Slashdot Mirror
AMD To Hide MHz Rating From Consumers
pezpunk writes: "Tom's Hardware is reporting here that AMD's next-generation Athlons will be identified by model number rather than Mhz rating. This means that an Athlon will be designated an "Athlon 1600" even though it's only a 1.4Ghz part. The true clock speed of the chip will NOT be shown either on the chip itself or even in the BIOS. Apparently, they're desperate to compete with higher-clocked Pentiums in the minds of consumers -- proof that even the underdog can pull dirty marketing tricks =("
Don't they already have a P rating or something of the sort? It's from way back when they did the k5's I believe.
It makes sense to me. Lower MHz Athlons are always compaired to higher MHz P4's in benchmarking and stuff. It just proves the MHz isn't everything.
Which is the marketing scheme? The faster MHz? Or the better chip????
Good quote, too many chars. Seriously, the slashdot 120 char limit sucks!
What next? Car model names? Oooooh the AMD "Mustang SHO"! Buy one and get laid everyday! Howzabout the AMD "Shilznatz" for the Thug in us all - faster than a Glock 380.
D
The first, last, and only tech news site on the net
I fail to see how this will help. It seems to me that it will only confuse the consumer. You take the one piece of data that the average buyer uses as a benchmark (the MHz rating) and completely obscure it.
It seems to me that the consumer would be better served by AMD advertising in plain language why their chips are better than the competition's.
Look at it this way, if you went to the gas station and the pumps were only listed as "Formulas One, Two, and Three" instead of octane ratings, you'd likely buy the cheapest one instead of the one best suited to your needs.
"Tell me doctor, with all of your defenses, are there any provisions for an attack by killer bees?"
I suppose that will still work correctly right? Isn't that taken from kernel callibration routines instead of BIOS?
On my Athlon 700:
cpu MHz : 700.044
If it helps AMD get the market share and laurels they seem to deserve, great! Maybe it will force Intel to be more innovative in their architecture design sessions than they are in their marketing sessions.
It's not really that bad of a "dirty trick" since an 1Ghz Pentium Processor and a 1Ghz Athlon Processor are not exactly equal... but because they are both "1Ghz" they ARE equal in the minds of many consumers.
Benchmarks usually place the like-clockspeeded Athlon at slightly faster then it's Intel competitor... but it becomes hard to market that.
Hiding the clock speed from the BIOS though... going a bit too far.
Also, whatever 'P' rating you rate it at is meaningless. An Insel chip may be faster at integer math, slower at memory access and floating point while an BMD chip may rock at floating point but be terrible at other things. Plus, are we comparing against the PQ3 or the PQ4 Insel CPU?
No, keep the information about Mhz right on the CPU. Ideally, keep the FSB and multiplier as well. But just don't use this as your selling point.
Oceania has always been at war with Eastasia.
AMD should go for this all they way! After all, we all know how well trying to hide a chip's REAL speed rating worked for Cyrix! oh, wait....
Seems to me that a marketing ploy like this will not work against a marketing giant like Intel...
Intel will simply exploit the fact that the Athlon "1600" is not a 1600Mhz chip.
The average consumer(read non-slashdotter) will see the "True" 2000 beside the Athlon "1600" and will obviously go for the higher numbered chip.
Apple has tried to educate the consumer about the reality of clock speed, and they failed. What makes AMD think they can achieve a different result?
I am become Troll, destroyer of threads
I don't think clock speed is important. It should be printed on the computer box someplace but it doesn't need to be part of the marketing or product name.
Clock speed hasn't mattered to me since about 100MHz. Just get a current PC, and your computer will be fast enough for the popular applications (MP3 for instance).
Of course power users will care, but average joe doesn't..it's hard to compare MHz to MHz these days anyway.
If I don't know what the clock speed is on my chip, how am I expected to set the jumpers on my motherboard?
FreeBSD's boot process will still tell how fast :-).
it is clocked
Wouldn't this strategy defeat the purpose of this ruling? Those same questionable vendors can come out of the wordwork, and say that they just sold you a 1.4ghz AMD chip, when in relality, you've just got a 1.2ghz overclocked to 1.4ghz? Without the ability to see both the chip model # *and* the chip speed in the bios, it will be very hard to proof that you get what you ordered.
I agree that stupid consumers are infactuated with high clock speeds that lead to this problem, but AMD chips, from my experience, seem to stand on their own in terms of quality and performance compared to Intel, and need not hide behind this strategy to effectively compete. Besides, if anything, they have to woe the OEMs and not the ones buying speciality-built computers, and last I checked, many of the OEMs are still Intel-based.
"Pinky, you've left the lens cap of your mind on again." - P&TB
"I can see my house from here!" - ST:
...consumers can't get it through their heads that clock speed is not even close to being everything. Intel has proven a willingness to more or less lie about the speed of their processors (got look at some Tbird vs P4 benchmarks and tell me I'm wrong there).
As long as the public continues to see things based solely on the clock speed, AMD can't win unless they:
1.) try to educate consumers better (not gonna happen because cpu design is complex)
2.) fight dirty and do Intel's tricks right back to them.
I'm not too happy about it either, but there's little else AMD can do. At least there's one good thing: it's only a model number. Unlike Intel, they're at least not lying about clock speed.
-Kasreyn
Kasreyn: Cheerfully playing the part of Devil's Advocate to hairtrigger
I can't say that I disagree with their inclination to not want to use MHz, when it doesn't mean anything; even less now that the Pentium 4 is up to 2GHz. Talk about bloat.
I don't agree with their attempt to give out these "model numbers" that looks suspiciously like higher clock rates. Cirix tried the same thing a while back. It only confused people in the end and (IMHO) increased the consumer's reliance on MHz as the single metric on which to base purchasing descisions on.
We really need a good benchmark to comapare these, but that is a very old story....
room101 -- how much can you stand before they break you?
(they always break you eventually)
Arrrrggg, just when I thought AMD would actually have a chance in the marketplace vs. Intel they go and do a stupid thing like this.
::Insert Marketing Stuff::"
This is the tactic of a loser. Look where it got Cyrix. What they *should* be doing is emulating Apple, and run a lot of ads expostulating on the "Myth of the Megahertz". This has the double bonus of getting them airtime and also slamming Intel without mentioning Intel outright (or even *with* mentioning Intel, that's fine). They don't even need to get into technical details, just say stuff like "In the most demanding benchmarks, our processors come out ahead. They are more efficient, and better able to perform the tasks that will launch you into the Internet Era.. etc. etc.
If they want to be seen as a serious competetor in the business arena, this is NOT the tactic to take. Bogus "power ratings" are just that. Bogus. I had just started to genuinely *like* AMD as a company that put out a good, solid product with a minimum of BS. Man, I'm so pissed off about this. Grrr!!
"He's more machine now than man, twisted and evil."
This certainly raises an eyebrow, but I cannot see how it provides any value to me (or the average consumer). It makes things even more mysterious! Given a choice between something that is well known, with a published clock rate, or a 'second tier brand' that hides information, I'd think it would give Intel and even BIGGER advantage. Unless, of course, for some stupid reason, Intel decides to do the same thing.
But really, for the AMD fan, this is an insult. Hopefully their marketing and PR people know some sort of angle to this beyond the obvious that will magically capture market share by removing its Mhz rating.
"Check it out: the new Athlon 1600!"
"Excuse me? Yes, how fast does this processor actually run?"
