[[[Intriguing. Are you saying even 'big' developers perhaps even Adobe aren't optimizing apps like Photoshop to do 'several' things and are instead lazily allowing the minimum amount of work to be done?]]]
Negative. I *know* for a fact that Adobe is taking full advantage of the G4. Chris Cox is a sharp guy and is well ahead of the curve with respect to crossplatform optimizations be it Wintel or Mac. In fact, he is one of the leading AltiVec-specific programmers on the planet. Anyway, here is an interesting snippet taken from another message board in response to some arrogant poster.
[[[Excuse me?
Adobe (that's me) and Intel spend more time and effort optimizing Photoshop for Windows than for the Macintosh!
The AltiVec code in Photoshop was done by one Adobe engineer, mostly in his spare time. (those of you who know my work habits can stop laughing now)
The MMX and SSE code in Photoshop were done by 2 Adobe engineers and at least 6 different Intel engineers, analyzed by several Intel performance experts, code reviewed by yet more Intel engineers, etc.
Intel and AMD are constantly working with Adobe to improve Photoshop's performance on *their* systems.Yes, Photoshop 4.0 had some bad code -- but put the blame where it's due: MSVC++. The code was even worse on the Macintosh at the time. And lots of people have made claims about Photoshop's code quality. But none of them have matched Photoshop's quality and been any faster (well, outside of myself the hardware vendors working with me). In short: you haven't got the faintest clue what you're talking about.]]] -Chris Cox\t
[[[the G4's potential to do 'more' things per cycle\t]]]
That's just it, developers have to be willing to take advantage of the features the G4 has to offer. They need to add as much parallelism as possible. It isn't clear that they are doing that. I suspect that in many instances the code was simply ported and tweaked a little to run on the Mac. Code that was originally designed to take advantage of x86. It's possible that the programmers are more familiar with x86 and therefore tend to be x86-biased or x86-centric with respect to their coding habbits/techgnique.
[[[Will it start at 2 gig? ]]]
Who really knows? It's only a number used mostly by marketing departments. It isn't important. What's important is the "work done".
I'm certainly not qualified to say if what you've said is true or not. I can say that it makes quite interesting reading. One thing I can speak a bit on is advertising and marketing. Simply put, if a 800mhz G4 has say the equivalent speed of a 1.6GHZ P4 then why not do what AMD has been doing for years. Simply name the speed of G4 processor the equivalent speed to the correlating P4 /Athlon cpu (of course I understand that differing cpu architectures do not make this cut and dry). AMD has not run into any public perception problems as a result of naming their cpu speeds in an equivalent P4 speed naming scheme. In fact often AMD purchasers have looked at this naming scheme as getting more for less. Come on Apple, there are people out there that would love to use your systems if only they could be led to the water (and they have not been led to the water). I've done what I could personally to turn my own coworkers on to the mac (1-TiBook & 1- Imac Convert). It's kinda like buying a house. Nice house but....Location, Location, Location. Now i'll let the experts continue.
[[[You said "The PowerPC has a lower clockspeed - sure. But it also has lower latencies, larger
caches, more functional units, more pipelining,..."I thought the PPC had less pipelining than the pentium?- Mark]]]
Well, actually, that was Chris's comments. However, I think you are confusing "pipelining" with "pipes" or rather "depth or the stage of the pipeline. Or perhaps Chris was referring to the fact that the current crop of G4's have increased levels of address pipelining. I don't think it was meant to imply that the G4 has a "deeper" pipeline or a pipeline with more stages.
From Apple's website:
[[[The performance advantage of the PowerPC G4 starts with its data pipeline. The term
"processor pipeline" refers to the number of processing steps, or stages, it takes to
accomplish a task. The fewer the steps, the shorter -- and more efficient -- the pipeline.
Thanks to its efficient 7-stage design (versus 20 stages for the Pentium 4 processor) the
G4 processor can accomplish a task with 13 fewer steps than the PC. You do the math. ]]]
I don't want to divert this thread too much but ithas just occurred to me no one is really expecting the G5 next week, why. Because last year SJ had no good new after the hype got out of control it was a compleate anti-climax. Huge damage was done to Apple sales as people held off expecting great things and held off and held off.
