously-unheard-of levels, esp. at introduction. WTF?


Hold on a minute sport. I'm not complaining about the low chip prices. I'm also not comparing AMD and Intel products, only the different paths they have both taken into the future.
The point of that comment was to illustrate the fact that Intel IS feeling the heat from AMD. Why else would they release a new part at such a low price? Where in my comment do I complain?

In the early 80's, IBM tried with various new technologies to wipe out the clone computer manufacturers. It didn't work. The strategy was a blunder on their part. It seems that whenever one company has a HUGE part of a market and the competition catches up, they try to redefine the game. AMD surpasses Intel with mainstream Windows performances so Intel says, &quot;oh, that's old technology, we're moving on.&quot; Problem is the mainstream public and software engineers have not moved on.

I'm only suggesting that Intel may not have the might to sway the market as it hopes it does. RAMBUS is another example. I'd call it a blunder. If you put enough of these blunders together bad things can happen. Ask IBM.

 

[grant2]

<< It is evident from your comments that you are not very knowledgeable about CPU architecture. >>



It's funny how you dodged the actual question... let me ask from a different perspective:

Please explain why I should care that my processor has &quot;superior architecture&quot;? (please note- that's a very vague term which is why i put it in quotes)

At the moment, I only care about more tangible aspects of my processor... price, performance, compatibility. It wouldn't bother me one bit if the thing was powered by a hundred elves juggling abacuses as long as my rig ran fast.

 

[grant2]

What i mean to say is,

if we can get the same price &amp; performance from different methods, we do we &quot;need&quot; one or another?

 

[gunf1ghter]

Because you hit bottlenecks.

I'm not going to be more specific. If you don't believe me or if you think that a 4 gig chip based on todays technology would be a godsend then more power to you.

 

[BFG10K]

fkloster:

What crack pipe you been hitten'. I'll take that bet...

No crack pipe and no betting.

Just run as many cross-platform benchmarks (3D games) as I have and read through as many developer plan files as I have (Carmack, Sweeney etc). Also be sure to check out Moto's and AMD's spec scores.

You'll see that I'm right.

 

[grant2]

<< Because you hit bottlenecks.
I'm not going to be more specific. If you don't believe me or if you think that a 4 gig chip based on todays technology would be a godsend then more power to you. >>



If both CPUs have equal performance, they will become the bottleneck in a system under the same conditions.

I'm not disbelieving, i'm just asking you to explain why I &quot;need&quot; a certain kind of CPU if a different kind could give me the same price, performance, and compatibility.

 

[gunf1ghter]

What do you mean by &quot;equal performance&quot;

there is NO such thing.

you might get parity in a specific application but tests have shown definite differences even at the same clock speed between the P3, P4 and AMD K7 architectures.