Intel, OEMs Face Lawsuit For Megahertz Marketing

prostoalex writes "A group of PC owners filed a lawsuit against Intel, Gateway and HP, stating that companies spread misleading information about Pentium 4 processor performing faster than Pentium 3 or Athlon. The complaint alleges that 'the Pentium 4 is less powerful and slower than the Pentium III and/or the AMD Athlon.' PC World has more details in its story." I wonder if the same litigants have a suit against the USPS for ads leading one to expect prompt service from courteous, competent employees.

Comments

It's true even on the P4 Xeon level.

[t0qer]

Just recently I had a neighbor hire me to do a concept animation of a machine he was going to build. I used truespace 5.2. It was insanely detailed down to individual links on the bicycle
chain drive.

The poly count got so high that my P4 was going to take 3 days to render it. My computer could hardly handle moving around in the scene anymore. I told the neighbor I had brought the scene as far as it could go on my P4
and I couldn't go any further without a new machine. He gave me $2500 to work with so this was what I built.

Dual Xeon P4 2.0ghz
1 Gig RDRAM
Maxtor 80gig IDE drive
DVD-R(by his request)

The system definetly cut the rendering time down, to 24 hours,but something just didn't feel right about the new render time. I could
have bought 2 more p4 1.4ghz and accomplished the same for less. What really got me was when my friend rendered the scene on his single athalonMP 2200.

14 hours

A single athalonMP 2200 was smokin my dual xeon setup! Well, this is all it took for me to write off intel forever. Intel fuck you and your shitty CPU's, you've lost my trust forever!

Anyone that is even considering using a Intel solution as a renderstation, please don't waste the money. You can do a lot more with a lot less using AMD.

Re:It's true even on the P4 Xeon level.

[Sivar]

Applications that are "specifically optimized for the Pentium IV" generally are identical but use the SSE2 instructions. Sometimes they are compiled around some of the P4's rediculous weaknesses, such as it's incredibly slow handling of bit shift instructions. (8 clocks, IIRC on the Pentium IV. The Athlon can do up to three per clock

Anyone can tack on vector instructions to a CPU. The problem is the underlying architecture of the P4, which isn't as easy to fix.

The AMD Hammer series will have those same SSE2 instructions AND a superior architecture (to even the AthlonXP).

Where will the Pentium IV be then?

The facts ...

[Monkelectric]

At the time the P4 came out, AMD's Athlon processors were SMOKING Intel, and that is Intel's fault, they certainly could employ enough engineers to destroy AMD. Fact is, they got lazy. Previously AMD processors hadnt been as stable as Intel and they could still sell on that point but to AMD's credit by the time the Athlon came out it was a stable platform (still had a couple minor issues). And Intel was worried.

What Intel has been doing to make chips faster ever since the 486 has been adding more execution units. The 386 had 1 execution unit, 486 had two, PII and PIII had 4, and I *think* the P4 had 8 units? ... Anyways, this is really like putting more tires on your car. It SOUNDS like more, but you ain't goin any faster, the fact is that 4 execution units and 4 wheels is about as many as people will ever need. The problem is, that it becomes impossible to schedule instructions for 8 units, having 8 instruction units is essentially saying, your code should have 8 seperate threads [using the term threads loosely] that dont depend on eachother to avoid interlocks ... *IMPOSSIBLE*. Second of all, intel stretched their pipeline to 40+ stages, this means that the penalty for pipeline stall, branch perdiction miss, context switch, etc is *HUGE*. AMD's Athlon pipeline was a lean 7 stages.

Why did Intel do this? They were scared because AMD beat them at their own game. Intels self esteem was damaged -- So they launched an agressive marketing campaign, and used these tactics to maniupulate the marketing metric, MHZ. Ceartinly sleazy.

You'll notice now that Intels best P4 is faster then AMD's best part right now -- they've backed off the agressive advertising. However, they burned enough geek karma that I'll never buy intel again.

To remedy the situation, processors ratings need to be measured in IPC*MHZ [instructions per cycle] for both integer and floating point operations. Then it would be pretty clear to consumers what was going on.