The Council of Better Business Bureaus has recommended (based on a tip by Dell) that Apple Computer discontinue comparative performance claims regarding its Power Mac G5 desktop.
Apple Told to Halt ‘World’s Fastest’ Claims for G5
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
see subject
Not to say Apple’s advertising wasn’t somewhat misleading, but doens’t Dell kind of have a conflict of interest in this?
Most advertisements I see make incredible and ludicrous claims.
It would be a conflict of interest if Dell was in charge of the BBB. But they are not. Dell simply tipped them off, and the BBB agreed that Apple’s claims went to far. I agree with the BBB, btw. Apple’s marketing department is completely immoral.
Come on now really, do you really believe most of the crap that they say in advertising? Why is it that they are singling out Apple?
And when was the last time Apple used that sloagan anyway? I only remember it being used when the G5 was launched.
Quality products and services my ass. My Dell OptiPlex nearly rattled its main fan off the case it was so poorly designed. Not to mention the fact that it performed very poorly for its specs (1.7Ghz P4, 256MB RAM). With all of the customer service complaints about dell they need to be bitchslapped over this.
If it is a lie than they should be asked to stop telling it, for the sake of the consumer.
Oh yeah, I really want Dell of all people looking out for my best interests. This coming from a company who makes some of the worst machines I’ve ever seen, and they are worried about consumer interests, HA!
“Come on now really, do you really believe most of the crap that they say in advertising?”
I’m not even a Mac user and I agree. This is going to backfire and make Dell look like whiners.
I agree that Dell isn’t really looking out for the best interests of the consumer, but I’m glad someone made a stink about that. The claim was pretty shakey when it was first made, and within a month it was no longer true at all(thanks to the opteron).
Yeah but Apple didn’t use that line for more than maybe a month or two. It has been nine months since its release.
The Council of Better Business Bureaus has recommended (based on a tip by Dell)…who where tipped by intel….
or told by intel to complain.
I think marketing is itself completely immoral.
These days marketing is all about convincing people that they want something that they otherwise would not want. I’d wouldn’t say Apple is worse than others really. Apple is bad at blowing tech specs out of proportion, sure, but do you really think Pepsi make’s people act younger? Or those colognes really make people instantly desirable to *everyone*?
I’d say this is good, so long as the BBB continues to go after all the OTHER false advertising around. How about sneaky car ads that show models you can’t acutally get in a given country? I don’t trust ANY ads anymore. The half of them are blatent lies.
I laughed when I first saw Apple’s claim. I can build a machine that would kick a dual G5’s 4$$ for half the price. Go BBB, go Dell!
I bought my iMac G4 from a local Apple store. The guy who sold it to me looked me in the eye and said that this iMac G4 700 MHz is 4 times faster than a Pentium 4 system.
But if I were at a Gateway store or a CompUSA, they know I’m getting a PC so now their tactic is to sell me extended warrenty and accessories I don’t need/want (when I bought my first PC, the salesman insisted I buy a surge protector from them
.
Business is business. Everyone has their own way of selling their products and making money. I have no idea why Dell would want to call Apple out on their claims. Dell’s customer base is use to buying $400 PCs, so Dell doesn’t have to worry about them defecting and buying $2000-3000 G5 PowerMacs.
You guys are all so negative. All I know is the Internet is faster with a Pentium 4, and Microsoft sees the potential[0] in all children.
[0] potential sources of perpetual income, that is.
You can build a dual 64 bit CPU workstation that has a bus speed on 1/2 the chips clock speed? That comes with a UNIX core OS that can run Photoshop, Office, almost an Linux app, will eventualy be able to emulate Windows via Virtual PC. And can run professional industry standard apps like Final Cut, Shake, SoundTrack, Logic, etc.
Where do I sign up? Hey can you build me one too?
Actually coming from a visual effects background I have seen a lot of Mac users protest to anyone that tried to point out the false claims Apple made around the world. A constant response was “Well PIXAR purchased G5 systems so they must be the fastest and most powerful 64-bit system” or “You guys are just jealous”. Failing to realize that Apple owns PIXAR and is the only highend post-production studio using the G5. As well that third party tests proved the G5 is not the most powerful or fastest as Apple wanted everyone to believe. It was false stating they were first to release a 64-bit CPU when AMD had already released 64-bit processors prior to the G5. Let’s also realize the G5 is a desktop, not a workstation. At least not one I would consider since it does not come with pro graphics (ie: FireGL/Quadro FX, WildCat).
When it comes to consumers purchase decisions they base purchasing items on advertising and word of mouth. Every company tries to stretch the truth a little to boost sales. In Apple’s case though they out right lied to the consumers in an attempt to boost their profits. What I find more appauling is that it took another competitor to point out the errors to the BBB. Why didn’t the BBB take action months ago when this advertising first started? Apple should be made to post a retraction to the media services appoligising for any confusion their ads may have caused to consumers.
The reason I single-out Apple’s marketing with particular vehemence is because of the way they use tech-specs to confuse clueless people. I have a NewsWeek with an Apple add in it that boasts about how the G5 can have 200 instructions in flight at the same time. The way they word it, it makes the reader thing that Intel chips can only process a few instructions at a time (which is true) while the G5 can process 200. None of it is a complete lie. The G5 does have a reorder buffer of 200 instructions and an Intel CPU does only process a few instructions per clock. But comparing the two doesn’t make any sense, because the G5 can also only process a few instructions per clock and the P4 has a reorder buffer of 126 instructions!
I consider this to be a worse form of marketing bullshit then Intel’s claims that “the P4 will make your internet faster!” Apple’s claims, because they sound technical, makes clueless contingent of the Mac community think they actually know something about computers, and come onto web-forums where they can shove their ignorance in everyone else’s faces.
Its good marketing, no doubt about that. Makes the reader feel smart, and manages to bash the competition in the process. But its completely immoral from my point of view.
“In Apple’s case though they out right lied to the consumers in an attempt to boost their profits.”
In the benchmarking, did not the dual G5 show itself to be the fastest at SOME tasks (obiviously, not all, but some)? So isn’t that enough to make the claim even if it’s not all the way true?
“Failing to realize that Apple owns PIXAR…”
Actually, both companies are public, and Pixar has NOTHING to do with Apple. Steve Jobs, of course, is CEO of both. But that’s where the partnership ends.
The artists use G5s, sure, but the render farm is all Intel. For now.
I also think they should ban claims along the lines of “America’s favorite coffee!” How do they know that? Did they do a survey? Where can I find their methodology to see if the results they obtained are statistically significant?
If you’re going to have laws about false advertising, at least enforce them properly.
Did anyone complain when Intel said the Pentium 4 would make the Internet faster?
Alpha.
http://www3.sk.sympatico.ca/jbayko/cpu5.html#Sec5Part5
?
First, Apple does not own Pixar. Steve Jobs is the CEo of both companies but he is mostly hands off at Pixar. Pixar also uses a Linux render farm. I also believe that the 64bit AMDs were not actually is desktop systems at the time the G5 was released.
And to Rayiner, it may be immoral in your words but my point is why single out one comapany. If you honestly look at any companies marketing you are going to find many discrepencies. In this case Dell felt the need to whine about something that hasn’t been used in commercials or advertising for a long time now. It was not smart of Apple to make the claim in the first place simply because of how fast the industry moves. But it should also not have been made as big of deal of as people have made it. I can tunr on the radio or TV and soon enough I’ll hear about the worlds best..whatever. Its a common marketing practice. And as far as you making comments about Mac users not knowing about computers then you again need a wake up call, I know a lot about computers and that is the reason I moved to the Mac OS X platform. It was more flexible than anything on the PC, allowing me to run my Linux programs, native OS X programs and Windows programs all side by side. Before throwing insults around maybe you should stop to think.