So Apple has a huge problem with these and other boards because false expectations lead to declining sales, nobody wants to buy last years model. Secondly Steve like supprises.
Is it possible that there are several people posting to the community delibertly negative but informed information just to suppress expectation. I remember last year one very knowledgeable post (From whome I can't remember) saying that there was no way we would see the G5 until 2003. It was really well written and cogently argued.
However, go back to the top of this thread, then look at some of the info we have seen from Dorsal. Add the fact the the G4 has made it into just about the whole range of Apple products, and I begin to wonder if this mad conspiricy (Sp) is possible. It just may be that we will see the G5 and it will be a supprise, there is no doubt about that.
I expect Jaguar to ship with any new hardware announced next week, regardless of wether it is G5 or not. What ever new Powermac is announced, I think they will be shipping in 3-4 weeks with Jaguar, delivery of hardware will be in August and that IS late summer.
None of this really matters. What really matters is that Apple's top end machine gets its hat handed to it by Intel & AMD machines that cost 1/2 as much in 3D apps. Look <a href="http://www.blanos.com/benchmark/"; target="_blank">here</a>. Try to find a single benchmark where the dual gig Mac gets more the 1/2 or 1/3 the score the top machine gets-- then realize that in many cases, the top scoring machine costs 1/2 what the Mac does.
It might make you guys get all warm and fuzzy inside thinking that all Apple's performance woes would be solved if software developers would properly optimize their code, but think of it like this: In 20+ years of "Microcomputers", PCs, and shrinkwrapped software, the overwhelming majority of developers haven't made any attempt to optimize their code. That's not going to change just to haul Apple's fat out of the fire. Apple needs machines with more bandwidth, and higher clock speeds, YESTERDAY.
BTW-- It's rather disingenuous to site RC5 as proof of the G4's superiority. RC5 simply takes advantage of the fact that Altivec provides a barrel shifter operation. RC5 (and BLAST) are abberations, and shouldn't be taken as an inidicator of the G4's potential.
If you want to learn something about the problems the G4 has with bandwidth limitations, hop on over to <a href="http://www.arstechnica.com"; target="_blank">Ars Technica</a> and do a search in the OpenForum for a poster named BadAndy. He's an embedded programmer who works with Altivec.
[[[None of this really matters. What really matters is that Apple's top end machine gets its hat handed to it by Intel & AMD machines that cost 1/2 as much in 3D apps. Look here. Try to find a single benchmark where the dual gig Mac gets more the 1/2 or 1/3 the score the top machine gets-- then realize that in many cases, the top scoring machine costs 1/2 what the Mac does.]]]
I believe that I already addressed this in an earlier post. Still I could take something like SETI and distributed.net (see above) and say the same thing. It has to do with code.
[[[the overwhelming majority of developers haven't made any attempt to optimize their code. ]]]
This statement is complete BS. Forget what Chris Cox has stated. EVERY meaningful software developer optimizes the code they write.
\t
[[[BTW-- It's rather disingenuous to site RC5 as proof of the G4's superiority. RC5 simply takes advantage of the fact that Altivec provides a barrel shifter operation. RC5 (and BLAST) are abberations, and shouldn't be taken as an inidicator of the G4's potential.]]]
Why not? They are in fact real-world.
[[[If you want to learn something about the problems the G4 has with bandwidth limitations, hop on over to Ars Technica and do a search in the OpenForum for a poster named BadAndy. He's an embedded programmer who works with Altivec.]]]
Yeah... ArsTech is the place... lol And BadAndy is the expert LOL! Funny I never see him on any of the AltiVec forums I visit.
[quote] I believe that I already addressed this in an earlier post. Still I could take something like SETI and distributed.net (see above) and say the same thing. It has to do with code. <hr></blockquote>
No, you didn't address it. People who need a 3D rendering program aren't going to write it themselves, they're going to buy it off the shelf. Here's a real world example of Apple's machines getting the shit beat out of them. Deal with it.
