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

[damocles]

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

That's my favourite


Simple advice- don't mess with non-prescription drugs

 

[Noriaki]

I dunno damocles, the SSE2 one is pretty good but what about this one:

<< The P4 and Rambus may mature well but it does not have much time as Intel wants to ramp 64 bit cpu production and dump 32 bit asap. >>

Considering Intel has clearly announced that for a copule years at least the Itanium IA-64 is going to be Server and High-End workstation only, and that the x86-32 will continue in the desktop space for some time yet, that one is pretty ridiculous.

Although it doesn't crop up as often as the SSE2 one, becuase AMD fanboys use the SSE2 to advocate AMD over Intel...which happens a lot!

 

[Wingznut]

Favourite... I'm not sure I can pick just one. Noriaki's is probably my favorite, but this one is good, too.

&quot;it's too bad they didn't leave it there a couple more months and release it the way they had planned to. Instead, they rushed it off the drawing board...&quot;

 

[Duvie]

I agree that the sse2 is a non factor in terms of the p4 running as an effective computer...It does well without the optimization and the sse2 optimization will make it shine a bit more when it becomes widely available in popular apps.

However it is usually the intel fanboys that say &quot;wait until programs start using the sse2 coding&quot;...Many just got tired of the sad excuses by intel fanboys defending why in benchmark testing at many many reputable sites that the 1.7 p4 could only stand even with the amd 1.33 ddr. So don't think that started as an amd bash of p4...Remember amd fanboys had numerical data of the benchmarks on their side they didn't need to make excuses...

now with the future amd chips coming with sse optimization and a bit more AMD too will benefit from this, so it won't be just a boost for Intel...

 

[Noriaki]

Sure the P4 1.7Ghz only hangs with an AMD 1.33GHz....

That's fine....on release a PPro-150 was only about as fast as a P120 or 133....that's about the same Ratio....

The P4 will ramp up and improve in speed.

Besides...the P4 has a huge pipeline, it was never meant to match the Athlon clock for clock. That's why it's available 400Mhz higher speeds.

And it hasn't even really started to Ramp yet.

 

[Duvie]

I agree with you...I am just stating that sse2 was not an amd fanboy saying...

We have been over that battle of mhz for mhz wars and how different architecture can't be compared, blah blah blah...

I know the p4 will ramp up better in the future...will it mean a p4 2gig actually beats an amd 2 gig...probably not...too much to make up...

Sure sucks to be that early adopter who gets stuck with a slow performing low end mhz p4...I think Intel probably should have started this line at 1.5 or 1.7 and forgot about the 1.3 and 1.4 and let the tualitin p3 chip fill in the gap...

 

[LocutusX]

<< The P-Pro couln't take a P5.... >>





<< That's fine....on release a PPro-150 was only about as fast as a P120 or 133....that's about the same Ratio.... >>



This is a rather misleading &quot;myth&quot;. The only reason it couldn't match the performance of a P5 was because the platform used in these comparisons were 16-bit - such as Windows 3.1, or even Windows 95 (which had just been released).

Switch the OS to a Unix or Windows NT and the PPro really began to fly. At the very least there was a 20% improvement in performance over an equally clocked Pentium Classic, assuming 32-bit OS and 32-bit applications.

The only &quot;optimization&quot; Intel did to ensure the Pentium-II did not suffer from the same problem (since at this time Intel woke up and noticed that Windows 95 had a significant amount of 16-bit code) was by the addition of a segment register cache. This cache was originally removed from the PPro because the designers felt that by the time the PPro was released, all OS's and apps would be 32-bit and therefore not need one.

The problem is a bit different with the P4. At least with the PPro it was inevitable that everyone was going to move to 32-bit eventually... the impetus was already in the industry and in the market. I'm not so sure about the P4.

 

[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?



<< What do you think? Is Intel gaining ground, holding even, or sliding? Not just in terms of sales, but technology. Sales is today, but technology effects tomorrow's sales. >>

In terms of sales, it's pretty much evened out with Intel holding an 80/20 overall. In desktops, esp. higher end, and obviously esp. in geek machines, it's probably more like 75/25 or 70/30 desktops as a whole, and 60/40, maybe 50/50 in the DIY/geek machine category.