A lot of the comments on here seem to indicate that people think of the BBB as an officially sanctioned regulatory agency.
They are not. They are an independant organization that responds to complaints and arbitrates between the complainant and the complainee. the results of this process are then recorded and the business’ status as a member of the BBB is decided accordingly.
They do not do independant investigations.
Advertising, to me, is hilarious. Almost all advertisers at least imply their product are superior to all others. I guess there must be some line the BBB thought Apple crossed by claiming to be the “ultimate”. Apple probably would not have felt compelled to make an advertising campaign like that had it not been for the long, long G4 drought of pitifully small speed increments. It is odd that Dell made an issue of this after the ad campaign had stopped (if that was the case).
with signs in their windows saying…..
store 1 – best pizza in the world
store 2 – best pizza in the country
store 3 – best pizza on this block
I’ve soured to Dell a long time ago, every machine of thairs I have owned was a lemon. I tend to agree however that Apple was not only a little shady with their advertising but down right flagrant. I’m actually starting to sour to Apple computers as well since their product track record continues to circle the drain. No updates in over 6 months seems to suggest a struggling company though they are supposed to be doing quite well. I was at one point quite hot on buying a Rev.B dual G5 but now starting to question my ideas. Apple has noone to blame but themselves.
Dell forced to stop claiming that they make computers.
The latest three we got from dell: One has twice now decided to deny that it has a NIC. One (mine) had the HD go down two months into using it. The third… lets just say after the first two, it isn’t being used for anything too critical.
Also, my G4 450 runs circles around them. It’s a bloody 2.19 GHz lump of plastic. I would have never have claimed that a G4 is four times faster than a P4, but by clock speed 4 x 450 = 1.8 GHz vs. 2.19 GHz. Better make that five or six times faster.
I think Dell specializes in taking good chips, making bad computers, and bugging people to upgrade.
I would tell Apple to complain about Dell’s “Rock Solid” statements or the customer service claims or the refund statements. It is my wife’s and my opinion that those claims are completely false and we are still waiting for the reason they refused to return the sales tax so I can send that to the state comptroller’s office.
(we own no computer related stocks/funds and have friends that are employed at Dell or it’s new name)
>No updates in over 6 months seems to suggest a struggling company
Apple has released updates every 6 months or so for as long as I can remember. Nothing unusual
Well, for one thing, it’s a well-constructed, blazingly fast machine that’s ready for a 64-bit OS and applications; runs commercial-grade OS X apps side-by-side with popular Open Souce programs running under Xfree86; supports a suite of silky-smooth, Apple-only software like iPhoto, Final Cut, and Garage Band — well, one could go on and on. Plus Macs really hold their resale value. It’s amazing to see two or three-year-old G4s on eBay pulling 75% of their original purchase price.
I’m biased, having finally broken open the piggybank and anteed up for a G5. But I’m a longtime Linux user, and spend time on a reasonably quick Win XP box at work. It’s always a pleasure to get home and fire up the Mac. To me, at least, it was well wortyh the money.
Thanks for the correction. I’m just a little frustrated with Apple advertising gimmicks.
You’re pretty fast, but a bit outdated.
The G5 does have a reorder buffer of 200 instructions and an Intel CPU does only process a few instructions per clock. But comparing the two doesn’t make any sense, because the G5 can also only process a few instructions per clock and the P4 has a reorder buffer of 126 instructions!
I consider this to be a worse form of marketing bullshit then Intel’s claims that “the P4 will make your internet faster!” Apple’s claims, because they sound technical, makes clueless contingent of the Mac community think they actually know something about computers, and come onto web-forums where they can shove their ignorance in everyone else’s faces.
What the Apple marketing claims that the G5 can have 200 instructions in flight and the P4, according to you can have only 126 in it’s reorder buffer. Which is true albiet worded to favor the g5.
That is worse that claiming that the internet will be faster with a pentium 4. or centrino technology will allow me to work from mount everest and the beach. Give me a break. Which is taking advantage of the cluelessness end users?
The Apple G5 cluster is faser than both the opteron based one (same clock speed more nodes) and xeon/itanium based ones with lesser number of nodes (top500.org). Are you telling me that is Apple marketing too?
I remember apple claiming to have the world’s first 64-bit personal computer. If I recall the opteron is a workstation and server chip. AMD’s desktop line is the Athlon64 which didn’t come out till much later. So techincally there is nothing wrong with what apple claimed when it release the G5.
We all know that marketing and advertising is nothing more the propaganda with glitter based on half truths. It’s been that way since is inception in the early 20th Century.
Who cares what Apple claims. Dell claims to be America’s choice PC, which hogwash when in actuality the home-built PC is the most popular choice amongst Americans and abroad!
If the P4 takes a “narrow and deep” approach to performance and the G4e takes a “wide and shallow” approach, the 970’s approach could be characterized as “wide and deep.” In other words, the 970 wants to have it both ways: an extremely wide execution core and a 16-stage (integer) pipeline that, while not as deep as the P4’s, is nonetheless built for speed. Using a special technique, which we’ll discuss shortly, the 970 can have a whopping 200 instructions on-chip in various stages of execution, a number that dwarfs not only the G4e’s 16-instruction window but also the P4’s 126-instruction one.
Is this the apple marketing claim you are talking about? Wrong. That is ARStechnica’s article found here.
http://www.arstechnica.com/cpu/02q2/ppc970/ppc970-2.html
Additionally, the PowerPC G5 features processing innovations that optimize the flow of data and instructions — meaning the PowerPC G5 can pump through more than 200 in-flight instructions at a time, a whopping 71% more than the 32-bit Pentium 4.
That is Apple’s website.
http://www.apple.com/g5processor/executioncore.html
How is that misleading?
Well, this all reminds me on a german saying: “getroffene Hunde bellen”, which means in english somehing like “hitten dogs bark”.
I think, just because Dell complained about apples advert, there *must* me something true with it, else Dell would have silently ignored it. (btw, is Dell selling *ANY* 64 bit Workstation at all?).
Btw, somebody here said that “he” could easily build a compareable cheaper machine… I really doubt that.
They’ll be able to re-run the campaign this summer when the Dual 3Ghz 970fx starts shipping
What the Apple marketing claims that the G5 can have 200 instructions in flight and the P4, according to you can have only 126 in it’s reorder buffer. Which is true albiet worded to favor the g5.
Another reason not to put technical literature in advertising. The size of a re-order buffer is an engineering trade-off. A larger re-order buffer is a bad thing to have if it takes away transistor resources that could be spent elsewhere. See, a reorder buffer only needs to be as big as it needs to be. The G5 has more execution units than a P4, and thus it makes sense that the reorder buffer should be bigger. From an engineering point of view, comparing these sort of technical details is like taking somebody’s quote out of context.
That is worse that claiming that the internet will be faster with a pentium 4.
Note that they claim that the internet “experience” will be faster. If you do stuff like use Flash videos, online games, etc, that’s a true statement.
or centrino technology will allow me to work from mount everest and the beach.
When did Intel ever claim that?
Give me a break. Which is taking advantage of the cluelessness end users?