[quote] EVERY meaningful software developer optimizes the code they write. <hr></blockquote>
Must not be many meaningful software developers, then. Most developers are not like Chris Cox. As soon as the code works, they're done with it. Damn near nobody takes the time to optimize.
[quote] Why not? They are in fact real-world. <hr></blockquote>
Right-- real world abberations. The performance gains are not representative of what a typical parallelizable algorithm would gain by being optimized for Altivec.
[quote] Yeah... ArsTech is the place... lol And BadAndy is the expert LOL! Funny I never see him on any of the AltiVec forums I visit. <hr></blockquote>
Uh, care to substantiate your criticism? It really doesn't help your case just to dismiss it out of hand... Did it occur to you that he may go by a different screen name on other forums?
Ed M, thank you for some great reading. Particulary...the Cox speaks out.
Mr. Typewriter..? I put that on my 'favourites' list. I can't say why...because I'd be at risk of closing this excellent thread... (The best thread I've see on 'insider...)
Keep up the good work.
Lemon Bon Bon
PS. Ed M...some of the links point to the G5's integer and fpu's being built from the ground up. Okay. How many integer heads? How many fpu? Better branch prediction? I kinda expect better mhz than we've had over the last couple of years...but even so...it seems that there'll be more emphasis on 'throughput' than 'bruteput'...
Y'know the 'leaked' G5 spec scores a while back...do you have a link to them? Do you have any inkling to their validity?
Article says products should be available in early 2003.
"The specification defines both a single differential link in each direction between devices and support for ganging four links together for higher throughput applications. On the system level, designers can connect parallel and serial RapidIO devices through switches without using special bridging functions."
Parallel RIO is useful for motherboard device connects but the serial RIO looks like it is adapted for longer lengths. This may be an application of Gigawire; to connect several machines together in clusters.
Guy's you're talking about caching, and bandwidth like there the things that make the computer work. this is only part of the equation. The reason the G4 is faster (twice as fast) is because of processes per cycle. That's why Apple talks about gigaflops instead of megahertz. Lets make this simple shal we. you have 2 chips an Intel Pentium 4 at 100 MHz, and a G4 at 100 MHZ (I know these chips don't come in 100MHz speeds, but bear with me) the Pentium 4 works 5 processes per cycle while the G4 does 10 processes per cycle. This means that in 1 second the P-4 did 500 processes, while the G4 did 1000 processes. this means that the G4 is superior. The P.C. industry has increased it's bandwidth to make up the difference in processing speeds. It doesn't speed up a computer that much, but when configured with faster video and audio cards they make a big difference. However; this technology can sometimes bog down a system, and shorten their life because of the additional wear and tear on their electronic components. This isn't the best solution. Apploe needs to find a way to widen the bandwidth on the motherboard, without degradeing the componenets. Kind of your happy medium, and that's ehat they have done, and will probably continue to do.
<strong>I don't want to divert this thread too much but ithas just occurred to me no one is really expecting the G5 next week, why. Because last year SJ had no good new after the hype got out of control it was a compleate anti-climax. Huge damage was done to Apple sales as people held off expecting great things and held off and held off.
So Apple has a huge problem with these and other boards because false expectations lead to declining sales, nobody wants to buy last years model. Secondly Steve like supprises.
Is it possible that there are several people posting to the community delibertly negative but informed information just to suppress expectation. I remember last year one very knowledgeable post (From whome I can't remember) saying that there was no way we would see the G5 until 2003. It was really well written and cogently argued....
I could be wrong.</strong><hr></blockquote>
I don't expect the next generation Mac processor before MWSF at the earliest, and if the rumors are true about Apple canceling the Motorola version not until next year's MWNY. But I will be surprised if Steve doesn't come out and 1) announce what Apple's plans are regarding next gen processors/architectures and 2) give the public a preview of what's coming a year or so down the road. He may be between a rock and another rock; he doesn't have a choice. If he doesn't people are just gonna bitch and moan, but worse, there will just be too much speculation on the company's future. If he does there will be naysayers who come out and connect poor future Mac sales to these announcements, but I can't sewe how telling its customers that there are faster machines in the pipeline a year down the road will cost them biz. It could have the opposite effect; people will realize that if they need a new mac now there's no point in waiting for the proverbial G5, and just go out and buy the best Mac that's currently available.
