Sign in with
Sign up | Sign in
Your question

True Performance Matters

Last response: in CPUs
Share
October 20, 2001 3:42:06 PM

Reading Toms MPF 2001 it seems most companies (INCLUDING INTEL) agree that the true performance matters rather than pure Ghz. Which is why it irritates me that Intel continue to market thier P4 processors on Ghz alone, while with Itanium it doesn't matter.

AMD have a right to start the 'True Performance Initiative' to get the common guy / girl on the street in line with us geeks and the rest of the industry. Whether through marketing or teaching it is important to get that message across.
However I believe the PR rating that AMD have used in thier XP processors is wrong I think they did this to act as a stepping stone on to what will become the 'True Performance Initiative' which will be supported by industry worldwide. (Which AMD reckon will be when Hammer comes out).

I believe eventually AMD will drop the 'XP PR Rating' and move onto the 'True Performance Initiative Rating' when it becomes an accepted standard.

Most companies in the Server / MP business seem to agree on this anyway e.g IPC ratings (Instructions per clock) etc. However the oppisite seems to be true in the Desktop sector which is why I think AMD are right to head the intiative with backing of other companies.

Providing that this intiative is correctly done with full support and backing of the MAJORITY of other CPU manufacturers i.e IBM, Motorola, Apple, AMD, Sun, ARM,
Intel - (will be reluctant)

Whether it succeeds or not (which I doubt it will with Intel not in favour of it) what matters is that it IS a step in the right direction.

Sprinter vs Triathlon is not the way to go (as Tom stated in his artical).

PLEASE NOTE: I am specfically talking about the CPU, and NOT a system as a whole.


<font color=purple>~* K6-2 @ 333MHz *~
I don't need a 'Gigahertz' chip to surf the web just yet ;-)</font color=purple>
October 20, 2001 4:05:59 PM

OK this is sincere reply, not intended for any flaming. You say Intel agrees that performance matters yet they still market more Mhz, Ghz? Will someone please show me where or how Intel is saying "more Mhz is better". In the Blue Man ads I know they didn't. And in the new set of ads they have ready they don't either. Just because they are truthful and still label their processors with Mhz rating deosn't mean that they are marketing it that way. So I await some examples. Thank you
October 20, 2001 4:13:29 PM

<A HREF="http://www.cpuscorecard.com" target="_new"> The CPU Scorecard</A> is the closest thing to comparing the true overall performance of every *86 processor. As you can see, the Athlon XP 1800+ is faster overall than the Pentium 4 2 GHz.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
Related resources
October 20, 2001 5:05:19 PM

Well that's nice and all, but I still want someone to point out to me where Intel advertises more Mhz is better. I mean everyone seems to bitch about it so it has to exist somewhere doesn't it?
October 20, 2001 8:59:12 PM

It's not a direct advertisement by Intel; it's a common misconception that Intel conveniently refuses to correct. Intel caters to that misconception without either affirming or denying it.

I believe the technical term is "deceit by omission."

Kelledin

"/join #hackerz. See the Web. DoS interesting people."
October 20, 2001 9:23:58 PM

Actualy, the move to the Pentium 4 processors was to allow for the increase in the Mhz/Ghz. Specificly to make the processors look "Sexy" to your common man.

When it was released, on a per clock cycle basis, the PIII Beat out the PIV, quite handily too. Check out http://www4.tomshardware.com/cpu/00q4/001120/index.html and take a look. While much of the hardware around the PIV has improved, it's still built for a high clock speed more than it is built to be an efficiant processor.

It will be interesting to see if the Athlon XP line makes it to 2 Ghz, and what sort of performance it will have.

60 FPS, 70 FPS, 80 FPS Crash!
Daylight comes and I have to go to work :frown:
October 20, 2001 9:29:15 PM

So how does Intel cater to this deceit? So what they have an inefficient architecture right now, that requires high clock cycles. I really don't see any deceit. They don't claim to beat Athlons clock for clock and they don't advertise or "deceit by omission" just because they have high clock cycles.
October 20, 2001 9:36:43 PM

I don't think I ever saw Intel ads saying they're faster than cpu a b or c, like AMD does, right now Intel's ad campaign is based on performance where it counts.

It's too bad not many programmers share the same skill level as Id software's John Carmack, most programs are not written good and when optimizing for the P4 it shows how sloppy the programmer really was.

"<b>AMD/VIA!</b>...you are <i>still</i> the weakest link, good bye!"
October 20, 2001 9:48:02 PM

Quote:
right now Intel's ad campaign is based on performance where it counts.


Yeah, we all know the general populace only uses computers for Quake 3, Video Editing, and Photoshop.

Laugh.




:tongue: Have you ever tried cooking an egg on your HSF? Tasty. :tongue:
October 21, 2001 5:52:06 AM

Intel doesn't directly advertise their product as being "faster," they let common misconception take care of that. You don't see Intel employees saying, "Our CPUs are faster than Athlons," but you don't see them saying, "Well...clock speed really isn't everything," either (unless you leave them no room for evasion). And honestly, the average guy who walks into a computer store wanting to buy a new computer isn't even going to ask such a question--he's going to judge the value of a system by the big numbers on its ad. That's what Intel caters to.

Do you even know what "deceit by omission" means? It means indirectly perpetrating a lie merely by what you <i>don't</i> say. That's actually more serious than it sounds; a certain company called Rambus Inc. has been convicted of fraud based largely on that concept.