Both are taking advantage of the cluelessness of end users, no doubt about that. My point, however, was that Apple’s marketing is more of an affront to the technically minded, because it makes the end-user *think* that they are not clueless. Moreover, people are used to that type of marketing, were the vendor implies that wearing Axe deodorant will have women fawning over you. They can recognize that and weed it out. I would argue that nobody honestly believes that to be true, and that vendors know that. Apple showing misleading numbers, however, is a different situation. Most people do not have the knowledge come to the true conclusions in this case, and there is no way that Apple marketing department believes that most people are smart enough to do so.
“Dell claims to be America’s choice PC”
Isn’t “America’s Choice” the name of the sub par generic items available in supermarkets? If so, it’s a perfect claim.
You have to wonder why the largest PC maker would waste its time with a company that’s on its way out? I smell fear.
I’ve soured to Dell a long time ago, every machine of thairs I have owned was a lemon. I tend to agree however that Apple was not only a little shady with their advertising but down right flagrant. I’m actually starting to sour to Apple computers as well since their product track record continues to circle the drain. No updates in over 6 months seems to suggest a struggling company though they are supposed to be doing quite well. I was at one point quite hot on buying a Rev.B dual G5 but now starting to question my ideas. Apple has noone to blame but themselves.
Maybe with your new G5, you’ll be able to run spell check – just an idea! And really, Apple’s benchmarking was done by a third-party using the processor agnostic GCC compiler with settings not completely understood by this here guy. No, this here guy likes real-life hands-on tests – here, the G5 kicks some booty. Git yer G5 – you’ll be happy….
Additionally, the PowerPC G5 features processing innovations that optimize the flow of data and instructions – meaning the PowerPC G5 can pump through more than 200 in-flight instructions at a time, a whopping 71% more than the 32-bit Pentium 4.
That’s it! That’s entirely misleading, because the words “pump through” imply to any sensible person “complete,” or “go through,” or “execute.” Instead, the appropriate word would be something along the lines of “juggle.” Does juggling more instructions at a time allow you to complete them faster? Maybe, maybe not. Beyond that, I’ll reiterate my comment about engineering trade-offs. Highlighting details like this, without providing a reasonably complete picture of the situation, is akin to taking somebody’s claims out of context.
“or centrino technology will allow me to work from mount everest and the beach.
When did Intel ever claim that? ”
It’s in one of their TV ads. Funny thing the Everest IMAX film crew used Apple G3 laptops on the mountain.
I have read the complaints that Java the Hut is “The world’s best coffee” and Apple saying they have the fastest computer are the same thing. Nonsense.
Coffee, or any other food for that matter is a subjective taste. Wrold’s best to who? If one guy says it after drinking your coffee, then it’s true. At least to that guy.
Apple saying they have the world’s fastest computer is a testable claim that is not subjective. If you run 10 benchmarks and 5 of them slaughter the competition than you have the fastest computer at those 5 tasks, not the fastest computer period. They need to be slapped for it. It’s a false claim.
Wether it’s the BBB or someone else, there needs to be someone enforcing standards in advertising. Go BBB, and if Dell deserves it I hope they ride Dell’s butt for it too. And someone has to complain… this time it was Dell, next time it might be you. Don’t you want someone to listen and take action?
Apple saying they have the world’s fastest computer is a testable claim that is not subjective. If you run 10 benchmarks and 5 of them slaughter the competition than you have the fastest computer at those 5 tasks, not the fastest computer period. They need to be slapped for it. It’s a false claim.
The great thing about benchmarks is, like standards, there are so many to choose from. I’m sure Apple or Dell could come up with 10 benchmarks each that show their computer is faster. Granted, if the Dell lasts long enough to run the benchmark remains to be seen.
My friend, a recent purchaer of a 12 iBook; was trying to convice me that the G5 was the worlds fastest computer….he was using the brochure as proof; I sat him down and talked him out of his insanity. It isn’t to say that that the G5 isn’t a beutiful and wonderful Computer…but it is lying to say it is the worls fastest. Oh just to show the power of advertising, my frined is very computer literate; a professional programmer, and all around nerd; in fact he is the one who got me started playing with computers.
Now as to those compaering the Worlds fastest… campaign to various Wolds Best.. campagins; best is not a gaugable aspect. I say that Grolsch is the best beer; but others might say Guiness. Fastest on the other hand is guagable. When it was said that the McLauren F1 was the fastest street leagle production car…it was. No other car could do 0-60 in 3.1 sec. That is fastest. Computers are harder to guage, but still to say that one is fastest you should have reall proof (a hard thing to vcome by).
Well those are my rants (sorry bout the spelling and grahmer , been up since 3:00 am)
“I think, just because Dell complained about apples advert, there *must* me something true with it, else Dell would have silently ignored it. (btw, is Dell selling *ANY* 64 bit Workstation at all?). ”
That’s pretty weak logic. Apple misrepresented the performance of *Dell’s* product in their advertising.
BTW, Apple isn’t selling a “64-bit workstation” – they’re selling a 32-bit consumer desktop that just happens to use a 64-bit chip.
“Btw, somebody here said that “he” could easily build a compareable cheaper machine… I really doubt that.”
You can doubt the sky is blue, too.
I believe it could honestly be established that the G5 is indeed the world’s fastest PC.
Most of the benchmarks are so selective about what they actually test. They completely ignore the reality of using a complete program; measuring just the processing of a few numbers, rather than total throughput, or working with an interface to get to the section to be measured.
Take this example from the PC Magazine review of the G5 which shows the Mac to be marginally faster than a Wintel at Photoshop:
http://www.pcmag.com/article2/0,1759,1274230,00.asp
[i]At these larger image sizes, although the Wintel test times were quite good, both the G4 and G5 computers proved more adept at distort functions like wave and pinch. Moreover, on the Windows system, loading the controls often took a minute or more. If these times are added back to the actual test times, both Macintosh computers would have clearly outperformed the Windows-based computer.[i]
How nice that they could ignore the reality of working with the operating system and measure only discrete components of a program.
Then watch the video from last year’s Apple Developer’s Conference with Photoshop runing twice as fast on the Power Mac G5 than on 3.06 GHz Dual Xeon workstation. Did they use operations that would favor Mac? Yeah, of course I’m sure they did. But watch it. It’s real world, and a fair test.
Saying the Apple Dual G5 is fastest is a measure of the whole machine; not just integer, floating point, or vector performance of the IBM G5 chip.
Just do what other companies do…
“Compared to the LEADING BRAND, Apple’s computer get your clothes whiter.” (er, that is, “are the fastest personal computer.”)
“Then watch the video from last year’s Apple Developer’s Conference with Photoshop runing twice as fast on the Power Mac G5 than on 3.06 GHz Dual Xeon workstation. Did they use operations that would favor ”
Apple has never lost a bakeoff on their own stage. Never.
Yeah. And its not more than 200 instructions, since the re-order buffer can hold up to 200 instructions. The more than is misleading too.
But that’s marketing for you.
“That’s pretty weak logic. Apple misrepresented the performance of *Dell’s* product in their advertising.”
Yea for sure. I just ignore that, ok?
I really doubt that you can seriously compare two totally different architectures.
“BTW, Apple isn’t selling a “64-bit workstation” – they’re selling a 32-bit consumer desktop that just happens to use a 64-bit chip.”