"It's a 1600!"
"Yes, I know that, but how fast is it? in megahertz?"
"It's equivalent to a Pentium at 1600 Mz."
"Okay, but how fast does it run?"
"I don't understand the question, sir."
"How many megahertz does this processor run at?"
"Perhaps you're not familiar with what we call 'The Megahertz Myth'...."
"I'm thoroughly familiar with it, I've worked in hardware for fifteen years. I just want to know how many megahertz this particular processor runs at."
"It's equivalent to a...."
"No, I don't care about that. What's the clock speed?"
"It's faster than a...."
"That's nice. What's. The. Clock. Speed?"
"Would you like to see some comparisons to...."
"Never mind, I'll just go check out the Motorola booth."
What? That just doesn't make sense. Why wouldn't Intel just go, "oh look, it works for AMD, now let's market our 2GHz PIV's as Pentium 3000s" ????
Why oh why do you think this would cause Intel to work harder on their architecture?
Well at least Microsoft are honest. You definitely do need a 2000MHz CPU to get tolerable performance out of NT5.
-- SIGFPE
If AMD thinks that the BIOS won't be revealing the true speeds of the CPUs, they are on crack. I guarantee you that right after these CPUs hit the market, ABIT will release a BIOS update for all of their mobos supporting the chip. This update will show the true CPU speed, giving ABIT an edge in the overclocking market. To compete, ASUS will do the same with their BIOSs/motherboards, which have a hard time against the cheaper ABIT mobos. After that, EPoX will do it for the value oriented segment of the market.
And then it will end up a standard feature on all the AMD mobos out there....
Or you could have used an oscilliscope. Sheesh, not to metion the fact that different processors process different numbers of instructions per clock cycle.
"Prefiero morir de pie que vivir siempre arrodillado!"
>>After reflecting on this result, we think that >>Intel is using both the falling edge and rising >>edge in an attempt to better market their
>>products.
you're a retard.
You haven't proved anything - you've just proven that the code runs in the same time. Intel MAY be running at twice the MHz of the PPC, but the PPC maybe doing twice the work per clock cycle. Go read an elementary book on CPU architecture. Look for words like 'pipeline', 'cache', 'system bus', and anything other remotely technical to educate yourself.
>>Final Conclusions: After doing some scientific
>> analysis that
funny - you haven't done any analysis, and haven't prooved anything. Please let me know what 'scientific' school you graduated from so i can avoid it like the plague.
This reminds me of an old Kids in the Hall sketch: "Having spent 6 months in the merchant marines, and speaking a little conversational french, I THINK I KNOW A THING OR TWO ABOUT THE RECORDING INDUSTRY..."
You are living proof that a LITTLE knowledge is a dangerous thing.........
Wasn't the 5X86 Cyrix? AMD went from the 486 to the K5 if I recall.
yes we did do science in here. we looked at the physics of waves and waveforms in doing our analysis. perhaps you didnt read the whole post.
The One Rule Of Chess You'll Ever Need: Don't play someone who carries a kit in their bookbag.
The normal user sees clock speed as a measure of performance. Clearly it isn't, and so AMD is moving to model numbers that use higher numbers. Tom says a A1600 is "as fast" as a P4 1.6ghz, however this still relys on clock speed as a measurement!
They need to move away from clock speed and to real world output. I think a good idea would be do name their CPUs after something like the number of FLOPS or MIPS the processor is capable of, much like Apple has done (except that AMD and Intel are both x86 for the sake of this argument, and so it might actually have an effect), unfortunately, neither Apple nor AMD has the market share or reputation to start a new trend, especially since the Intel PR machine has the "clock speed" crown and is likely in no hurry to reveal how weak a P4 has to be in order to reach the higher clock speeds.
--------
It's OK to be social, just don't tell anyone about it.
What are you talking about? The length of the pipeline is irrelevant here. A 2GHz P4 runs at 2GHz. How are Intel "inflating the MHz"?
...and sell them under the name PowerHouse 2200's.
Outdoor digital photography, mostly in New Engl
Actually, you're partly wrong. The ALU in the P4 is double clocked, it really does run at 4GHz in a 2GHz P4.
What I find amusing is that the same people who bitch that we shouldn't judge a processor by its clock speed are the same people who bitch that Intel's processors are slower at a certain clock speed than AMDs.
Who cares? The big question is overall performance. Intel made an architectural choice for the future, not for short-term performance gains. The trade-offs that they have made now are going to allow them to grow to much higher clock speeds in the future while AMD has a harder and harder time of it.
Sometimes it's best to just let stupid people be stupid.
This is not a dirty marketing trick. This is a (admittedly, stupid) counter to Intel's dirty marketing of their bloated speed ratings.
Here is an article on ZDnet discussing the issue. In it, an independent analyst notes that the P4 is 20% less efficient (does 20% less work per clock cycle) than the the P3. This means that MHz comparisons are no longer comparing apples to apples, and therefore meaningless.
As others have said, this obfuscation won't serve AMD in the long run, but they are the "victim" of this marketing war, not the perpetrator. The true victim is Joe consumer, who buys a chip because it has higher MHz, instead of having a metric which actually measures computing power.
Cyrix did the same thing a number of years back.
As a matter of fact, a quick search shows that they got in hot water for this tactic as this Register Article shows.
This makes complete sense and I was wondering when it was going to happen. I would suspect that we will see Intel doing the same.
/. knows which processors to buy, and we don't just pick our processors based on a marketing name. We all look at benchmarks etc and do our comparisons there. The average joe doesn't get it anyway, no mater what the hell the write on the box.
.02.
If you look at the Intel road map, specifically, to the next generation chipset, the IA64, you will see that it is slated to come out at something like 800mhz. No general consumer is going to pay a premium for a 800mhz chip, even though a IA64 at 800mhz will knock the socks off a P4-2ghz.
The consumer has been trained that MHZ are THE measuring stick of processors. As a rule of thumb on like processors that works. IA64 changes all of that, and marketing has to change as well.
I don't know what everyone here is getting all worked up over. Anyone (just about anyone) who reads
That is my
akeRoo
I agree on your first point, which is why neither of them should do it. The fact that AMD did it first is, erm, interesting...
On your second point, if consumers really ever do that, then good luck to them, and may the best chip manufacturer win.
Did you friend with the EE degree ever think to check the clock input going into the chip? Seems to me the world would be astounded to find out that their 1GHz pentium has a 500MHz clock.
I'm sorry.. when the P4 2ghz can't outperform an Athlon 1.3ghz, I don't think that AMD is doing the dirty marketing.
Rating computer systems (specifically) and chips (more generally) by mhz is absurd. There is more to computer speed than mhz, as we've seen by the various cache and bus differences between the chips available (even using the same manufacturer). A 450mhz P2 will frequently outperform a 500mhz Celeron. When you make the jump to the P3, the change actually widens the gap.
AMD moving away from the mhz game is an excellent move, but they really need to come up with some way of letting the public have some idea of how fast their chips are compared to the competition..
Sancho
If you don't like MHz comparisons, why just make up new numbers to compare to MHz ratings? Why not start marketing on a whole new metric, like MIPS or MFLOPS?
If AMD were to start selling processors based on MFLOPS I suspect Intel would have to publish their own numbers. It would be obvious to consumers that the two ratings were not comparable - that is, if you see an ad with a "1200 MHz" machine and a "35 MFLOPS" machine you don't assume the former is 35 times faster.