[quote] The reason the G4 is faster (twice as fast) <hr></blockquote>
Groan <img src="graemlins/oyvey.gif" border="0" alt="[No]" /> Sorry, but the G4 is seriously bandwidth choked... If it had a QDR bus like the P4 does, then maybe it could approach its full potential. Who knows? If the G4 had a 400MHz (effective) bus, then perhaps we'd see what it's capable of. But it doesn't. The simple fact is that the G4 spends a lot of time waiting to get data to crunch-- because of its slow bus. If it had a higher bandwidth bus, we'd see a good size increase in performance.
[quote] (I know these chips don't come in 100MHz speeds, but bear with me) <hr></blockquote>
What you've failed to acknowledge is that the P4 has never been offered at a speed as low as the fastest G4 that is available now. In fact, the fastest P4 has a clock speed 2.5+ times what the fastest G4 is available at. To top it all off, a P4 machine running at that speed is going to cost half what Apple charges for the dual gig Powermac.
[quote] this means that the G4 is superior. <hr></blockquote>
Yes, on a cycle-by-cycle basis, the G4 is more efficient, but that's not the point. There are P4 machines available which take any machine from Apple to school, for less money. They make up their lower efficiency per cycle with high clock rates. You started your response by complaining that we weren't considering "the whole equation". You can't just focus on processor efficiency, either. Clock rates, bandwidth and a slew of other factors go into it as well.
I urge you to look at the link typedesigner provided. It should convince you of just how much Apple's offerings lag PCs.
<strong>I urge you to look at the link typedesigner provided. It should convince you of just how much Apple's offerings lag PCs.</strong><hr></blockquote>
I looked over the performance stats on that link and got the opposite reaction. The dual 1-GHz G4s hold their own pretty well and even come out ahead on tests like RC5 and Photoshop (only beaten by dual processor Athons and P4s).
[quote]<strong>What you've failed to acknowledge is that the P4 has never been offered at a speed as low as the fastest G4 that is available now. In fact, the fastest P4 has a clock speed 2.5+ times what the fastest G4 is available at. To top it all off, a P4 machine running at that speed is going to cost half what Apple charges for the dual gig Powermac.</strong> <hr></blockquote>
I think you need to check out <a href="http://www.aapltalk.com/shootouts/"; target="_blank">System Shootouts</a> before saying how much 'cheaper' P4/Athlon machines really are.
[quote]I looked over the performance stats on that link and got the opposite reaction. The dual 1-GHz G4s hold their own pretty well and even come out ahead on tests like RC5 and Photoshop (only beaten by dual processor Athons and P4s). <hr></blockquote>
To each his own, I guess. The dual gig PM gets its nuts chopped off, sauteed, and served to it in Maya, Lightwave, Cinebench, Mathmatica, and SETI. Personally, I'm a heavy Lightwave user, so that's what's important to me.
[quote]I think you need to check out System Shootouts before saying how much 'cheaper' P4/Athlon machines really are. <hr></blockquote>
Uh, I think you need to find a site that at least trys to be impartial for price comparisons. I mean, c'mon-- aapltalk is claiming that a dual gig Powermac is categorically equivalent to a 3GHz P4, and they haven't updated the comparisons for $3k+ machines (and others) since the Powermacs were introduced back in Feb. Don't you think some things may have changed since then?
"To each his own, I guess. The dual gig PM gets its nuts chopped off, sauteed, and served to it in Maya, Lightwave, Cinebench, Mathmatica, and SETI. Personally, I'm a heavy Lightwave user, so that's what's important to me."