Wrong at all! Just because most of Panther isn’t 64 bit right now, at least the kernel *HAS* to be 64bit, and thats whats important, the machine is 64 bit, ahd has not only some asm commands which are 64 bit (like some NEC VR cpus). You really would notice that when compiling a Linux kernel for a G5. (this will only boot when compiled for 64 bit, like for example on Ultra Sparc or PA-RISC) (so with Panther ATM you have a 64 bit kernel and 32 bit userspace. Thats right now just the same with Linux on for example a PA-RISC machine.)
Another reason not to put technical literature in advertising. The size of a re-order buffer is an engineering trade-off. A larger re-order buffer is a bad thing to have if it takes away transistor resources that could be spent elsewhere. See, a reorder buffer only needs to be as big as it needs to be. The G5 has more execution units than a P4, and thus it makes sense that the reorder buffer should be bigger. From an engineering point of view, comparing these sort of technical details is like taking somebody’s quote out of context.
Apple’s marketing literature didn’t compare the reorder buffers you did a few posts ago!!
The Allocator allocates a Reorder Buffer
(ROB) entry, which tracks the completion status of one of
the 126 uops that could be in flight simultaneously in the
machine.
For you kind information the P4 can have 126 uops inflight. http://www.intel.com/technology/itj/q12001/pdf/art_2.pdf. The ROB tracks status of the uops and is 128 entries long.
the 970 can have a whopping 200 instructions on-chip in various stages of execution, a number that dwarfs not only the G4e’s 16-instruction window but also the P4’s 126-instruction one.”
Various stages of exection being the key word here from ARStechnica’s article.
or centrino technology will allow me to work from mount everest and the beach.
When did Intel ever claim that?
Have you ever seen thier “unwire” centrino commercials on TV?
Wee-Jin Goh
Yeah. And its not more than 200 instructions, since the re-order buffer can hold up to 200 instructions. The more than is misleading too.
“Group Completion Table entry: The GCT is the 970’s equivalent of a reorder buffer. The GCT has 20 entries for keeping track of 20 active groups as the groups’ constituent instructions make their way through the ~100 execution slots available in the execution core’s pipelines. Regardless of how few instructions are actually in the execution core at a given moment, if those instructions are grouped so that all 20 GCT entries happen to be full then no new groups can be dispatched. ” ARStechnica
Each group is 5 iops.
Actually the first 64Bit consumer machine would be an Atari Jaguar. Not a G5 or AMD.
This just makes it look like Dell are running scared and can’t compete. Plus it’s amazingly petty of Dell when so many other ads make unsupported claims.
It would be nice to see some extensive benchmarks testing the G5s, Athlon 64s, and Opterons. There is http://www.pcworld.com/news/article/0,aid,112749,pg,8,00.asp but one benchmark isn’t enough to convince me; however, I really think the G5 isn’t the fastest by a longshot, and especially not worth the high price. Additionally I wouldn’t really want to buy into the Apple monopoly and seemingly lousy market.
The BBB thing is good, and maybe more companies will call each other out on their bullshit (hopefully the BBB won’t heed every petty complaint though).
Highlighting details like this, without providing a reasonably complete picture of the situation, is akin to taking somebody’s claims out of context.
I have a NewsWeek with an Apple add in it that boasts about how the G5 can have 200 instructions in flight at the same time. The way they word it, it makes the reader thing that Intel chips can only process a few instructions at a time (which is true) while the G5 can process 200.
If I am not mistaken it is you who highlighted the 200 inflight instructions claim and brought it out of context. In fact, you even went so far as to claim that the ppc970 has more execution units and thus should have a larger reorder buffer.
The G5 has more execution units than a P4, and thus it makes sense that the reorder buffer should be bigger.
The architectures of both those chips are different and you comparison of reorder buffers is invalid. The GCT can keep track of 100 iops.
IBM.com http://www-306.ibm.com/chips/products/powerpc/processors/
PPC 970 Features
• Instruction pipe
– 64KB L1 Inst cache, direct mapped
– 32 entry I buffer
– 8 instructions fetch / cycle
• Branch prediction
– Highly accurate dynamic prediction
• Dispatch, issue
– 1 group (4 + branch) / cycle
– Up to 20 active groups
– Up to 8 issue / cycle
– Over 200 instructions in flight ****************<<<<
• Data pipe
– 32 KB L1 Data cache, 2-way sa
– 32 x 64b GPR, FPR
– 32 x 128b VRF
– 512KB L2 cache, 8-way sa
– 8 data prefetch streams
I have presented to you multiple sources of how many inflight instruction(uops/iops) a p4 and ppc970 can have in thier pipeline, including intel’s own document about the P4. Yet you claim that Apple’s, ARStechnica’s and Intel’s definition.
Intel says p4 126 uops inflight
IBM says ppc970 200 instructions inflight
There is nothing misleading about the 200 number.
You really should read the following http://www.apple.com/g5processor/executioncore.html an then tell me what is factually incorrect about thier description of the ppc970’s pipeline.
Dell’s consumer base is $399.00, dude your getting a dell kind of people, why would they be whinning at Apple for selling more expensive computers…….
Dell also sells expensive workstations, such as dual 3.2Ghz Xeons. They’re probably more concerned about businesses deciding between Precisions and G5s than individuals.
I feel that I need to chime in here with something, unless I missed an equivalent post, has been overlooked. Also, I believe this is something important to point out to foreign friends, as I’m not sure exactly what Eugenia means here by “The Council of Better Business Bureaus.” There was a post earlier that said that they were not a regulatory commission, and I’m assuming he is an American like I am. The American BBB is not a Council, per se. So I’m not sure if Eugenia is talking about something from a home country, or from the US. Anyway …… if you’ve read through this little rant this far, let me point out one thing.
The American BBB is a *Business*, plain and simple. It is not a Federal Bureau, as the name can imply, it is not a Council of elected officials, it is not Federally (American Government) Commissioned to do anything. They are a business just like Apple, Dell, Chevrolet, Pizza Hut, and Bob’s Used Stereo Hut.
Many, many people do not realize that it requires only three things to become a member of the BBB (U.S.) These are:
1. To be in business a minimum of 12 months (this may have changed)
2. To be relatively complaint free while in business.
3. To be your annual dues.
Again, they are a business, they collect money to grant the right to display the BBB torch. Their job is to take complaints and file them against a company. But they rarely ever do anything about the complaints.
Ergo, a real problem here is home improvement. Companies coming into business, installing let’s say replacement windows. The windows mess up, you call the BBB, you file a complaint and the BBB says “Thank you for your call.” They then call the Window Installation Company (Bob’s Used Windows) and tell him he has a complaint against him.
Then …. they hang up. If BUW, Inc. (Bob’s Used Windows) decides they aren’t going to fix your problem, then they aren’t going to fix your problem. And this has in fact led to the BBB creating a couple of other programs that are invite only with which they *will* get involved and send an arbitrator out to take care of the problem.
But basically, all I’m saying, is that *I* could call and say “I’d like to make a complaint against Dell.” and the BBB would happily take the complaint down, then call Dell and then …. hang up and go on about it’s business of taking more calls.
Actually the first 64Bit consumer machine would be an Atari Jaguar. Not a G5 or AMD.
LOL, we’re talking CPU’s here, not a few components or buses. The Jag used a 68000, which was 16 bit (Tempest 2000 _still_ rocks though).
Actually the atari jaguar had 5 processors, 3 of which were general enough to be considered CPU’s, the least of which was a 68000:
it had:
a 32-bit RISC programmable GPU,
a 64-bit RISC Object PU,
a 64-bit RISC Blitting PU,
a 32-bit DSP,
and a 16-bit 68000 running at 13.295 MHz
“Yea for sure. I just ignore that, ok?”