/* The beatings will continue until morale improves. */
Excellent summary - exactly the way I feel. I was never pulled in by mhz alone anyway...
creation science book
Accordingly, the first chip released under this new nomenclature will be the 1600 MHz "Athlon 2.1GHz." AMD expects sales to improve immediately.
...think like your average consumer. Since if they know ANYTHING about their computers, its the speed, imagine them trying to buy software.
/.'ers here, I'm sure we all have tech-impaired family and friends (like that one who bought Max Payne to run on their 486...you know who they are).
"Okay sir, and how fast is your computer?"
"Its an AMD 1600"
"So...how fast is it?"
"Its an AMD 1600"
"Do you know how fast it is, in MHz?"
".......Its an AMD 1600"
The average consumer will now know even less. And while that might not mean much to the
Sure, it might be what AMD needs to compete with Intel's ads, but they should just launch their own ad campaign showing how the 1.4GHz Athalon performs just as well, or better than the new 2GHz Pentium IV in almost every non-SSE-related benchmark.
Bingo! Couldn't have said that any better myself. On a side note about the reason for removing the Mhz from their products, take this into consideration.
AMD's Athlon has always been notorious for smoking identically clocked Intel processors. PII, PIII, P4, it doesn't matter. The Athlon always outperforms in both integer math and floating point. So, has anyone ever considered that AMD's processors just do the same instructions in less clock cycles? I remember seeing old intel specs that stated integer adds took 4 clock cycles to complete. (I'm sure it's outdated by now, but the concept is still there.) These aren't RISC cpu's where a single clock means a single operation has been completed, these processors end up waiting between 1-50 clock cycles for every instruction they perform. Obviously the higher delayed instructions are the more complex (like MMX and SIMD), but by simply making common operations take less time, you'll get better performance.
I've heard that AMD is going to seriously start concentrating on this aspect by making instructions take less clock cycles, rather than playing the die size wars with Intel. If this is the case, I can understand how the marketing guys would want to do something to disassociate their chip's performance from the fancy number that the competition is going to be steadily raising in the meantime.
So before you cry about AMD selling out, think about the differences between their CPUs and Intel's CPUs, and what they have to lose if they don't drop the Mhz listings.
I'm confident that this marketing strategy will only be in affect for a year or so at the most. Just long enough for people to realize that Mhz isn't everything. Once Intel raises the bar a few more Ghz, and AMD is still right there keeping up if not beating them, they'll release their current CPU speeds and you'll all be awestruck how a 2Ghz cpu could possible put a 5Ghz cpu to shame.
I also did some scientific research. I wrote a similar program, and printed it out from both a 1 GHz Pentium III and a 500 Mhz G4. I then threw both copies of the printed program out a window. I was surprised to find that they landed at approximately the same time. This just goes to show that Intel is obviously counting its clocks TWICE instead of ONCE. I think this is backed up even further by the fact that the copy of the program printed out from the Pentium machine, fluttered as it fell a lot more than the G4 copy.
It's 10 PM. Do you know if you're un-American?
Its creepy that Intel seems like the honest platform now. A P4 2000 is actualy 2GHz. PC800 RDRAM is actually 800Mhz. But calling an Athlon 1.4 an Athlon 1600 is almost as bad as calling 266 DDR SDRAM PC2100!
A deep unwavering belief is a sure sign you're missing something...
OK, but if I can't compare the clock speed of an Athlon vs. a P4, I also can't compare two different Athlons. How do I know whether I should be model 1600 or model 1800? It would sound like saying Windows 2000 is better than Windows 98 because the number is higher. Not telling how to compare two of your products, will likely decrease your sale.
Also, now Intel can say: "Our latest P4 beats the crap out of an Athlon XYZ" and people won't know that they compared it the the slowest model.
Opus: the Swiss army knife of audio codec
And, you know what? Within a week, we all sigh with relief, because the old units never worked anyway!
When was the last time you heard the MIPS or FLOPS rating for a processor? When the RISC processors came out, and scored 100 x the nearest CISC chip, we suddenly started hearing how worthless those ratings really were. (Which was true, only the people saying it had been using them to crush the competition under their feet, the previous week.)
What's the FLOPS rating for a Pentium IV? Anyone seen it listed on any of Intel's adverts? Curious, that.
Truth is, there -is- no meaningful number you can use, to describe a processor. Applications will vary so much in performance, depending on how well they exploit the various caches and pipelines, that any value you get will be useless for any realistic comparison.
Worse, the bottlenecks for the main memory, the PCI bus, any local busses, etc, ad nausium, are so much more significant than the processor. Sure, building a faster chip will earn lots of green bits of paper, whereas building a better motherboard will simply earn lots of whining from hardware manufacturers.
The reality is, though, that processors today would be perfectly adequate, if the support hardware were up to scratch. (Anyone remember the problems the 486DX-50's caused? Those worked at 50 MHz, direct. Great design, but the hardware needed to run it killed it. The 486DX2-66 was really just a DX-33 with some fancy over-clocking. The support hardware was all standard stuff. That's why it caught on.)
It's time to take another look at that hardware, though. I doubt it's changed much since the DX-33 days, except with a few extra levels of caching. It's still convection-cooled, for the most part. The connectors are still badly designed and cheaply made. Sockets are built to be easy for plebs, not easy on components.
Compare this with a VME or VMX bus, where the backplane alone costs more than most top-end PCs and where ease-of-use can go jump in a lake. These are systems where customers can afford to pay, and don't want to pay for junk.
I'm not saying PC manufacturers should suddenly switch over to VMX-style architecture (128-bit busses can get a little interesting, and besides, I've some PCI cards I'd like to keep using!), but it's time to do some re-designing. If a user wants to be babied, they're not going to handle hardware installation, anyway. They're going to go to a shop. Providing idiot-proof systems is simply driving up the number of idiots and driving down the performance of computers.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
IBM refers to their own mainframes's performance using an obscure value like 'Relative performance units'. They use an arbitrary baseline from one of their own models - call that 'Relative performance 1.0' and then proceed to not only NEVER publish any other vendors numbers but tell you that any other vendor's machines are not comparable and that no other benchmark can be compared or correlated. In a similar vein if you use some of the Lotus benchmarking tools for Notes and publish the results they can sue you.
The last PC I put together was AMD based. Before that, I had always gone Intel. What moved me over was the fact that I could get more performance for my dollar. As long as the processor is powerful enough to do what I want, I don't really care about the clock-speed. Yeah, having a more clock speed can bring "bragging rights", but every day we are shown that clockspeed is almost meaningless in terms of performance.
;)
In my mind, however, hiding the clock-speed rating is equivlant to hiding the version number on software. It's no longer Windows 6.0, it's Windows XP, or 2000, or Windows "The Version that Makes Windows Good(tm)."
This whole processor coverup thing started with Intel and their "Pentium" series. It does make business sense, but it can tick off tech-savvy people. Why? The average consumer thinks "Processor" and not "80586 200MHz CODENAME CPU". Consumer understand brand names, and brand names help companies develp identies and products. That is why it is now Windows 2000 and Windows XP: it creates a sub-brand of the real product.
Think about this: Windows NT 5.0 and Windows NT 6.0 versus Windows 2000 and Windows eXPerience. The version numbers make it sound like a simple "upgrade" while the brand name make them seem like completly seperate products. It may be just enough to convince people that it is a world of change, regardless of what is actually in the box.