As for the future, Intel has it in spades over AMD. Not so much in the desktop arena, they are definitely shifting some focus away from that to concentrate on higher margin chips - servers, specialized DSPs (gotta love the StrongARM+DSP+FLASH all in one killer chip), mobile markets (StrongARM here again, plus other future low power stuff). AMD has no competition for these areas, excepting for servers, and that is very weak ... only up to 4 CPUs, am I not correct?


Bottomline, AMD has started a dangerous price war in the CPU market. That is why you see Intel reacting the way they are. The price war has basically &quot;commoditized&quot; the CPU market, cutting profit margins from 50-60% to 20%, even less. This (commoditization) happened to the memory market a long time ago, and a little memory company named Intel decided to exit that business in search of more profitable things. Along came this cash register company that needed a CPU and ...

AMD is willing to sacrifice profit margin to gain some marketshare. Intel is playing along for now, and it hurts them far less financially. But their whole company history has been one of changing for the future. That is what you are seeing now at Intel. Within 5 years, less than 50% of their business will be on the desktop. They will be more like a TSMC combined with an in-house design service than the Intel you see today.


One last note - about Intel &quot;supposedly buying equip. to produce RDRAM&quot;.

'Spose it's partly true, though that's stretching it a bit. The fact of the matter is that Intel bought (up to 20%)/invested in Samsung, the company, which does a lot more than just produce RDRAM.

But then that raises a very interesting point ... Samsung, arguably one of the finest memory producers for a long time (dang, them PC100 -GH chips are sweet !!! ), and IIRC, the largest producer as well, is the driving force behind RDRAM production at this point. Hard to imagine a company like Samsung would be involved if they did not see (eventual) potential in producing RDRAM. Pricing is no problem to them, that can be cured with mass production, to which Samsung is no stranger. That only leaves the latency problem. Now THERE is a hurdle ...


Look out Wingz, you need your hip waders in here!

 

[Vrangel]

Midnight Rambler,
I hope you used recent rally to sell INTC .
Intel had a good decade-long run, its over.

As for hope that it'll penetrate other markets its a pipe dream.
They will only lose money there. Intel has been a monopoly long enough
to forget how to compete just like IBM in the eighties.

It will take new outside management team to change things.
If it ever happens. It did happen at IBM, but not at DEC or Xerox...

 

[Wingznut]

Over? You don't think it'll ever go back up again?

 

[PliotronX]

I don't think the P4 is a blunder. It's in the middle of a transition into a new &quot;era&quot; of applications/games. The more I read into the P4, and actually have had a chance to play around with a P4 on my friend's machine, I kinda feel embarrassed for knocking the P4.

The benefits of extra long pipelines and extraordinary amounts of bandwidth may not be immediately visible to you, but you should look a bit deeper. Intel knows what they're doing, don't ever think otherwise...

 

[Vrangel]

Wingznut,
I updated my post above.

IBM stock was down from 1987 to 1997.
It took a full decade to recover.
DEC and Xerox never recovered...
Good luck.

 

[Wingznut]

How Intel is reacting to this &quot;recession&quot;, is extremely encouraging. And the companies/technologies that they've been acquiring is also very interesting.

Personally, I'm really enjoying soaking up all the Intel stock that I can right now. Heh, but what else do you think I'd say.

 

[Vrangel]

I know a guy who put his retirement money in INTC at $47.
Oh well , I guess he will have to work a little longer...

 

[gunf1ghter]

People that put all their eggs in one basket deserve what the market hands them.

 

[Darksamie]

I think the competition has them scared. It looks like they are just trying to win the clock speed race against AMD to look good in the marketplace rather than put the money into R&amp;D and get something on the market that is a good quality product.

I really don't know who thought of all the ideas behind the Pentium IV chip, but they must have been pretty damn smart....the only problem was that they cut half these ideas out of the chip before it went into production....