Notice who’s product Apple compared the G5 too.
“I really doubt that you can seriously compare two totally different architectures.”
Of course you can. Well, *I* can.
“Wrong at all! Just because most of Panther isn’t 64 bit
Hardly.
“right now, at least the kernel *HAS* to be 64bit, and thats whats important, the machine is 64 bit, ahd has not ”
Without 64-bit memory addresses, it isn’t a 64-bit machine.
My Octane is a 64-bit machine. Your G5 is not.
“only some asm commands which are 64 bit (like some NEC VR cpus). You really would notice that when compiling a Linux kernel for a G5. (this will only boot when compiled for 64″
Linux isn’t OS X. Apple doesn’t sell a 64-bit desktop.
” bit, like for example on Ultra Sparc or PA-RISC) (so with Panther ATM you have a 64 bit kernel and 32 bit userspace. Thats right now just the same with Linux on for example a PA-RISC machine.)”
That’s not a 64-bit OS. Sorry.
I agree that Dell has customer services issues (see http://www.lparky.com/blog/PermaLink,guid,cc9535f7-6bf6-411a-adf0-7… ).
However, they are generally good, in my experience, in making up for their mistakes (see http://www.lparky.com/blog/PermaLink,guid,5c3a5c34-175e-4c00-9a3c-9… ).
All that said, whether Dell tipped off BBB or not, Apple’s ads were inappropriate and never should have aired. If nothing else, it strengthens the arguments of the anti-Mac folks that Steve Jobs is full of crap.
Look at the empirical evidence regarding the heat and power consumed by the PPC970 and make up your own mind:
— Giant heat sinks
— 600W (and noisy) power supply for a system that only takes 2 low power drives
— 80W per CPU power system for 1.4Ghz CPU
— Lack of any ability to scale past 2Ghz over the past 9 months, even with 90nm chips
— G5 server runs only one (1) PPC970FX due to heat/power issues
— IBM running their blade servers at 1.6Ghz, slower than any Apple system
— lack of Apple G5 laptop, G5 iMac, etc., all due to high power/heat.
And on the other side, there is no empirical evidence showing that Apple’s claims of low heat/power are true.
Apple deserves to be spanked by the BBB for their outrageous lies regarding performance. I wonder who will spank them for all the other lies they tell. It looks like the entire G5 is a sham and that Apple made another dumb decision by deciding to stick with PowerPC.
I still don’t understand when I see reviews from the media on the performance of workstations and then they use useless tests. No one should class a desktop or workstation speed based on how fast it can open a Word document or surf the Net. Using real world tests such as those found at http://www.spec.org/ is a good way of running speed tests. Another good example of what tests should be run to evaluate workstation performance can be found here http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_88…
If you guys/gals are interested in seeing how the G5 stacks up against other procs for rendering with Maya and Mental Ray see this link http://zoorender.com/ Click on “Benchmarks/Render Benchmark/Mental Ray Render 5.0 and above”. It’s interesting seeing where Apple’s fastest dual G5 placed even with 4GB of PC3200 DDR400 system RAM.
Dell is cheating their customers by supplying SXGA+ screen which is far inferior than what they advertise in their specification. They supply inferior quality Hitachi screen in all WSXGA computers. Never going to buy Dell again. Look here:
http://forums.us.dell.com/supportforums/board/message?board.id=insp…
I bought my first Mac with the advent of the G5. Before doing so I did a lot of online research comparing high end AMD’s, Pentium 4’s, and Apple G5’s. the specs were reasonable close and I looked at high end systems from both Dell and Alienware. Since I qualified for a discount from Apple the Dual G5 came in a bit cheaper than the rest. And while it might not be the worlds fastest desktop computer, it was close enough for me. Anyone who believes all the claims made by any computer company should email me. I have a bridge here in town that I am willing to sell cheaply. The G5 is probably the best engineered computer system I have seen to date. The case is a marvel and the inside is totally uncluttered with cables. Personally, I think Dell is just jealous because they couldn’t desing a computer like this.
Bill
i dont see why dell should complain. if apple is lying and dell is is loosing business because of it? does that mean dell should lie about performance too?
If I am not mistaken it is you who highlighted the 200 inflight instructions claim and brought it out of context.
They were already out of context in the ad!
In fact, you even went so far as to claim that the ppc970 has more execution units and thus should have a larger reorder buffer.
The G5 does have more execution units. The P4 has two regular integer, one limited integer, one floating-point, one floating-point move unit, and two address-generation units. The G5 has two integer units, two floating-point units, two two load/store units, two vector units and a branch unit. The G5 also dispatches more instructions per clock cycle (4+1 vs 3), and its execution resources are more general and easier to utilize in parallel. For example, the P4’s AGUs don’t come into play except for indirect memory accesses.
The architectures of both those chips are different and you comparison of reorder buffers is invalid.
Wasn’t that my point? They have a reorder buffer size appropriate for their respective architectures. Thus, comparing the two makes no sense.
There is nothing misleading about the 200 number.
I never said there is. What I said is misleading is comparing the size of the reorder buffers in the two processors. Specifically, Apple’s comment that “the PowerPC G5 can pump through more than 200 in-flight instructions at a time, a whopping 71% more than the 32-bit Pentium 4.” Its taking the numbers out of context. The size of the reorder buffer is a design trade-off, and taking it out of context is idiotic. Its like Honda comparing its S2000 engine to a Porsche’s and saying that their engine runs at a much higher RPM. Beyond that, the words “pump through” clearly imply to me “executing” or “completing” not “reordering.” The fuzziness of the term changes the meaning *significantly*.
By Jk
This just makes it look like Dell are running scared and can’t compete. Plus it’s amazingly petty of Dell when so many other ads make unsupported claims.
right on. its so obvious why are we all debating about the subject?
I never said there is. What I said is misleading is comparing the size of the reorder buffers in the two processors. Specifically, Apple’s comment that “the PowerPC G5 can pump through more than 200 in-flight instructions at a time, a whopping 71% more than the 32-bit Pentium 4.” Its taking the numbers out of context. The size of the reorder buffer is a design trade-off, and taking it out of context is idiotic. Its like Honda comparing its S2000 engine to a Porsche’s and saying that their engine runs at a much higher RPM.
Haven’t I made it abundantly clear already. According to Intel and IBM the manufactures of said processors (P4 and PPC970), they have 126 instructions and over 200 instructions inflight respectively. Therefore the comparison was accurate.
You are inaccurately extrapolating what inflight instructions mean to the size of the P4 reorder buffer. I have provided adequate technical info to verify my point. The ppc970 have a 20 group GCT, each group being 5 instructions, which is smaller than the P4’s 128 entry reorder buffer. Thus invalidating your argument that 200 inflight instructions means the PPC970 has a 200 entry reorder buffer which it clearly doesn’t. Like I said the P4 and the G5/ppc970 are different architectures and the number of inflight instructions need not be the same as the reorder buffer entries. Yet you assumed that since the p4 has 126 uops inflight and a 128 entry ROB that both architectures would have inflight instructions = ROB size.
The cumulative number of instructions in a G5 cpu’s pipeline including the 100 entry gct and executiong units is over 200, according to IBM,APPLE and all the articles and university papers talking about the ppc970. And the P4s pipeline can have 126 instructions at any given time, this fact is also from Intel’s P4(netburst) microarchitecture paper. All the relevant links are in my previous comments.
