A quiz for us all

[audiyoda]

Being that this is my 900th post, I'd like to submit a quiz. Two questions, no prize except that you'll have the knowledge that you really know your stuff.

Question #1:

Who said the following and where:

The only measure of performance that really matters is the amount of time it takes to execute a given application. Contrary to a popular misconception, it is not clock frequency (MHz) alone or the number of instructions executed per clock (IPC) alone that equates to performance. True performance is a combination of both clock frequency (MHz) and IPC:

Performance = MHz x IPC

Question #2:

Who said the following and where:

this design effort....focused on delivering an average instruction executed per clock (IPC) that was within approximately 10% to 20% of the P6 micro-architecture...

I'll give you a hint: Both quotes come from white papers, so you'll need to tell me the title of the white paper (where).

-Craig

[audiyoda]

Oh yeah, the attached graph goes with Question #2.

-Craig

[DrZaius]

I know the first one, it's a paper from Intel titled Inside the NetBurst Micro-Architecture of the Intel Pentium 4 (pdf (http://www.intel.com/pentium4/download/netburst.pdf)). I remember that intro from looking through it a while ago, and a quick search on Google confirmed it. As for the second one, I'm guessing it's from Intel as well?

Congrats on the 900th post! 100 more and make the big 1K.

[dead_eye]

Steve Jobs in the apple mhz myth video? Bunch of BS.

[audiyoda]

Good job Dr. Z! Now we've established that Intel does indeed believe that pure horsepower does not make a CPU...why then did they come out with the P4?

-Craig

[audiyoda]

Sorry dead_eye...but in a roundabout way, you're kinda-somewhat-sorta close.

-Craig

[audiyoda]

Oh come on -- no more guesses? I thought for sure HAL would get in on this.

-Craig

[Nelreem]

i woulda thought that the first was from jerry sanders about the new Athlon XP. i read something like that on the AMD website when i first heard about the XP. The next one is probably from intel. ill go check their article archives

[HAL9000]

I would, but after a weekend of re-arranging the entire house to be painted, hauling out garbage, getting rid of other junk, work, dentist appointment, 2 all day service calls, selling my old beater wagon, etc, I've let myself get a bit run down to the point of getting a good cold, and just don't have the energy to figure it out :(

[audiyoda]

To help HAL's cold, let me finsih this off---

The second quote is indeed from Intel:"this design effort [Pentium 4] focused on delivering an average instruction executed per clock (IPC) that was within approximately 10% to 20% of the P6 micro-architecture..."
-"Desktop Performance and Optimization for Intel Pentium® 4 processor"-Intel Corp., Feb 2001

Funny -- Intel actually is admitting that the Pentium 4 is "within 10% to 20%" of the Pentium 3. Now how sad is that? This is the actual graph that goes with the statement:
http://www.msu.edu/~carlencr/fullgraph.gif

So Intel has purposely put a CPU to market that is a dog to a comparable P3. By their own dogma if a 1.0Gig P4 existed it would be up to 20% slower than a P3 1.0 Gig CPU. But then by their own accounts a CPU should be measured not jsut by MHz/GHz, but by multiplying it's throughput (IPC) with clock speed to determine actual CPU performance. Something Apple/Motorola have done for over a year and AMD is now doing.

Just stirring up the pot -- Rather have stew than anything. :)

And get better HAL, We Wuv U :p

-Craig

[DrZaius]

Good posts audiyoda. My guess as to why Intel released the Pentium 4 when it has so few improvements over the Pentium 3 is marketing. People like big numbers after the word "Pentium," regardless if it is worth the extra money.

[highrisemech]

I don't know but it sounds like something Toaster would write.

[HAL9000]

Thanks Craig.

I really don't find it that unusual that the P4 at its release to be a bit of a dog in comparison to the PIII. What you will most likely see happen is as the PIII reaches it's maximum potential, performance gains per megahertz gains will reduce substantially while the P4 on the other hand will start to show performance increases as it reaches its "sweet spot". It too will then drop off as it maxes out its potential, and by then, a new core is released. This also happens with the AMD chips, but isn't nearly as noticable as they are taking a far more aggressive stance and alter the core more frequently than Intel does to keep the tight competition.

[Demosthenes]

I think the Pentium III was a pure marketing scheme, not the Pentium4. Still to this day, I ask myself, what is the Katmai core? The real Pentium III's are the Coppermines and Tualatins. The Pentium4 is an architectural revamp worthy of a new revision number. Maybe they are releasing a lower standard CPU so the PentiumV will look really cool. Yeah, right.

Respectfully,

Demosthenes

[jotto84]

the first question sounds EXACTLY like the question AMD was using in their quiz entry for winning an Athlon XP, in their huge giveaway last month

[audiyoda]

Considering I got them from a friend who works for AMD, that wouldn't suprise me.

-Craig

[highrisemech]

A lot of Intel's moves are money oriented. They have stiff competition from AMD
who by the way Intel almost pushed out of the cpu market years ago AMD was really struggling to hang in there. I'll bet Intel's stock holders make heavy demands. New definitly does not mean better. AMD has hung with the big guy on the block by means of innovation and quality. Where as Intel knew it could respond to AMD's attacks through name recognition. But now it seems that AMD has that same recognition and also becoming well known for a quality product.
Thus Intel has responded with the Pentium 4. I think where it will go from here given the fact that AMD and Intel seem to now be on equal footing, will now be
well improved products from both companies. If Intel stumbles now they will move into second place behind AMD for the first time and I do believe this will happen. AMD has won its place through vigorous R+D. Some people differ with this opinion but I still believe that us the consumer have benefited from thier stiff competition.

[audiyoda]

As long as uninformed computer shoppers continue to purchase those 'great' deals from the likes of Dell, Gateway and IBM that tout a P4 SDRAM system for $799, AMD will always be #2. Being #2 isn't all bad -- it keeps you lean and mean while Intel continues to sit back and work off name recognition.

I think within the next 5 years we will see AMD equal to even ahead of Intel in the general consumer's eyes. When those three computer makes get out from under the Intel shroud and starts to offer an alternative, consumers will jump at the chance to get equal/better performance at a lower price.

-Craig

[HAL9000]

Originally posted by highrisemech
A lot of Intel's moves are money oriented.

Isn't that the point of a successful business?

[audiyoda]

Partially. But complacency at the top has killed many a king of the hill. And that's the trap Intel is in right now. That's wat killed many dot.coms -- they thought they could ride the tide of technology, telecommunications and the internet -- not so. Some certainly survived and survived well, while others died terrible deaths -- there didn't seem to be any in-between. Question is, will Intel do the same? Now that it looks like Motorola may get out of the chip business and Apple seeming to sway the way of AMD if that happens, Intel could become a lame duck.

-Craig

[HAL9000]

True enough, I don't argue that. If Intel was to make to many mistakes and AMD enough right moves, a switch of power could easily exist, my point was a bit more basic than that in the sense that any move you make in business comes down to the all mighty dollar. Even in large corporations where they keep a location that is losing money alive, it's all about a tax write off that in the end will put a couple more dollars in the bank.

[audiyoda]

Agreed.

-Craig

[highrisemech]

I also agree Hal but in your quest for that mighty dollar you can't lose sight
of the future dollars.

[HAL9000]

Also agreed, but unless you know the entire details of a companies marketing strategy, you can't define one thing as a mistake as it may be an instance of let this go a bit here, but gain huge there.