Back to processors, I think AMD is going to try to make some brands - focus on the name and image and push aside the gritty technical details. IE. "The AMD WhizBang(tm) processor is as powerful as the Intel Pentium 4 2000MHz." It's marketing... pure and simple.
I don't think it will matter what they call it or how fast the CPU runs. Independant benchmarks will show the true performance of the processors. This could be a good thing in that it may get ordinary consumers to become more informed about speed vs performance. IMHO, an informed consumer is much better than one that simply buys the one with the bigger MHz rating.
Price, Quality, Time. Pick none. What, you thought you had a choice?
Salesmen
Flat out. That's the problem with computers these days is that the salesmen don't know the inside of the computer from the outside.
Example: Back when Intel started making S370 Celerons, I went and asked a computer store clerk if it was Slot1 or S370. He said it was slot1. "All the older socket processors are too out of date and just don't perform as well." After going back home and looking on the internet to find out that the computer was a S370 rather than Slot1, I didn't trust that salesman again.
I figgured I'd try again last week with the Pentium IVs. I found another sales clerk and acted like a potential college student needing a computer for college. I asked him which was the fastest processor. "Oh, hands down, the Pentium IV! I mean, they just released a 1.8GHz chip, when all AMD has is a 1.3 GHz chip." I figured I'd play this out..."But is it worth the money? I mean, that Athlon system is $400 cheaper!" His response? "Well, if you need the cheaper system for college, pick the Athlon. But if you really want those games to shine, pick the Pentium IV. All that money is for the faster processor and faster memory." I just had to get out of there before I blew my top over his faster memory claim with RAMBUS.
Look at it this way: If you're ever gonna go out and buy a car, look over the lots to see what you like and what looks nice. But for crying out loud, NEVER take for granted what the dealers say, because they're out there to sell. If you want to know how the things honestly perform, find someone who already owns one and ask them! Or go to your local mechanic (everyone should have one, just like everyone should have a neighborhood geek whenever they need help with their computer) and ask them what they think about that specific model car.
That's why Cyrix and Intel both have to crank out these pathetic "P" ratings in order to satisfy market competition. The people who sell the products in the stores have no other choice.
When I read this earlier on ZDNet, "industry experts" were saying that AMD couldn't get consumers to abandon MHz ratings for instructions per clock cycle. Personally, I don't see a problem with them trying to change comparison factors as long as the numbers are meaningful, which 1600 doesn't sound like it is.
I'm sure their marketing team could come up with something like 1.9 giga-doodles for a 1.4 MHz cpu. Obviously something a little more sexy would be needed though.
If AMD changed schemes, then what geek here would not buy them because of it? We know what they're referring to. But I guess the average consumer couldn't compare giga-doodles to GHz on their own. But AMDs marketing could again jump in with stickers & posters for retail stores and OEMs. Something that specifically states what the giga-doodles of this AMD is vs. the giga-doodles of similarly priced P4. That's definitely not illegal and would be better received by the geeky population at least.
The worst thing for AMD would be for Intel to start spinning it as:
"They can't keep up, so they HAVE to change the numbering. We're actually leading the way."
as opposed to AMD's take of:
"We want you to know that our chip X is comprable to intel chip Y."
This space for rent. All reasonable inquiries will be entertained at proprietors discretion.
...why not just use SPECfp/SPECint or some other established performance-rating criteria?
Exactly, in the same way that (in some places anyway) you pay for natural gas by the number of megajoules rather than the volume, because the volume can vary and be misleading.
What part of what I said are you disagreeing with?
I beleive what I told Hard_Code could also apply here:
thank god there are people like you out there. as you can see, i am getting a lot crap about this. my hope was for people (intelligent people) to think objecctively and possibly do their own research and see for themselves how some processors act.
The One Rule Of Chess You'll Ever Need: Don't play someone who carries a kit in their bookbag.
Your posting, while well thought-out, is technically nonsense.
I write this as a person with a bachelors in Computer Engineering who is currently completing masters in EE.
First of all, the waveform in question is a square wave, not a sine wave. So I don't see how pi comes into...well, anything. You go though some basic trig to prove that f=c/L (frequency=speed of light divided by wavelength). So what?
Some architecture triggers on the rising edge, some on the falling. It does not really matter.
The rumor you heard about Intel architecture "counting" both rising and falling edges is silly; what counts is the number of pulses, not the number of rising and falling edges.
Now, there may be a basis to that rumor in that some architectures where the CPU runs at a multiple of the bus speed and triggers on both the rising and falling edges. The older Athlons, for example, run at a 200Mhz clock speed. But the external CPU bus runs at a 100Mhz clock speed.
Does this mean that AMD is cheating? That they are "claiming" 200Mhz when it is only 100Mhz?
No. What it means is that the Athlon triggers on the rising edge, then half-a-period-later it triggers again on the falling edge. Assuming that the Athlon triggers on a rising edge, this could be accomplished by inverting the clock, and ORing the signals together (although it is not that simple, you get the idea). So for each external 100Mhz clock pulse, the CPU fires two internal clock pulses. And the speed is doubled. So your 1Ghz Athlon runs externally at 100Mhz with a 5x multiplier. Inside it runs at 200Mhz with a 5x multiplier. 200x5=1000. See?
And since you trigger every half-period, you cut the time of the period in half. f=1/T, where f is frequency and T is period. So when you cut the period in half, you double the frequency.
That is why the new 266Mhz FSB Athlon chips need to have the external clock speed set at 133.
So why not just run the PC board at 200Mhz and forget all this silly clock-doubling hardware? It is not that easy. Desinging a glass-epoxy PC board to work at 100Mhz is hard; 200Mhz even harder. As you go higher in speed, harmonics in the microwave regions begin to creep in and most digital designers are not ready for that sort of variable. Plus, it raises the cost of everything in the PC. Remember that your PCI ports still run at 33Mhz on most machines....
Okay, so why is the Mac faster at the operations you used in your tests? It's a different architecture! You are comparing apples (pardon the pun) to grapefruit here. It's like saying that if a 10-cylinder diesel truck is at 4000RPM and a 2-cylinder moped is at 4000RPM, they should be going the same speed.
How many CPU cycles does each operation take on the G4? How many CPU cycles do those same operations take on the Intel? What about differences due to setup and OS lag? Is the compiler optimized for the CPU? If so, is it using out-of-order execution? That is the sort of thing you need to know for a test like this. The same operation may take 10 cycles on the Intel and 1 on the G4. So, for that operation, the G4 would be ten times faster. If an operation takes one cycle on both machines, the Intel would be twice as fast as it has twice as many cycles per second. Cycles Per Second, or CPS, is also known as Hz. And the Intel chip is running at twice the clock rate as the G4. Trust me. What it is not running at is twice the speed, since operations on the two machines take different numbers of clock cycles to complete.
the MHZ of a processor means absolutely squat.
It's been rehashed time and time again, and silly enough people keep clinging onto the MHZ speed as a performance rating. It means nothing and indicates nothing. everyone with a clue knows this and everyone that ever owned a cyrix 586 or 686 processor really knows this. you'll never get real ratings out there (print the mips and mflops on the chips!!!!) but even then that means nothing with the addition of huge pipelines and multiple pipelines.
I say just market as follows.
Athalon 4.2 - It's 4.2 times faster than the Pentium 4 (or whatever)
Do not look at laser with remaining good eye.