Beyond that, the words “pump through” clearly imply to me “executing” or “completing” not “reordering.” The fuzziness of the term changes the meaning *significantly*.
Yes, I am glad you said “to you they imply” all of the above. Which is exactly what they mean according to the manufactures of either chip and many other sources and they all seem in agreement with what Apple said in thier literature (read linked papers previously). You really must come up with real facts and not your personal interpretations.
I don’t think any of Apple’s claimes are outlandish or any worse than other manufacturer’s embelishments of facts. Like Intel’s propogation of the megahertz myth via advertising that more GHz means better/faster, which I think they shot themseleves in the foot with and are trying to go the AMD route by naming new chips with series numbers.
“These days marketing is all about convincing people that they want something that they otherwise would not want. ”
And this has changed how in the past 3000 years?
I am certain that if you went to a market or fair in ancient Babylon, somebody would have tried to sell you something you didn’t need.
You are inaccurately extrapolating what inflight instructions mean to the size of the P4 reorder buffer.
You’re correct on that. I shouldn’t have mentioned the reorder buffer. To be clear, like the ad, I’m only talking about in-flight instructions.
Yet you assumed that since the p4 has 126 uops inflight and a 128 entry ROB that both architectures would have inflight instructions = ROB size.
Right, but that doesn’t really change my point. My point didn’t concern the distinction between the re-order buffer and the number of in-flight instructions. I was trying to get at the fact that the G5 has more in-flight instructions because its a wider CPU with more functional units. Comparing the number of in-flight instructions between the CPUs in marketing literature doesn’t make sense, because its an implementation detail, not a useful measurement. Moreover, the way its phrased (“pump through”) it makes it seem like the G5 executes 71% more instructions in a given period of time, which is clearly not the case.
To reiterate my car analogy: the Honda S2000 engine achieves its power output be readlining at nearly 9000RPM. However, they don’t say on their marketing literature: “well, our engine red-lines at 9000RPM, vs Porsche’s which redlines at 7000RPM. That’s nearly 30% more RPM!” That would be a ridiculous comparison of statistics, because ultimately RPM doesn’t mean anything — horsepower and the output curve does.
Like Intel’s propogation of the megahertz myth via advertising that more GHz means better/faster
I’ve never seen an Intel ad that plays up the P4’s higher clock speed. And the megahertz myth is not a myth. More megahertz, all else being equal, means better performance. Contrary to what all the conspiracy theorists would like to believe, Intel did not design the P4 the way they did just so they could tout huge clock-speed numbers. They did it because they bet (rightly) that customers would favor better performance on streaming-style multimedia code over branchy-integer code. A deep, highly-clocked pipeline doesn’t suffer nearly as high a penalty when running the former type of code, and gets a large performance boost from the higher clock-speed.
to the guy who said buying macs was a monopoly.. yeah because if i buy a mac it can only run.. mac os x, and *nix. if i buy a pc it can only run… windows and *nix.
so basically you’re saying apple has a monopoly on os x? that’s fine. but they do NOT have a monopoly on the OS any more than windows on x86. the only restricted choice you get is not being able to run windows.. oops! sheesh, wrong about that too. try virtualpc sometime. it’s fantastic on g4s, and coming out soon for g5s.
so with a mac i can run.. os 9, os x, *nix, and windows software all without leaving os x. on a windows machine i can run.. *nix (via virtualpc) and windows without leaving windows.
the difference with the top one is os x is a shit ton better than windows, and i get native performance from my *nix and almost-perfect native os 9 software.
sigh you people. these processor wars are aways pointless. the g5 is plenty fast for most needs. you get it for the operating system, NOT for the hardware. i would rather purchase a $2000 dual 1ghz g4 than a POS windows box.
‘ve never seen an Intel ad that plays up the P4’s higher clock speed. And the megahertz myth is not a myth. More megahertz, all else being equal, means better performance. Contrary to what all the conspiracy theorists would like to believe, Intel did not design the P4 the way they did just so they could tout huge clock-speed numbers. They did it because they bet (rightly) that customers would favor better performance on streaming-style multimedia code over branchy-integer code. A deep, highly-clocked pipeline doesn’t suffer nearly as high a penalty when running the former type of code, and gets a large performance boost from the higher clock-speed.
Give it up. The PIII at a the same clock speed is faster than the P4. Why oh why did the pipeline stages just increase every iteration. AMDs have always been faster at lower clockspeed than P4s. Clockspeed tends to help integer performance greatly but not throighput or FP. Itaniums at lower clockspeeds are faster than p4s at nearly twice the clockspeed, I just relesed X cpu and broke the X Gigahertz speed barrier was AMD vs Intel in the 90’s main marketing tactic. AMD gave it up and started versioning thier CPUs XP XXXX+ to one up intels gigahetrz claim and went with the series versioning scheme with the opterons with intel following suit.
Please read the myrad of atricles on Pentium-Ms being faster than mobileP4s at lower clockspeed largely due to a smaller pipeline. I think Clockspeeed as a marketing term fits you car analogy a lot closer than the number of inflight instructions. You are right all else being equal a higher Megahertz means faster. By that I mean a 900MHz UltraSPARCIII is most certainly slower than a 1200MHZ USIII, but that is not the ineterpretation of the Megahertz Myth most people have. But you seem to have a penchant for interpreting things differnetly.
They didn’t bet (rightly) that customers would perfer high GHz machines, that’s why a naming convention change was announced recently for the same P4 line. Becuase the “customers” grew smarter or intel’s marketing department said “damn we have three/four lines and our high gigahertz chips are producing lower /equal performance to our other series it’s time to change the perception amongst the clueless people who walk into frys and was a PC with 40GB memory(HD space)”.
I think intel plays more with marketting and clueless users than Apple. Helll they put “intel inside” logos on boxes till than most PC buyers couln’t care less what cpu ran in their boces. What cpu is in you cars onboard computers, or VCR or you buildings cooling system? Exactly. Marketing,
May be you just have a thing for Intel and Microsoft. I must admit you are a well disguised creature under the bridge.
To reiterate my car analogy: the Honda S2000 engine achieves its power output be readlining at nearly 9000RPM. However, they don’t say on their marketing literature: “well, our engine red-lines at 9000RPM, vs Porsche’s which redlines at 7000RPM. That’s nearly 30% more RPM!” That would be a ridiculous comparison of statistics, because ultimately RPM doesn’t mean anything — horsepower and the output curve does.
That’s a great anology to prove my point. But your comparisons are wrong.
Let’s say inflight instrcutions = Number of cylinders and RPM = Clockspeed.
Porsche has a 6 cylinder engine and the s2000 a 4 cylinder one. G5 200 inflight instructions and P4 126. The Porsche redlines at 7000 RPM and the S200 at 9000RPM. G5 redlines at 2Ghz and P4 at 3.2 GHZ (approx). However, the porsche uses a wider Engine with more displacement and at lower RPMs produces more HP than the Honda engine. Prosche can surely say its 6 cylinders is better than honda’s 4, wouldn’t you? The honda engine needs to be rev’ed a lot higher than the porsche Engine to produce similar HP.
The P4 needs to be rev’ed a lot higher due it’s narrow and deep pipeline to produce performance numbers similar to the G5 at lower clockspeeds. The honda may produce higher HP at 10000RPM than a Porsche engine but at the expense of inefficiency and more noise. Similarly a G5 cpu is a more efficient design owing to its wider pipeline than a P4 in terms of heat dispation and power consumption while producing smilar performance. There fore the width of the pipeline and inflight instructions is a valid comparison point.