<strong>To each his own, I guess. The dual gig PM gets its nuts chopped off, sauteed, and served to it in Maya, Lightwave, Cinebench, Mathmatica, and SETI. Personally, I'm a heavy Lightwave user, so that's what's important to me.</strong><hr></blockquote>
The Maya benchmark isn't very valid. They used an older version on the Mac that didn't even support dual processors versus a newer PC version that does support dual processors. Not exactly scientific there.
The chart for Lightwave doesn't seem to make any sense. It shows the G4 at 100 and the P4 at 161, but looking at the data that was supposedly used to generate the charts shows the following test times. 127, 16, 8, and 10 seconds for the G4 and 127, 8, 4, and 7 seconds for the P4. Not exactly sauteed here is it?
Cinebench also doesn't make use of dual processors, so you would expect the G4 to fall behind.
They only have scores for an 800 MHz G4 on the Mathematica test. Who knows what a dual 1 GHz would really do?
Comments
Negative. I *know* for a fact that Adobe is taking full advantage of the G4. Chris Cox is a sharp guy and is well ahead of the curve with respect to crossplatform optimizations be it Wintel or Mac. In fact, he is one of the leading AltiVec-specific programmers on the planet. Anyway, here is an interesting snippet taken from another message board in response to some arrogant poster.
[[[Excuse me?
Adobe (that's me) and Intel spend more time and effort optimizing Photoshop for Windows than for the Macintosh!
The AltiVec code in Photoshop was done by one Adobe engineer, mostly in his spare time. (those of you who know my work habits can stop laughing now)
The MMX and SSE code in Photoshop were done by 2 Adobe engineers and at least 6 different Intel engineers, analyzed by several Intel performance experts, code reviewed by yet more Intel engineers, etc.
Intel and AMD are constantly working with Adobe to improve Photoshop's performance on *their* systems.Yes, Photoshop 4.0 had some bad code -- but put the blame where it's due: MSVC++. The code was even worse on the Macintosh at the time. And lots of people have made claims about Photoshop's code quality. But none of them have matched Photoshop's quality and been any faster (well, outside of myself the hardware vendors working with me). In short: you haven't got the faintest clue what you're talking about.]]] -Chris Cox\t
[[[the G4's potential to do 'more' things per cycle\t]]]
That's just it, developers have to be willing to take advantage of the features the G4 has to offer. They need to add as much parallelism as possible. It isn't clear that they are doing that. I suspect that in many instances the code was simply ported and tweaked a little to run on the Mac. Code that was originally designed to take advantage of x86. It's possible that the programmers are more familiar with x86 and therefore tend to be x86-biased or x86-centric with respect to their coding habbits/techgnique.
[[[Will it start at 2 gig? ]]]
Who really knows? It's only a number used mostly by marketing departments. It isn't important. What's important is the "work done".
--
Ed M.
You said "The PowerPC has a lower clockspeed - sure. But it also has lower latencies, larger
caches, more functional units, more pipelining,..."
I thought the PPC had less pipelining than the pentium?
- Mark
Junkmein
caches, more functional units, more pipelining,..."I thought the PPC had less pipelining than the pentium?- Mark]]]
Well, actually, that was Chris's comments. However, I think you are confusing "pipelining" with "pipes" or rather "depth or the stage of the pipeline. Or perhaps Chris was referring to the fact that the current crop of G4's have increased levels of address pipelining. I don't think it was meant to imply that the G4 has a "deeper" pipeline or a pipeline with more stages.
From Apple's website:
[[[The performance advantage of the PowerPC G4 starts with its data pipeline. The term
"processor pipeline" refers to the number of processing steps, or stages, it takes to
accomplish a task. The fewer the steps, the shorter -- and more efficient -- the pipeline.
Thanks to its efficient 7-stage design (versus 20 stages for the Pentium 4 processor) the
G4 processor can accomplish a task with 13 fewer steps than the PC. You do the math. ]]]
--
Ed
So Apple has a huge problem with these and other boards because false expectations lead to declining sales, nobody wants to buy last years model. Secondly Steve like supprises.
Is it possible that there are several people posting to the community delibertly negative but informed information just to suppress expectation. I remember last year one very knowledgeable post (From whome I can't remember) saying that there was no way we would see the G5 until 2003. It was really well written and cogently argued.