 

[Vrangel]

BTW Intel outspends AMD in R&amp;D by 10 to 1 or so.
I mean AMD shouldnt even exist by now, nevermind gaining market share.
I hope no one is insulted but its fairly typical bloat for a monopoly.

 

[BFG10K]

fkloster:

and NOW we have a motorola G4 733 that absolutely spanks ANY Intel or AMD part in clock for clock performance

Err, no it doesn't. For 3D games a P3/Duron/TBird spanks a G4 on a clock-for-clock basis. And a G4 733 has about the same SPEC rating as a 750 MHz TBird. The difference isn't as large as the Mac zealots make it out to be.

Not to mention a 750 MHz TBird is pocket change while a 733 MHz G4 costs practically the same as an entire PC system.

 

[ST4RCUTTER]

For a company that spends so much more on R&amp;D than their closest rival, you'd expect them to have more patents, not less.

 

[Finality]

I do not believe it will be a big flop. The P4 was made for speed not efficiency. Once we start passing 2Ghz (compared to current gen AMD) the P4 should be a clear winner.

The only downside to the P4 is the way it was tied to Rambus is its only downfall. It will be intersting to see what the future holds.

 

[joohang]

<< 99% of buyers only consider clockspeed when they purchase. >>


Ever thought of it this way?

Although it will take time and Intel could &quot;recover&quot; by then, people may begin to believe that Intel CPUs are &quot;slower&quot; than AMD CPUs.

But then somehow I doubt that average computer users will ever think beyond clock speed.

 

[zzzz]

Its not a blunder. People will keep on buying them. They would have probabaly introduced it without AMD's competition. What worries me about P4 is the socket change that is going to take place(Is this true? I have no idea).
Also I would like them to drop the prices of current P3's to T-bird level

 

[martysweb]

AMD is out to kill Intel and Intel is out to kill AMD. But AMD has been realy quite latly and Intel has been making alot of noise.Remember when your in a bar fight its the quite one's you got to watch out for..Also Intel has been raping people for years with there high prices and now there droping there prices??I dont think they can make it having to compete in the price war game because after years of raping people they dont know how..

 

[shathal]

I wouldn't know about AMD vs. Intel (though I can imagine their fantasies) - but I got it from VERY good sources that Intel is actually interested in keeping AMD &quot;alive&quot;. Intel isn't stupid, and looking at Microsoft's &quot;dominance&quot; in the OS-market, it doesn't want to deal with anti-trust &amp; all that stuff. As such, Intel is happy with competition - it just prefers to be winning as well .

The P4 isn't a blunder - (You don't think they made the P4 because the lost the 1 GHz race, do you? Do you know how LONG CPU design-cycles are?) and it has nothing to do with the 1 GHz lost race (they're going to win on 2 GHz and up I would say). Again, the move from P3 to P4 is very much like the move from 486 to Pentium. The architechture was radically different, and both needed some optimised code to &quot;really come to bloom&quot;. The P4 is a replay of this.

If you need another example - think of the GeForce3. There's a nice, speedy card. However, it will only be able to show off its TRUE unlocked potential once the propper DirectX 8 effects are used. It's a bit faster than the GF2 Ultra, but there's a lot more quality in it - however, one has to USE it, to notice the differences.

Hope this clarifies a few things.

<Flame-Safe comment I hope .>

 

[shathal]

Re: Research &amp; Promotion.

One &quot;problem&quot; I think we're all having is that the companies (BOTH Intel *AND* AMD) are thinking like a business (well - nothing &quot;wrong&quot; with that, they're not charities after all). They don't want to throw out ALL new technologies available in a single burst, only to then require a few years to come up with something new.

They release &quot;novelties&quot; gradually - so the P4 might come in these steps:
* 0.13 (that'a a given).
* 133 MHz/533 MHz FSB.
* SSE3 (or whatever)
* Another die-shrink
* And so on...

It makes sense from a selling point of view. Damn shame from an End User point of view, but that's what companies do. The same goes for car-companies - they've had better engine-technologies for years - they just release them a bit at a time... .