Kelledin

"/join #hackerz. See the Web. DoS interesting people."
October 21, 2001 6:10:06 AM

Quote:

It's too bad not many programmers share the same skill level as Id software's John Carmack, most programs are not written good and when optimizing for the P4 it shows how sloppy the programmer really was.

Ja, id has some great programmers. That must be why the AthlonXP <A HREF="http://www.anandtech.com/cpu/showdoc.html?i=1543&p=10" target="_new">reams the P4 so badly in RTCW.</A> :lol: 

Kelledin

"/join #hackerz. See the Web. DoS interesting people."
Anonymous
a b à CPUs
October 21, 2001 9:46:12 AM

"While much of the hardware around the PIV has improved, it's still built for a high clock speed more than it is built to be an efficiant processor."

The thing is that the PIV is still a fairly new design. As I understand it, it just needs to mature a bit, ie specifically: it needs to improve it's chances of making cache hits for the data it requires, and that would allow it to perform much more efficiently. Bigger caches could be a poor mans way of getting there, but probably better prediction will be what is needed in the end. We'll have to see what they come up with next.
October 21, 2001 10:56:20 AM

oh wait!

intel just anounced that the Mhz mith is broken!
they just told every one that acrtually the AMD chips were faster then the P4's so dont buy them!
lol, right..this is what you expacted?

<font color=green>
*******
*K.I.S.S*
*(k)eep (I)t (S)imple (S)tupid*
*******
</font color=green>
October 21, 2001 12:10:59 PM

AMDMeltdown, have you been keeping up with the reviews? The Athlon XP is generally as fast or faster in tasks that were traditionally the forte of the P4. It was data-prefetch and SSE that gave the P4 it's performance all along.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
October 21, 2001 9:18:23 PM

Well thank you for clearing up that whole "deceit by omission" there Mr. Dictionary. Yeah I know what it means and it has two things to do with Intel's marketing...jack and sh!t. Listen just cause Intel still keeps it honest by showing a Mhz rating they aren't hiding or deceiving anything.
October 21, 2001 10:28:57 PM

I've been to Intel seminars and so on and they are quite clever. They claim to be "50% faster", or "optimized", etc, but they only compare to themselves. It's usually them comparing a P4 to a P3 1 Ghz.

They don't have to say "we're 500 Mhz FASTER", it's the fact that they use bigger numbers rather than benchmarks to compete.

Quote:
An advanced microarchitecture and clock speeds reaching 2GHz enable the Intel Pentium 4 processor to meet business computing demands today and in the future.

An optimized processor core and powerful new instructions make the Intel Pentium 4 processor a peerless engine for multimedia, communications, and engineering.


<font color=red>God</font color=red> <font color=blue>Bless</font color=blue> <font color=red>America!</font color=red>
October 21, 2001 10:36:08 PM

ONLY if the software is written for it....
to me MMX, 3dNOW SSE & SSE2 feel like cop out options. father than make more efficient cores they have all new instruction sets.
(and both amd and intel are guilty of that)

P.S. i noticed Serious Sam has 3dnow+ optimisations... anyone reading this with a P3 or P4, does Serious Sam show SSE or SSE2 optimisations too?

Religious wars are 2 groups of people fighting over who has the best imaginary friend.
October 21, 2001 10:49:44 PM

Well I find it VERY decieving. I went to a Microsoft event which was being sponsored by Intel and it was a huge freakfest of Intel and Microsoft people claiming that WindowsXP is made for the Pentium 4. Yet WinXP is made for both AMD and INTEL processors. I think they'll say anything as long as someone writes them a check. If I was to take Intel seriously and not be a smart consumer then I would totally IGNORE AMD and buy a P4 since it's 87% faster than the P3 1 Ghz. They are hoping that the average consumer will completely ignore AMD so AMD has taken action to set forth another performance standard. They're both being very smart about this.

<font color=red>God</font color=red> <font color=blue>Bless</font color=blue> <font color=red>America!</font color=red>
October 22, 2001 12:55:04 AM

Ja, and in case you didn't notice, a lot of us are bitching about AMD's PR rating. Neither company is completely honest; it goes both ways.

Kelledin

"/join #hackerz. See the Web. DoS interesting people."
October 23, 2001 10:11:11 AM

Guys, AMD only decided to call for performance ratings when they fell behind in the GHz war, remember who was first to 1 GHz? I have no problem with performance ratings, but it needs to be by an independent panel, not some large number the company's marketing dept pulled out of thin air. And all this bollocks about Andersens auditing the benchmarking methods!!! That's like asking shipbuilding company to audit an aircraft builder, i.e. they know Jack about what AMD will show them.
October 23, 2001 11:22:52 AM

Tejas, they were fighting the p3 then, and back then mhz MEANT something, when the p4 came out, intel changed the way the game was played, and AMD is merely trying to make sure that the ignorance of the computer buying public does not give intel an unfair advantage.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
October 23, 2001 12:31:33 PM

When AMD reached 1ghz, AMD's performance per cycle was only marginally better than P3 or essentially the same. So, 1ghz on an Athlon meant the same performance probably within a percent, as a P3 at 1ghz.

When the P4 reached 2ghz, it was not much of a revelation because the Athlon at 1.4 had nearly the same performance, and the Athlon at 1.5 outperforms the P4 2ghz in many cases.

So what's the issue?



When all else fails, throw your computer out the window!!!
!