I agree with you completely, and I know exactly how the 2 CPUs differ, thankyou. My point was that the Intel chip runs at 2GHz, which is exactly what Intel says it runs at... It's not inflated in the sense that the parent poster was trying to imply, and certainly not in the way that a 1.4GHz processor is to be marketed as a "1600". Yes, technically they didn't say it was anything to do with the clock speed, but it still seems a bit dishonest.
It's meant to provoke thought in the realm of making decisions on what chips are faster.
I use the analogy from the sine wave (which does have a wavelength of pi) to demontrate the definition of frequency in a waveform. Compare that to squared off circuit waveform of the clock cycle, and you'll see my argument.
I also suggest that Intel is doubling the sped on paper. A 2 GHz chip is really just a 1 Ghz under the definition of frequency. Intel decided to count both the rising and falling edges in an attempt to "double" the speed of their chip.
Perform benchmark tests of your own. Another guy says he's done that and got similar results as my tests. You'll then see why Apple and AMD are attacking Intel's marketing of only using the clock speed.
The One Rule Of Chess You'll Ever Need: Don't play someone who carries a kit in their bookbag.
So, you would define smart as admitting defeat to the competition where you were once ready to pass in the race?
.6 bigger than AMD's number.
Here's the problem: AMD's been whoopin Intel in the benchmarks, so Intel fires back by overrating it's processors. It's so dumb, it works, because consumers are dumb. So, even though Intel's P4 machines cost $400 more to make just to get the same performance, they're attracting customers who take that number and sleep soundly with it at night knowing it's
AMD's own problem is that by doing this, they admit defeat to Intel and say, "Even though we want to get out of your shadow, we're going to play the game by your rules, because you still are the market leader."
Tell me then, what's going to happen when the 64-bit processors start coming out? Is AMD going to I-rate (Itanium-rate) their processors, even though the Itanium is a newly-designed processor that only emulates x86 instructions while the Sledgehammer computates them directly? That's just going to make us even more i-rate (excuse the pun).
AMD has the money and the processor to get themselves out of Intel's shadow. But so far, I have only seen one real commercial that AMD set forth (two years ago) to punch the Athlon hard. It only ran for about a month, and made them look wierd (is technology supposed to be about a gameshow where if you lose, you get hit by a train?). If AMD wants to remain in Intel's shadow, they should go ahead and follow a P-rating system. But if they truly want to become their own company and actually compete, they need drop this charade and start promoting how well their processors do, rather than what they're rated at.
It would show high-level results of independent testing of appropriately chosed AMD and Intel based systems (and not a whole raft full -- just one or two popular sets). With a couple of graphs and summary numbers, the consumer would see that an Athlon 1.4GHz performs comparably to or better than an Intel 1.8GHz, or whatever.
Don't overwhelm them. 1 page of reasonably large text and pictures should suffice.
The harder part (once control has been wrested from the brainless marketers) would be to get the sheet into salespersons, saleswebs, etc. hands for presentation.
Customers, even non-technical ones, are very adroit to smelling BS, and very adverse to it. I think an honest approach might work better than the proposed obfuscation.
Why don't chips compete on power consumption and battery life?
I think we can all agree that the latest and greatest chips are grossly overpowered for the average consumer, even the average gamer.
So in this age of power crises in California, why not sell laptops or desktops that are smaller and consume less power? I personally want a laptop that will run eight to ten hours on a battery.
Right now, I have a ThinkPad 570 that has every feature I want. It has a Pentium II Mobile at 366 Mhz. I can watch DVDs (granted, I have a hardware decoder PCMCIA card), browse the web, check email, even play games (Fallout Tactics) and I have no complaints at all. Battery life is two to three hours, depending on what I'm doing.
Meanwhile, Intel and AMD are releasing gigahertz processors for laptops. Why? Laptops are not gaming machines. Laptops are for a portable office. Most usage is email, word processing and internet access. By designing what is now a Pentium III 1.13 Ghz to instead be 500 Mhz, you could save money and power (while still making use of the SpeedStep features to further reduce clock cycles while on battery).
Truly "on the go" laptops could be smaller and lighter with longer run times. High end "desktop replacement" laptops could still use the full speed processors and the powerhouse video cards which spank my Voodoo 3.
Desktops could likewise be smaller, using the same features. Most desktops are available with build-in everything, so expansion bays/slots could be kept to a minimum.
Another advantage of this is that one could create silent computers, similar to the Apple G4 Cube. Less heat generation means less fans and that means silence.
Those who want to overclock are going to buy the high end processors anyways. But those building an MP3 server/player to integrate with their TV/stereo are not going to need a 2 Ghz processor. A 500 Mhz Pentium III (0.13 micron process) would simply need a heatsink and some airflow.
I welcome the day when megahertz is something you need to look to the "technical specs" page (and I mean technical).
www:~ > cat /proc/cpuinfo | grep MHz
cpu MHz : 1009.003
Gee, whatever will I do? :)
This whole discussion boils down to 2 points:
1) Hiding the Mhz from the masses is good
2) Misleading people about clock speed is bad
So why name a 1400Mhz PC as a 1600? That sounds like "lying" about the clock speed. Instead, name it an Athlon 6000? Name the 1500Mhz part Athlon 6500. That way, no one will make the "Mhz equivalency" mistake that hurt Cyrix, but the frequency is still hidden.
Who here bought an HP 600 or a Canon 720? No one, because manufacturers never made the mistake of naming printers by DPI. But I bet some people have an HP 624C.
The best solution would be a standards body, started by a tech reviewer, (like Tom's hardware or Anandtech) to assign each chip maybe 3 numbers that indicate it's performance in 3 key areas. Perhaps applications, games, and server. Then the consumer can easily browse the shelves looking at whichever number best applies to them. If the rating is independant, then we don't care if it is proportional to the Mhz or what, it is a valid usable measure for the consumer. Isn't that what we want?
I just assumed that "Athlon 1600" meant 1.6 Ghz. I like AMD's idea of ditching clock speed since it's irrelevant. I don't agree with a model number that closely resembles a clock speed. There are morons who work at Best Buy and other computer chains that are going to tell consumers that the machine is 1.6 Ghz. Why? Because it just makes sense and they don't know any better than the consumer. Then Best Buy will get sued for false advertising when someone figures it out. I would personally like to see model numbers like "A0108" (Athlon released Aug 2001). Any guess that it represents a 1.4 Ghz chip? Not really.
There is no reasonable defense against an idiot with an agenda
:wq
Let's see numbers that really can be compared across different platforms, such as:
How many minutes to compile these particular 100,000 lines of C++ code?
How many gigaflops?
That that is is that that that that is not is not.
I remember very clearly that a good number of people were angry with Cyrix when they found out that their 5x86 200 wasn't running at 200 MHz -- rather it only scored the same as a Pentium 200 in some benchmark. I'd think a company like AMD would remember that, but it looks like the old "those who forget the past are doomed to repeat it" adage is correct.
I still love AMD's product, but this new marketing is a shame, imo.
The only way to make it all true and fair would be for the manufacturers to start marking the chips with a number that more accurately identifies the actual capabilities of the chip.
The two most common things in the Universe are hydrogen and stupidity. -- Harlan Ellison
I'm sure AMD will release the clock speeds in the technical documentation. They have nothing to hide from the people who are aware clock speed isn't the only factor that determines how fast a processor can crunch numbers. And if they don't, they'll lose the power-users who currently flock to AMD in droves, which would bring them back to where they started.