However, go back to the top of this thread, then look at some of the info we have seen from Dorsal. Add the fact the the G4 has made it into just about the whole range of Apple products, and I begin to wonder if this mad conspiricy (Sp) is possible. It just may be that we will see the G5 and it will be a supprise, there is no doubt about that.
I expect Jaguar to ship with any new hardware announced next week, regardless of wether it is G5 or not. What ever new Powermac is announced, I think they will be shipping in 3-4 weeks with Jaguar, delivery of hardware will be in August and that IS late summer.
I could be wrong.
It might make you guys get all warm and fuzzy inside thinking that all Apple's performance woes would be solved if software developers would properly optimize their code, but think of it like this: In 20+ years of "Microcomputers", PCs, and shrinkwrapped software, the overwhelming majority of developers haven't made any attempt to optimize their code. That's not going to change just to haul Apple's fat out of the fire. Apple needs machines with more bandwidth, and higher clock speeds, YESTERDAY.
BTW-- It's rather disingenuous to site RC5 as proof of the G4's superiority. RC5 simply takes advantage of the fact that Altivec provides a barrel shifter operation. RC5 (and BLAST) are abberations, and shouldn't be taken as an inidicator of the G4's potential.
If you want to learn something about the problems the G4 has with bandwidth limitations, hop on over to <a href="http://www.arstechnica.com"; target="_blank">Ars Technica</a> and do a search in the OpenForum for a poster named BadAndy. He's an embedded programmer who works with Altivec.
[edit-- fixed link]
[ 07-09-2002: Message edited by: Gamblor ]</p>
<snip>In short: you haven't got the faintest clue what you're talking about.]]] -Chris Cox
<hr></blockquote>
I like this Chris guy. In which forum did you find him comments?
PS: Interesting reading. Thanks. Always nice to learn something new.
I believe that I already addressed this in an earlier post. Still I could take something like SETI and distributed.net (see above) and say the same thing. It has to do with code.
[[[the overwhelming majority of developers haven't made any attempt to optimize their code. ]]]
This statement is complete BS. Forget what Chris Cox has stated. EVERY meaningful software developer optimizes the code they write.
\t
[[[BTW-- It's rather disingenuous to site RC5 as proof of the G4's superiority. RC5 simply takes advantage of the fact that Altivec provides a barrel shifter operation. RC5 (and BLAST) are abberations, and shouldn't be taken as an inidicator of the G4's potential.]]]
Why not? They are in fact real-world.
[[[If you want to learn something about the problems the G4 has with bandwidth limitations, hop on over to Ars Technica and do a search in the OpenForum for a poster named BadAndy. He's an embedded programmer who works with Altivec.]]]
Yeah... ArsTech is the place... lol And BadAndy is the expert LOL! Funny I never see him on any of the AltiVec forums I visit.
Ignore this troll.
--
Ed M.
No, you didn't address it. People who need a 3D rendering program aren't going to write it themselves, they're going to buy it off the shelf. Here's a real world example of Apple's machines getting the shit beat out of them. Deal with it.
[quote] EVERY meaningful software developer optimizes the code they write. <hr></blockquote>
Must not be many meaningful software developers, then.
[quote] Why not? They are in fact real-world. <hr></blockquote>
Right-- real world abberations. The performance gains are not representative of what a typical parallelizable algorithm would gain by being optimized for Altivec.
[quote] Yeah... ArsTech is the place... lol And BadAndy is the expert LOL! Funny I never see him on any of the AltiVec forums I visit. <hr></blockquote>
Uh, care to substantiate your criticism? It really doesn't help your case just to dismiss it out of hand... Did it occur to you that he may go by a different screen name on other forums?
Anyway, check out <a href="http://arstechnica.infopop.net/OpenTopic/page?a=tpc&s=50009562&f=8300945231&m=8790959504"; target="_blank">this thread at Ars.</a> He gives a pretty good intro to Altivec coding, and talks about many of the problems with Moto's implementation over the course of the thread.