Anyway you slice it Apple’s the marketing literature isn’t misleading no matter how much you want it to be.
The honda may produce higher HP at 10000RPM than a Porsche engine but at the expense of inefficiency and more noise.
Should be
The honda may produce higher HP at 10000RPM than a Porsche engine at say 7000RPM but at the expense of inefficiency and more noise.
Don’t you think we have to start holding business accountable here? I mean, if we just let them lie to us, manipulate us, and essentially screw us because, well, it’s business, then does that not mean we deserve to be screwed over by them?
It disturbs me how willing we are to excuse business anything. Money is not the most important thing in the world, perhaps we should make integrity and honesty the most important thing.
Hey, I believe Apple has a point to call their machines the most powerful and fastest Personal Compueter in the world, I mean, the Xeon and itanium Processors are not exactly “Personal Computer” CPUs, not to mention that their performance is lacking in a lot of ways, but the P4, well, that gets harder to gauge, though considering the Power Macs went from lagginf behind to neck in neck in one fell swoop, and if rumours about this years 3ghz models are true, will surpass the best Dell, HP, and Gateway have to offer easily.
Anyway, …
Hector
“Now, regardless of how you feel about the accuracy of Apple’s marketing claims (they are marketing claims, after all– how come no one got on Intel’s case when its ads implied that using a Pentium 4 would give you faster downloads?), those Xserve G5s must really have Dell quaking in its boots to make them pull a cheap stunt like calling the BBB.”
http://www.appleturns.com/
Could they stop Apple from claiming that OS X is “the world’s most advanced operating system” too? Please? Pretty please?
having the opportunity to cut/post edit a video animation on an apple powermac g5, i must say i am personally pretty disappointed. beat me with your zealot sticks, but osX is as prone to application crashes as windows. after effects crashed twice while i was trying to import a high quality compressed png sequence of 4000 single images. the quicktime rendering itself out to dvd (.m2v files) was far beyond mediocre. i just say terrible terrible artifacts in a cartoon style of animation. lets drop out the fact that for some insane reasons apple’s dvd drives cant burn dvd+r discs, now thats a good features/cost ratio.
this machine has around a gig of ramspace, yet it was sluggish to render something out of aftereffects in a batch rendering, while i tried to encode the resulting files to m2v using quicktime, the fact that you can only encode one file at a time in quicktime “professional” is also quite weak.
I’ve never seen an Intel ad that plays up the P4’s higher clock speed. And the megahertz myth is not a myth. More megahertz, all else being equal, means better performance. Contrary to what all the conspiracy theorists would like to believe, Intel did not design the P4 the way they did just so they could tout huge clock-speed numbers.
Just to provide “FACTS” to back up Intel’s marketing ethics( or lack there of).
Google search for “p4 megahertz” produced these two articles.
“Intel, PC Makers Sued Over P4 Performance
Customers seek class action status in charge that vendors misrepresented the power of Intel’s top chip.”
http://www.pcworld.com/news/article/0,aid,104075,00.asp
Intel plans to dispel the megahertz myth which served so well
“Could it be that the megahertz myth is really about to die? The phrase megahertz myth points, of course, to the long standing awareness that mere clock ratings do not convey comparative performance between CPU architectures. ”
http://www.arstechnica.com/news/posts/1079200740.html
Google search for “p4 desgin philosophy” produced these two articles.
Intel’s Pentium 4 processor
“Intel’s design philosophy gives the NetBurst microarchitecture a very distinct character. For instance, clock-for-clock performance, or the number of instructions per clock (IPC) the chip can process, should be relatively low. But clock speeds ought to be very high.”
http://tech-report.com/reviews/2001q1/pentium4/
The Pentium 4 and the G4e: an Architectural Comparison
“When the Pentium 4 hit the market in November of 2000, it was the first major new x86 microarchitecture from Intel since the Pentium Pro. In the years prior to the P4’s launch the P6 core dominated the market in its incarnations as the Pentium II and Pentium III, and anyone who was paying attention during that time learned at least one, major lesson: clock speed sells. Intel was definitely paying attention, and as the Willamette team labored away in Hillsboro, Oregon they kept MHz foremost in their minds. This singular focus is evident in everything from Intel’s Pentium 4 promotional and technical literature down to the very last detail of the processor’s design. As this article will show, the successor to the most successful x86 microarchitecture of all time is a machine built from the ground up for stratospheric clock speed.”
http://www.arstechnica.com/cpu/01q2/p4andg4e/p4andg4e-1.html
The link to the last article.
Dude,
Your post reveals your newbieness to AE.
yet u blame the G5 for ur skillz (lack of)!
LOL, dude, whatever.
If u be getting compression artifacts, hey why not look at your compostion settings?
AE hasnt crashed in a hella long time for me, so as far as anecdotal testimony here at OS news, mines betterthan urs since I use AE on a mac ALL THE TIME!
My uncle bought a Dell with integrated video and he can’t even run a video game ona NEW 2.4ghz Pentium!?! Oh yeah Dell makes perfect sense dissing the Macs since their computers are worthless. Well expect that the new crop og upcoming G5s will simply BLOW AWAY the Intel cpus Dell depends on. Maybe the AMD 64 bit cpus can keep up a reasonable pace, but the writing is already on the wall for Intel, and Dell too, since Dell makes JUNK!
The “advertiser’s claim, ‘the world’s first 64-bit processor for personal computers,’ could reasonably be interpreted to apply to workstations.
Is it just me, am I hallucinating the words “for personal computers” in Apple’s claim? I don’t see how the claim could reasonably be interpreted to apply to workstations when it explicity says “for personal computers.”
Look at the empirical evidence regarding the heat and power consumed by the PPC970 and make up your own mind:
— Giant heat sinks
— 600W (and noisy) power supply for a system that only takes 2 low power drives
— 80W per CPU power system for 1.4Ghz CPU
— Lack of any ability to scale past 2Ghz over the past 9 months, even with 90nm chips
— G5 server runs only one (1) PPC970FX due to heat/power issues
— IBM running their blade servers at 1.6Ghz, slower than any Apple system
— lack of Apple G5 laptop, G5 iMac, etc., all due to high power/heat.
And on the other side, there is no empirical evidence showing that Apple’s claims of low heat/power are true.
Apple deserves to be spanked by the BBB for their outrageous lies regarding performance. I wonder who will spank them for all the other lies they tell. It looks like the entire G5 is a sham and that Apple made another dumb decision by deciding to stick with PowerPC.
better check your facts again. I have a 1.6 G5 at work we dont even know it is on even with all energy savings on best perfomance. The heat perfomance is still better then any INTEL or AMD chip. The heat sinks do not have 2 high powered fans on them like a PC. Open the case and look at one.
Also the G5 Xserve does come in a dual settup. You are plain wrong and have no idea what the G5 is about.
All you PC lovers should be very happy with APPLE. They have adopted HyperTransport that AMD originally started.
The first generation G5s are a great leap for APPLE. It puts the machines back close to the PC market. That is undenible. It might not be the “fastest” overall, and that is where APPLE seems to consistiently walk the line.
“My uncle bought a Dell with integrated video and he can’t even run a video game ona NEW 2.4ghz Pentium!?! Oh yeah ”
Your uncle bought a non-gaming machine and then complained about the game support? Please.