Most of the root posts I'm reading here essentially say the following:
I guess my subject line is a really nice way of saying you guys don't get out much. Who the hell are you kidding?
The reason that there are marketing departments is because people are sheep. Every huge corporation in the world depends on it. If people weren't sheep, they wouldn't pull all of the shit that they do. Why do you think Intel has been trumpeting their '2 GHz' speed capability so loud, despite the fact that core clock is becoming extremely distant from the actual throughput of a processor?
Answer: people are dumb.
Be happy that you're actually sitting there thinking about the issue, regardless of which 'side' you are on, because most everyone else isn't.
Rock
By implying that it is more important than it actually is.
By implying that a 1.6GHz PIV is faster than a 1.4GHz Athlon.
Yes, the nick is flamebait
Grapefruit?? Don't you mean lemons? :-)
Now, let's do the same thing with CPI. Instead of "Megahertz GOOD!", let's all stomp our feet and say, "CPI BAD!" I'm thinking of that metallica parody here. Anyway, people understand golf scores, where lower is better -- they can be made to understand that lower CPI is better. So why doesn't AMD come out with an ad campaign saying, "The pentium 4's average CPI is 97, and ours is just 2. Therefore, our chip is FIVE TIMES as fast as a p4 at the same clock rate!!"
The problem with this is twofold.
Firstly, anyone remotely sane will pick out the "at the same clock rate" line and be suspicious. Machines with higher CPI ratings tend to have higher clock rates to compensate.
Secondly, CPI for _throughput_ is likely to be in the 0.5-1 range for almost all systems, due to pipelining (issue 1-2 instructions every clock on average, and no matter how long they stay in the pipe before being retired, your throughput is 1-2 IPC). In practice, stalls would kill this for really long pipes, but there will *always* be benchmarks that perform well. I've had to benchmark this kind of thing. You wind up with numbers all over the map.
In summary, I think the only benchmarks that will make sense will be those that more-or-less accurately represent the real workload the machines are going to be exposed to (gaming benchmarks for the gamers, office suite benchmarks for office workers, etc.). It isn't a surprise that these are the best benchmarks, but with the architectures being compared diverging, it looks like they're the _only_ valid benchmarks.
60GXP is a model line, not a model by itself. Yours would be a 20GB 60GXP, as opposed to a 30 or 45.
Nicotine free Amish .sig.
OK I'm still not making myself clear. There is no dishonesty in Intel's position here. They made their pipeline longer for technical reasons (which many people disagree with, whatever, that's not my point). They are not jumping around saying, "hey look our pipeline is longer than AMD's", are they? No. As a result of having a longer pipeline they can run at a higher clock speed. No, it doesn't make it run any faster than the AMD method (at least in the short term). But they never said it would. You just thought you heard them say it.
Why the hell should Intel use a 4 stage pipeline just because you think that's somehow more "honest"? Dude, they're trying to make fast chips! Let them make them how they want!
I think a few people have bought an HP 600
;)
-Shieldwolf
just = (My)Opinion.toCents();
Ok, some software like 3D rendering does have performance specs, but the vast majority of software has no attention paid to performance, by consumer or programmers.
PJRC: Electronic Projects, 8051 Microcontroller Tools
The clock speed DOESN'T have any direct bearing on the system's performance. What's dishonest about this? I mean, the Mustang 5.0 always had a 4.7 liter engine...what's the big deal?
Why yes, I AM a rocket scientist!
This must mean the Athlon 1600 just can't compete with the awesome power of the Atari 2600!
I'm headed for E-Bay right now!
---
You'd be surprised at the broadband connection available to things crawling around in your hair.
I won't be selecting "AMD" on 's CPU Type select box.
-- @rjamestaylor on Ello
I'm not so sure that you could call this a "dirty trick". The MHz myth was expounded on here in /. recently. It'll be a nice change of pace to see chips benchmarked by true performance rather than artificial measurments.
But that's just my opinion.
--
If I actually could spell I'd have spelled it right in the first place.
The standard is not CPI, it's MIPS: millions of instructions per second.
For people used to comparing Intel with SPARC, ALpha, PPC and other similar chips, the standard measures for CPU performance are MIPS and FLOPS (floating Operations per sec).
Essentially, one "MIPS" (or MIP) is 1,000,000 instructions per second. If you have a 1,000 MHz chip that does 4 instructions per cycle, then you get 4,000 MIPS.
Another chip that also does 4 instructions per cycle (.25 CPI) but that opeates at 200MHz only accomplishes 800 MIPS.
So you can't rely on Cycles Per Instruction as the only measure of performance. Rely on MIPS.
Why do we always see comparisons of 2.0ghz P4's vs 1.4ghz Athlons? Even a 1.4ghz P4 vs. 1.4ghz Athlon doesn't make much sense.
I would rather see a comparisons of CPU's at the same price point. "Intel's $150 CPU vs. AMD's $150 CPU".
Of course, maybe you would need to include the cost of RAM in that evaluation as well, since the P4 (for now) uses the vastly more expensive and stupid RDRAM. So I guess an even better comparison would be, "a $1200 Athlon system vs a $1200 P4 system". Of course then there's debates on how to equip the systems... the AMD CPU+compatible RAM would be like $300 cheaper, so where do you spend the xtra cash? More RAM? Better vid card? Still, although less precise, it's more of a real-world comparison.
OtakuBooty.com: Smart, funny, sexy nerds.
Well, anyone who really wants to know will know where to look for the true MHz numbers, maybe he'll also know what these numbers are telling him. Since the advent of the P4 comparing MHz numbers for different processor types has become utterly useless, and even before it was just good enough to tell if an Athlon and a P3 played in the same league. Anyone who wanted to know about the performance for a specific app looked it up, and i don't bother much about the types who have too much money to throw it after a 2GHz P4. Most of those would probably be served better with the framerate of q3demo for a reference, but honestly it's not my problem.
Nevertheless the MHz Hype never ended, first AMD used it when they crossed 1GHz, now intel uses it shamelessly with the differently designed P4, not bothering people too much with too many gooey details like pipeline length. Well, still MHz sells, and probably the shops will find a way to clue the customers up about that number if it helps sell.
The sad part about it is, that the MHz number hade quite a little use, for example to compare two Athlons only differing in MHz-rate, knowing that the last 100 MHz gave you a 4% performance boost for a specific ppp, you can't hope for more for the next 100 MHz upgrade. Also it was quite helpful to judge the difference between say a 100 and a 133 FSB, with the same MHz-rate. But as i said, the numbers will be there for those who want to know em.
"By the way if anyone here is in advertising or marketing... kill yourself." -- Bill Hicks
Intel got out of the MHz business a long time ago when it introduced the iCOMP to show its chips growing performance linearly even though clock speeds weren't keeping pace (we're talking way back in the 50-MHz pentium days when AMD wasn't an issue and Intel was just trying to keep the press from spanking it).
iCOMP is a combination of several benchmarks and some hincky math (I've seen the formula--eeugh). Its saving grace is that it's applied uniformly to all parts. Marketing weasels can't get around standardization.
The fact is, one benchmark, be it "how fast does it tick?" to "how long does it take to decode the human genome?" is a worthless way to pick a computer. Look at all the benchmarks you can find. Understand the systemic issues in each test setup. Evaluate the use cases you will put to the computer. And if someone waves a benchmark at you and says "bias! this has bias!" say, "of course, it's biased towards the processor that runs that software faster."