[quote] Ignore this troll. <hr></blockquote>
Yeah, God forbid someone inject a little reality into your fantasy world.
Ed M, thank you for some great reading. Particulary...the Cox speaks out.
Mr. Typewriter..? I put that on my 'favourites' list. I can't say why...because I'd be at risk of closing this excellent thread...
Keep up the good work.
Lemon Bon Bon
PS. Ed M...some of the links point to the G5's integer and fpu's being built from the ground up. Okay. How many integer heads? How many fpu? Better branch prediction? I kinda expect better mhz than we've had over the last couple of years...but even so...it seems that there'll be more emphasis on 'throughput' than 'bruteput'...
Y'know the 'leaked' G5 spec scores a while back...do you have a link to them? Do you have any inkling to their validity?
Cheers.
<a href="http://www.eetimes.com/sys/news/OEG20020709S0015"; target="_blank">http://www.eetimes.com/sys/news/OEG20020709S0015</a>;
Article says products should be available in early 2003.
"The specification defines both a single differential link in each direction between devices and support for ganging four links together for higher throughput applications. On the system level, designers can connect parallel and serial RapidIO devices through switches without using special bridging functions."
Parallel RIO is useful for motherboard device connects but the serial RIO looks like it is adapted for longer lengths. This may be an application of Gigawire; to connect several machines together in clusters.
Guy's you're talking about caching, and bandwidth like there the things that make the computer work. this is only part of the equation. The reason the G4 is faster (twice as fast) is because of processes per cycle. That's why Apple talks about gigaflops instead of megahertz. Lets make this simple shal we. you have 2 chips an Intel Pentium 4 at 100 MHz, and a G4 at 100 MHZ (I know these chips don't come in 100MHz speeds, but bear with me) the Pentium 4 works 5 processes per cycle while the G4 does 10 processes per cycle. This means that in 1 second the P-4 did 500 processes, while the G4 did 1000 processes. this means that the G4 is superior. The P.C. industry has increased it's bandwidth to make up the difference in processing speeds. It doesn't speed up a computer that much, but when configured with faster video and audio cards they make a big difference. However; this technology can sometimes bog down a system, and shorten their life because of the additional wear and tear on their electronic components. This isn't the best solution. Apploe needs to find a way to widen the bandwidth on the motherboard, without degradeing the componenets. Kind of your happy medium, and that's ehat they have done, and will probably continue to do.
:cool:
[ 07-09-2002: Message edited by: Maverick ]
[ 07-09-2002: Message edited by: Maverick ]</p>
<strong>I don't want to divert this thread too much but ithas just occurred to me no one is really expecting the G5 next week, why. Because last year SJ had no good new after the hype got out of control it was a compleate anti-climax. Huge damage was done to Apple sales as people held off expecting great things and held off and held off.
So Apple has a huge problem with these and other boards because false expectations lead to declining sales, nobody wants to buy last years model. Secondly Steve like supprises.
Is it possible that there are several people posting to the community delibertly negative but informed information just to suppress expectation. I remember last year one very knowledgeable post (From whome I can't remember) saying that there was no way we would see the G5 until 2003. It was really well written and cogently argued....
I could be wrong.</strong><hr></blockquote>
I don't expect the next generation Mac processor before MWSF at the earliest, and if the rumors are true about Apple canceling the Motorola version not until next year's MWNY. But I will be surprised if Steve doesn't come out and 1) announce what Apple's plans are regarding next gen processors/architectures and 2) give the public a preview of what's coming a year or so down the road. He may be between a rock and another rock; he doesn't have a choice. If he doesn't people are just gonna bitch and moan, but worse, there will just be too much speculation on the company's future. If he does there will be naysayers who come out and connect poor future Mac sales to these announcements, but I can't sewe how telling its customers that there are faster machines in the pipeline a year down the road will cost them biz. It could have the opposite effect; people will realize that if they need a new mac now there's no point in waiting for the proverbial G5, and just go out and buy the best Mac that's currently available.