“Dell makes perfect sense dissing the Macs since their computers are worthless. Well expect that the new crop og”
A Dell Precision is a far more capable workstation than a G5. Hardly “worthless”.
“upcoming G5s will simply BLOW AWAY the Intel cpus Dell
depends on. Maybe the AMD 64 bit cpus can keep up a reasonable pace, but the writing is already on the wall for Intel, and Dell too, since Dell makes JUNK!”
Dell doesn’t make “junk” in the real world.
“Is it just me, am I hallucinating the words “for personal computers” in Apple’s claim? I don’t see how the claim could reasonably be interpreted to apply to workstations when it explicity says “for personal computers.”
A personal computer is a workstation is a personal computer. Marketing doesn’t define reality.
Perhaps in the RDF…
I suggest you al use another anology, you all seem not to be comparing the correct items.
Honda engine more effienct at producing power ( ratio of cubic displacement vs BHP)
A higher revving engine with a broad powerband ( lots of torque) can be considered a “better” engine
Porsche’s flat 6 is a proven engine/ not for its normally aspirated versions (well the original carreras is the expection) which only produce a rother modest amount of power for the size.
Porsche made its name for wel handleing and overall package ( braking being a very good +) Not until when you get into turbocharged engines does Porche really start to shine in the HP area (917/10 was I think the climax. http://www.supercars.net/garages/PeP/44v2.html)
SO really the G5 is the Honda and the P4 is the Porsche what kills me here is that I like my Apple and my Porsche
Give it up. The PIII at a the same clock speed is faster than the P4. Why oh why did the pipeline stages just increase every iteration. AMDs have always been faster at lower clockspeed than P4s.
Who cares how fast the chip is clock-for-clock (except in certain power-constrained situations like notebooks and blade servers)? What matters is total performance. For most of its life, the P4 has been faster than the P3 or Athlon.
Clockspeed tends to help integer performance greatly but not throighput or FP.
Absolutely wrong! Assuming that bandwidth isn’t a limiting factor, clock speed helps integer and floating-point performance identically, that is, linearly with increased clock speed. If you extend the pipeline to increase the clockspeed, however, then the performance increase of the extra clockspeed is offset by the performance decrease of the extra pipeline bubbles. This decrease is proportional to the average number of missed branches in the code. Integer code has an extremely high branch density, often 25% or more. Floating-point code often has a branch density close to 10% or less. Ergo, there are many more missed branches in integer code, thus the performance impact of the longer pipeline is higher.
It must be noted that the G5 itself uses the P4’s “long pipeline” approach to an extent. It has a 16-stage pipeline, longer than both the Opteron and the G4.
Please read the myrad of atricles on Pentium-Ms being faster than mobileP4s at lower clockspeed largely due to a smaller pipeline.
So what? The P4 was designed to be fast at high-clockspeed, not at a lower clock-speed. Notebooks are one of those areas where the clockspeed vs IPC trade-off doesn’t make sense, because power usage is important. That’s not the case with desktops.
I think intel plays more with marketting and clueless users than Apple. Helll they put “intel inside” logos on boxes till than most PC buyers couln’t care less what cpu ran in their boxes.
So? The CPU is often the most expensive component of the computer. Branding it makes sense. It should be noted that Apple *names* their machines after the CPU inside!
Let’s say inflight instrcutions = Number of cylinders and RPM = Clockspeed.
But that’s the thing. inflight-instructions have much less bearing on performance than clockspeed. All-else being equal, performance increases linearly with clockspeed. That’s not the case for the number of inflight-instructions. The Opteron, for example, keeps far fewer instructions in flight than the G5, but is competitive in performance at similar clockspeeds. The Athlon was the same compared to the P4. So in this analogy, clockspeed should be the number of cylinders, and RPM should be the number of in-flight instructions.
It also must be noted that the 215 vs 126 number is apprently not a comparison of the same things:
http://www.xbitlabs.com/articles/cpu/display/powerpc-g5_12.html
Third paragraph: “It seems like PowerPC970 processor cannot boast any exceptional features as far as the number of simultaneously processed instructions is concerned. Quite opposite to what Apple says.”
An Apple G5 has NINE fans Charles, including TWO big fans that blow air on the G5 heat sinks. It has a *600W* power supply. The G5 heat sinks dwarf anything Intel has ever produced.
It would seem you are fooled by low noise which is not the same thing as low heat / low power. Apple spent a lot of money making sure the G5 wasn’t as noisy as the G4 because Apple was on their usual path of facing a class action lawsuit over the G4 noise.
If you stop and *think* for a moment, why is the G5 not in a laptop or iMac by now? Uhhh. Yeah. Because heat and power issues. Which are all well documented if you read anything other than the official Apple liestream.
<<< The artists use G5s, sure, but the render farm is all Intel. For now. >>>
Actually they already migrated the renderfarm to G5’s. Considering Apple owns Shake I kind of saw that coming.
Heat is almost directly proportional to the power consumption of the chip. The p4 chips suck over 100W. the current g5s aren’t a whole lot better, but they do suck slightly less than 100W last i remembered.
here’s where you can grab your ankles. the new 970 chip that’s coming in Xserve g5s is the 90nm version of the chip, which takes 50W at 2ghz.
so that means 2 2ghz g5s in thos xserves takes less power (and thus less heat) than one p4. if you scale it down to 1.6ghz, you can safely get this at around 18W if i remember right. maybe less.
the 600W power supply is just apple’s over-engineering. 600W is so that the machine can supply enough power with a beefed up video card that requires external power (9800), 2 hard drives, one cdrom, 5pci-x cards, and 8 ram slots. that’s a fair amount more than most motherboards. and i’m sure pci-x takes more power than pci, considering it’s a higher bandwidth bus.
“to the guy who said buying macs was a monopoly.. yeah because if i buy a mac it can only run.. mac os x, and *nix. if i buy a pc it can only run… windows and *nix. “
If you want a Mac, where do you go? To Apple, and you pay for it. If I want a “PC” where do I go? Msi, Asus, Abit, Gigabyte, etc… I’m not an economics buff, but it sure looks like Apple has a much bigger monopoly than any of the x86 manufacturers. And the x86 manufacturers compete for consumers with lower prices. I like lower prices, and given that the G5 isn’t a clear winner, and that there is less community activity on the Apple side, and there are less products for OS X, I’ll stick with x86.
I know about Virtual PC and it looks decent, but if it’s anything like VMWare it will only be Ok for non-3d applications.
“the difference with the top one is os x is a shit ton better than windows, and i get native performance from my *nix and almost-perfect native os 9 software.”
Windows 2000 is fine for me. I’ll never get to use OS X until it’s worthwhile to buy the PowerPC architecture.
People always mention how OS X crashes on them, then experienced users describe how–because they know more–it never have crashes. Well, I’m an experienced user of Windows 2000 and it runs nicely, it runs forever, it’s secure (not be default), it doesn’t crash on me… it’s going to take a lot for me to want OS X.
“sigh you people. these processor wars are aways pointless. the g5 is plenty fast for most needs. you get it for the operating system, NOT for the hardware. i would rather purchase a $2000 dual 1ghz g4 than a POS windows box.”
As I’ve said, Windows is excellent for me. I like to play games, browse the web, and use the occasional application. I’m not going to buy a G5 for $X extra dollars when it’s worse at running games (if they’re even ported), and obviously not significantly better at running apps and browsing the web. So the processor wars aren’t pointless for me at all.