--Blair
"If you don't run the software faster, you lose."
Apple is doing this now -they face dwindling market holding because they are oft-seen as the easy-to-use-but lacking power solution. Not so. Truth be told, Apple could blow away any PC proc based on a purely MHz rating too - this from a die-hard PC user - but this is not the forum for that debate...
Apple has taken the high road. They have begun to educate users en-mass about the problems of relying on the MHz rating. Sure we know better: Intel cannot benchmark equal to AMD on a MHz rating because they cannot run the same number of operations per second. Thats simple math.
The analogy I like is who has more light: if everyone 100 light bulbs but all mine are 100 watt and everyone elses are 60 watt, everybody can see that the 100 watt bulbs are going to produce more light, but it still seems like everyone is comparing the number of bulbs - "Its got to be brighter becasue they have more bulbs!"
AMD has gone the opposite way in the analog, like saying we're giving you 60 bulbs but the amount of light will be the same as intel's 100 bulbs. And most people are still stuck saying "ya but you're ripping me off for 40 light bulbs!" AMD needs to take a better look at how the big picture will appear to the public, are they looking for more light, or more bulbs?
"Quando Omni Flunkus Moritati" -- Red Green
Vintage computer games and RPG books available. Email me if you're interested.
Oh bugger. You meant the code the benchmarks rates the processor at, not the code the benchmark uses. I'm going back to the corner now.
Vintage computer games and RPG books available. Email me if you're interested.
Seems that this tactic has already proven it's ineffectiveness with Cyrix's PR ratings.
RFC2119
I guess you missed the whole point of my posting. Intel is not "doubling the speed on paper". The error on your benchmarks is in your attempt to compare CPUs of totally different architecture and instruction set.
The Intel CPU is twice as fast. I agree; it is very foolish to judge a CPU by its clock speed.
But that is what you are doing! You say that since a 1Ghz Intel runs a program in the same time as a 500Mhz Apple, the Intel must be running at 500Mhz!
The Intel is running at twice the speed, but on that architecture your program takes twice as long (in clock-cycles) to run! What is the problem here???
What we need is not to keep going on with this CPU Speed game of the fastest chips. What we need to do it get the CPU companys to work on increasing bus speeds of motherboards and all ports and buses. In theory if you have a computer with every bus at the same speed the computer would have no wait time. The little hourglass icon could be deleted off of your system cause there would be no waiting. That is what cause slow down today is the bus speeds not the CPU. If you are running a P4 2Ghz of AMD 1Ghz there is not much different both are powerfull enough to kill most anything app you try to run. if you drop the CPU to 600Mhz and increase all buses to 600Mhz that would kill any of todays processor. Tell me what you think on this subject
That is my thought. I could be wrong; gostf
Better make that A200108. What, were you born yesterday?
If was doing marketing for AMD, I'd stress this to hell and back. Show how Intel has "a lot of gigahertz" but not much "performance" or "power". Make it look like Intel was somehow trying to deceive you by quoting clock speed rather than how fast your applications run. Then show the prices for equivalently powerful chips. Maybe throw in a quip or two about whether you want to buy the name "Pentium" or buy computing power.
Do I get the job?
sub f{($f)=@_;print"$f(q{$f});";}f(q{sub f{($f)=@_;print"$f(q{$f});";}f});
Technically, Specfp/SpecInt measure the perormance of a particular system, running a particular OS... The CPU is often handicapped by the choice of motherboard, for instance. Although it might be useful for a typical computer shopper to see a standardized benchmark, most people are not interested in Quantum chromodynamics or computer chess
not really 'instructions per second'. folks want to know how much faster is it, given a fixed cost constraint, over the competition.
if you can say 'the equivalently-priced P4 to our K7 takes twice as long to scan and save your photos', that is a clear communication to the buyer.
only techies care about instructions/second and pipeline depth, and so on.. users can relate more to how much they have to pay on one brand vs. another to get the same performance.
measure performance in common terms, like the photo example I used. address things like: working on large spreadsheets, doing audio/video, doing math (and all that multi-thread processing used by the latest net.virus) - really just doing anything that's heavy on cpu can be used as a way to relate how much you have to spend to get that same job done in the same amount of time.
--
"It is now safe to switch off your computer."
Please tell me your are kidding about the above.
./ reader pound every tiny possible interpretation of a posting into the ground?
./ crowd seems to get off from picking at every single byte on a post: The details here were used as an example. This author assumed that the Athlon's logic triggers on a rising edge for purposes of this discussion. It may be a falling edge; please invert all logic cases if this is true. There may be minor errors in spelling and punctuation or grammer in this letter. There may be a typo or a mistake of another sort. Any specific examples used in this post are intended only as reference to the post, and not to apply to the world, internal combustion engines, the mating habits of Cyprinodon diabolis or anything else.
That's good; start a post with an insult.
\begin{editorial}
I am willing to bet you are one of those software-types who thinks he is an expert at everything and all people who don't agree with your opinion on the language, OS and architecture that is in fashion are idiots who should be banned from the use of a keyboard. I've seen it all before; I remember when CS people laughed at C; Pascal was the language of the future and you would run it on a real computer, a VAX. "Some decent CISC architecture for real programmers who know what they are doing."
That same crowd now laughs at the VAX and went from Pascal to C to C++ and now to Java. I'll step out of the way when the tide changes and they all start spouting that LOGO is the only way to program.
\end{editorial}
If you had put your ego in park and read my message carefully...perhaps giving me the benefit of the doubt rather than jump on every pedantic
point you could misinterpret, you perhaps could have seen what I was discussing.
Nope they don't. The slowest Athlon was 500Mhz.
I was referring to the speed of the FSB, as should have been obvious. Why must every
But it transfers 2 bits per cycle, making the effective data rate 200 MHz.
If I was as pedantic as yourself, I would probably complain here that "you must be some sort of idiot to think there are only two bits transferred per clock cycle" but I know what you are trying to say...however your "translation into writing sucked".
Yes, the transfer rate of the bus between northbridge and the Athlon is externally a 100Mhz clock bus with a data transfer on both clock edges...but that is a kludge to avoid having a real 200Mhz bus outside the chip because of the reasons I explained in my posting. But internally (that is, inside the IC) the FSB is multiplied to 200Mhz and the data appears at the normal points on the clock cycle.
The Athlon has used the both-edge kludge on the external bus since the first chips. In fact, on the module-type chips the L2 cache only operated at 200Mhz/2 or 100Mhz (in other words, half of the core CPU speed).
But at some point the data from the external bus has to actually be used. Do you think it is carried all through the chip at 100Mhz, even though the core speed is 200Mhz? What, do you propose they multiply it up in some sections while dividing it down in others?
The clock is multiplied as I described. This syncs up a rising edge to the presence of the data on both the rising and falling edge of the clock cycle. The CPU can then grab the data where
it expects it on the trigger from the clock.
Okay, now some disclaimers since the
Some people have used car analogies (the Mustang 5.0 with a 4.7 litre engine), but no one has made the key observation.
Chips are being sold based on the tachometer scale. The P4 revs at 2GHz. The Athlon revs at 1.4GHz. Most of the buying public thinks that the tachometer is actually the speedometer.
So now AMD is painting different lines on the tachometer. They are writing 1600 where 1400 used to be etched. It's going to confuse the mechanics all to hell without doing anything about the ignorance of the buying public.