Groan <img src="graemlins/oyvey.gif" border="0" alt="[No]" /> Sorry, but the G4 is seriously bandwidth choked... If it had a QDR bus like the P4 does, then maybe it could approach its full potential. Who knows? If the G4 had a 400MHz (effective) bus, then perhaps we'd see what it's capable of. But it doesn't. The simple fact is that the G4 spends a lot of time waiting to get data to crunch-- because of its slow bus. If it had a higher bandwidth bus, we'd see a good size increase in performance.
[quote] (I know these chips don't come in 100MHz speeds, but bear with me) <hr></blockquote>
What you've failed to acknowledge is that the P4 has never been offered at a speed as low as the fastest G4 that is available now. In fact, the fastest P4 has a clock speed 2.5+ times what the fastest G4 is available at. To top it all off, a P4 machine running at that speed is going to cost half what Apple charges for the dual gig Powermac.
[quote] this means that the G4 is superior. <hr></blockquote>
Yes, on a cycle-by-cycle basis, the G4 is more efficient, but that's not the point. There are P4 machines available which take any machine from Apple to school, for less money. They make up their lower efficiency per cycle with high clock rates. You started your response by complaining that we weren't considering "the whole equation". You can't just focus on processor efficiency, either. Clock rates, bandwidth and a slew of other factors go into it as well.
I urge you to look at the link typedesigner provided. It should convince you of just how much Apple's offerings lag PCs.
<strong>I urge you to look at the link typedesigner provided. It should convince you of just how much Apple's offerings lag PCs.</strong><hr></blockquote>
I looked over the performance stats on that link and got the opposite reaction. The dual 1-GHz G4s hold their own pretty well and even come out ahead on tests like RC5 and Photoshop (only beaten by dual processor Athons and P4s).
[quote]<strong>What you've failed to acknowledge is that the P4 has never been offered at a speed as low as the fastest G4 that is available now. In fact, the fastest P4 has a clock speed 2.5+ times what the fastest G4 is available at. To top it all off, a P4 machine running at that speed is going to cost half what Apple charges for the dual gig Powermac.</strong> <hr></blockquote>
I think you need to check out <a href="http://www.aapltalk.com/shootouts/"; target="_blank">System Shootouts</a> before saying how much 'cheaper' P4/Athlon machines really are.
To each his own, I guess. The dual gig PM gets its nuts chopped off, sauteed, and served to it in Maya, Lightwave, Cinebench, Mathmatica, and SETI. Personally, I'm a heavy Lightwave user, so that's what's important to me.
[quote]I think you need to check out System Shootouts before saying how much 'cheaper' P4/Athlon machines really are. <hr></blockquote>
Uh, I think you need to find a site that at least trys to be impartial for price comparisons. I mean, c'mon-- aapltalk is claiming that a dual gig Powermac is categorically equivalent to a 3GHz P4, and they haven't updated the comparisons for $3k+ machines (and others) since the Powermacs were introduced back in Feb. Don't you think some things may have changed since then?
<strong>What was the best thing before sliced bread?
</strong><hr></blockquote>
Pre-split coconuts?
YOUCH! :eek:
Gamblor,
Lemon Bon Bon
<strong>To each his own, I guess. The dual gig PM gets its nuts chopped off, sauteed, and served to it in Maya, Lightwave, Cinebench, Mathmatica, and SETI. Personally, I'm a heavy Lightwave user, so that's what's important to me.</strong><hr></blockquote>
The Maya benchmark isn't very valid. They used an older version on the Mac that didn't even support dual processors versus a newer PC version that does support dual processors. Not exactly scientific there.
The chart for Lightwave doesn't seem to make any sense. It shows the G4 at 100 and the P4 at 161, but looking at the data that was supposedly used to generate the charts shows the following test times. 127, 16, 8, and 10 seconds for the G4 and 127, 8, 4, and 7 seconds for the P4. Not exactly sauteed here is it?
Cinebench also doesn't make use of dual processors, so you would expect the G4 to fall behind.
They only have scores for an 800 MHz G4 on the Mathematica test. Who knows what a dual 1 GHz would really do?