The whole thing makes AMD look cheap.
How do they go about conveying the message that the Athlon 1400 is a six cylinder engine versus the P4 which is a four cylinder engine?
I remember reading biker magazines about ten years ago. In North America, you had 600cc bikes which redlined in the 16,000 RPM range. In Tokyo, 600cc bikes were not street legal. The biggest bike allowed was in the 250cc range IRRC. So in Tokyo you had these 250cc bikes that redlined at 28,000 RPM. You had to pull 10,000 just to cross an intersection. Isn't that a perfect description of the Pentium 4?
Back when all the other chip cloners were playing games with their product naming, I never bought any of those chips. I feel the same way about rebate stickers: the amount of mental effort required to read the fine print exceeds the expected return from redeeming the coupon (should you be so lucky that a cheque ever arrives). I tend to make my life simpler by looking at only those products "gimmick not included". If the sticker makes me work harder to decide what I'm buying, I don't buy it.
I think AMD is exchanging one bottle of single malt (customers who are influential and know what they are doing) for one Joe-sixpack of American beer. They are heading out into territory that will be fought on marketting terms rather than on technical terms. Good luck to them.
I remember another company that took this route: Gateway 2000. They started off with unbeatable price/performance, and then they veered into big screen TVs and five-disc CD changers. Look at Gateway 2000 now.
How long will it be before the AMD PR2100 and the AMD PR2200 are just the same chip in a different die color?
Ah but which one is better at mixing cement?
Yep, I've used an Apple. It was ten years ago, but I've used an Apple.
I think you mean a Mac, though, since only Macs have the little bomb you describe. All I have to say concerning that is if you're having problems then your configuration is having problems because my home computer is perfectly stable these days, be I in 9 or 10.
And I've had to deal with extension conflicts, but I've also had to deal with a corrupt boot partition on Linux and a corrupt registry on Windows. All seem to be around the same level of fun.
You're right, though, it's not because they were out-marketed in the chip department. It's because they were not made the standard by IBM, which people seem to underestimate the power of, even today. IBM made the PC *as we know it* and that's what stuck. Never mind there were PCs before it, it doesn't matter. IBM made theirs and that was the official first coming and what people had to match. So here we are. The better platform still (miraculously) idling by with near 10% retail marketshare at the moment and 5-7% installed base (IIRC). Someday we'll all see the silliness that is Windows and use Something Else but, until then, we're stuck with chips that needlesly pull more power than they should just to get a marketing edge, not a performance one.
I'm just still sitting here amazed that my 450MHz G4 can beat the snot out of many modern chips. Bad, bad design on Intel's part, mainly. AltiVec helps in certain cases, though. =)
CPU looooong ago stopped being a bottleneck. These days when you sit waiting for Internet Exploder 5.5 to appear after pressing the button, it's the disk that you're waiting for.
Sure, chuck in 128Mb of RAM, that'll cache the disk and solve the problem. Umm. nope. The CPU's are still 10 times faster than the RAM so it's still sitting around spinning, waiting for the data. And Windblow$ is terrible at managing it's paging so it still grinds away at the disk whenever you switch windows anyway.
Sure chuck in 10Mb of level one cache, that'll cache the memory and solve the problem. Have *seen* how expensive CPU's with massive L1 caches are?
My home system is an AMD Ksomething at 500MHz with loads of RAM and the fastest disks and SCSI bus I can afford. It still easily beats the latest and greatest 1.5GHz ATA/5400rpm based systems.
Deleted
AMD has basically no change in catching up mhz-wise until their next gen core, which is a long ways off..
Given the way Intel lets marketing lead engineering lately, perhaps AMD needs to do the same:
Replace the regular multiplier with one that first steps the clock up to 40Ghz, then back down to sane speeds. Call it a 40Ghz processor. Watch Intel switch back to "the clock speed doesn't matter!!!!"
If intel's chips are indeed twice as fast as apple's, then why do benchmarks (SPEC benchmarks) consistently show that Apple's chips can match intel's? I realize that RICS v CISC instruction set issue, piplining (Intel's are more than 2x long), and other architecture basics. The point of my post was to bring suspicion upon Intel and their clock speed counting.
As other people have mentioned, part of science is sharing ideas and thoughts with others in order to gain better insight onto the idea or problem, something you should be familiar with in your research for your masters. Many times there have been cases where a team has missed a crucial detail that undermines all of their other work.
Perhaps a better idea would be to run the code through CPU simulators of the PIII and G4 and get numbers on how many instructions are actually being executed in the entire code? I'm surprised no one has said that yet. If intel actually has twice as many instructions executed in the same amount of time, this would confirm that Intel's chips are running at 1ghz and apple 500mhz, but it would also demonstrate the efficiency in it's code execution and further provoke thought to the Hz myth.....
Just remember, we're only trying to figure out a rumor. We don't know the truth of it, and posting it here got back a lot of feedback.
The One Rule Of Chess You'll Ever Need: Don't play someone who carries a kit in their bookbag.
Actually this is not new.
When I upgraded to a Pentium-class computer, I bought an AMD PR-133 processor and thought I was getting a 133MHz processor. Only back home I discovered that it really was a 100MHz processor, which was approximately the same speed as a 133MHz Inter Penium.
I don't know have they used this trick in the meantime, but it's now totally new news.
But then again, it seems reasonable. Of course the public will only look at numbers in the name, not real benchmarks etc.
I doubt, therefore I may be.
Playing MP3's is an 'application' of a computer, like holding together Ducts is an 'application' of Duct Tape.
ReadThe ReflectionEngine, a cyberpunk style n
This was supposed to be some great russian bear cpu. I hacen't heard poop since over a year ago, what gives?
-You can cry, but you'll still die. There'll be no tears in the end.
Yeah, so? Isnt' the speed thatthe CPU runs at what's relevent? If some crazy cpu merchant sells me a 1GHz chip as a 1.2GHz chip and it runs stable at that speed, who cares if it's not "really" a 1.2GHz chip. Sure looks that way to me. If the system's not stable, then underclocking is one of my debugging steps before returning the chip. If it runs underclocked, then I take the chip back and bitch about it until I get a new one. Lather, rinse, repeat with a quality vendor, and quit worrying about what AMD/Intel sells their chips as. The different clock speed chips are the same damned thing anyway, but the faster ones have a more expensive guarantee.
:)
Though, I suppose you've got a point if it's really important for someone to know what AMD rated the chip at. I suppose that'd be useful in some environment, maybe, though I can't think of one off the top of my head...
It's somewhat unconventional, but I think of the frame buffer as a two-dimensional signal. Similarly, I think of an MRI image as a three-dimensional signal. If you look at the nuts and bolts of graphics processing, it's pretty much like any other signal processing: data ordered along axes, acted on by vector and matrix operations.
-- ;-)
Kuro5hin.org: where the good times never end.
One more issue is that the P4 has an AWFUL FPU. Yes, you're going to say that that won't matter as soon as software seriously supports SSE2. I call BS on that, because AMD never had the luxury of defining an alternative toe the FPU that would have a serious chance of wide-spread support. Intel, however, gets teh benefit of the doubt and everyone assumes it's ok for them to demand that all new software be written with THEIR proprietary instruction set, instead of having to do it well the old-fashioned way, where AMD kicked their sorry asses in FPU.
Nicotine free Amish .sig.