[dividend]

check this speed-measure:

Dual G4 1GHz Lightwave 3D Benchmark Comparison -
Just a quick note on Lightwave 3D v7.0b performance on a dual 1GHz G4 system running OS X. One of the popular LW benchmarks is the standard "raytrace" sample scene. Here's my score from the Dual 1Ghz, plus some other scores for the same test from a popular Lightwave 3D benchmarks site for other CPUs/Systems:

� Dual 1GHz G4: 121 seconds (120.7 seconds)
� Dual 800Mhz G4: 153 seconds
� Single P4 4 2.4GHz: 123 seconds
� Single P4 1.8 GHz: 187 seconds
� Single Athlon XP (@1.76GHz): 171 seconds
� Dual AthlonMP (@1.6GHz): 101 seconds
(another entry showed a dual MP @1.53Ghz reporting 92 sec.)
� Dual Pentium III 1Ghz: 150 seconds
� Quad Pentium III 850Mhz: 84 seconds
� Dual P4 Xeon 2.2Ghz (512K L2): 77 seconds

(I didn't even know a 2.2GHz Xeon existed.) Sources of the other systems is Chris's Lightwave Benchmarks site. (Tests were run according to the instructions there.)
��
taken from http://www.xlr8yourmac.com/

does this show quite clearly that the G4 is not THAT fast, although it is apparenlty very fast per mhz?

[larkost]

What you are seeing in these tests is two things:

The code probably was optimized for x86 processors. What is optimized for one processor family probably will not work as well on another family just because it is compiled.

The code almost certainly is not taking any advantage of the real strength of the G4, the Alitivec engine. The sort of code that it is probably using would be ideal for the FloatingPoint aspects of Alitivec/Velocity Engine, but they probably were in more of a hurry to just get something out...

This just points out a truism about benchmarking: it only tells you how fast each tes machine is at what you are doing, and is almost useless when you try to generalize...

Apple is not lying about speed, they are simply telling you how fast it works on PhotoShop and then letting you generalize that to everything else. In other words: Apple is letting you lie to yourself.

[Bigc]

Their lying because the Dual gigger is as fast as a 2.4 gig P4 x86 chip??

[11011001]

G4's kick major ass in RC5 cus of the altivec.

Anything really like this, computational stuff, tends to scream on G4's.

The graphic cards will also make a huge difference. I don't know, but would I be correct in assuming that you will not get the same performance out of a GeForce 3 made PC as a GeForce 3 Mac card? I mean they are probably built with PC hardware in mind, then made to work on a mac? I think I am probably wrong though. Enlighten me.

[ 03-31-2002: Message edited by: 11011001 ]

[michaelb]

It's a pretty well known fact that the G4s aren't the fastest things for raytrace rendering, particularly in the majority of 3D apps which have originated on x86.

But 3D on the Mac is still viable because final renders are slow on anything - that's why it takes about 50 PCs a few hours to render 30 seconds of Shrek's head or whatever - whereas the creation stage of modeling on the Mac is very usable.

With a decent 3D accelerator (GeForce 3/4 Titanium), G4s are perfectly good for 3D. LightWave and Maya are both going heavily after the Mac platform, and there are rumors that Apple themselves are looking to integrate a 3D solution into their video editing suite (or should that be "sweet").

[Leonis]

Yeap, 3D usually isn't Mac's strength.

But hey....we aren't just doing 3D

We have a lot more stuff to do on the Mac

[OverclockedHomoSapien]

Originally posted by dividend:
<STRONG>check this speed-measure:

Dual G4 1GHz Lightwave 3D Benchmark Comparison -
Just a quick note on Lightwave 3D v7.0b performance on a dual 1GHz G4 system running OS X. One of the popular LW benchmarks is the standard "raytrace" sample scene. Here's my score from the Dual 1Ghz, plus some other scores for the same test from a popular Lightwave 3D benchmarks site for other CPUs/Systems:

� Dual 1GHz G4: 121 seconds (120.7 seconds)
� Dual 800Mhz G4: 153 seconds
� Single P4 4 2.4GHz: 123 seconds
� Single P4 1.8 GHz: 187 seconds
� Single Athlon XP (@1.76GHz): 171 seconds
� Dual AthlonMP (@1.6GHz): 101 seconds
(another entry showed a dual MP @1.53Ghz reporting 92 sec.)
� Dual Pentium III 1Ghz: 150 seconds
� Quad Pentium III 850Mhz: 84 seconds
� Dual P4 Xeon 2.2Ghz (512K L2): 77 seconds

(I didn't even know a 2.2GHz Xeon existed.) Sources of the other systems is Chris's Lightwave Benchmarks site. (Tests were run according to the instructions there.)
��
taken from http://www.xlr8yourmac.com/

does this show quite clearly that the G4 is not THAT fast, although it is apparenlty very fast per mhz?</STRONG>

The G4 isn't so great at raytracing, yet based on these figures, it seems to keep up just fine with x86! WTF is the problem? So G4s can "keep up" on stuff that's optimized for x86's strengths, but on stuff that's optimized for the G4's strengths, the Mac wipes the floor with Wintels.

Seems to me that Macs are keeping up quite well, although I think that Apple needs to do better at marketing, and they need more MHz anyways. To gain marketshare Macs have to be BETTER than Wintels in every way, not just "keep up" with them.

[rampant]

Just being "almost as fast" with the very things they are good at for that price is not acceptable. Now if they were "almost as fast" for the stuff I do day to day, I'd have kept my ibook.

[dividend]

Ok, thanx for your input. to put things clear: i don't care that much about computerspecs, I hvae a Ti550 and 550 is more than i really need. I think a consistent interface and ease-of-use is more important than a few second here or there, though I guess it is different from those of you who work professionally with photoshop and fcp...

but i would note from the table that in this test, 1 Ghz G4 = 1.2 Ghz P4, which gives in THIS instance that PPC is 20% faster per mhz than x86. I have heard that cpus are good at some things and bad at others. Is there any others processor-demanding programme, except Photoshop, that is altivec-enabled and works for both plattforms? But I am surprised at what HUGE differences cpus can work at, but IF it is the case that the programme is not well written for the G4 then it is of course not of any particular interest.

How much is a 2.4 Ghz P4 anyway? Guess it is cheaper than a dual G4 Ghz? On a different note: some specs for my Ti550 and Ti500: (from: http://macspeedzone.com/html/hardwar...r_list_9.html)

PS, alitviec, PS, no alti iTunes encode
ti550 51 205 51
ti500 47 193 45

I have no idea what they did in photoshop, wouldn't understand it anyway, but it is interesting to see that more mhz can actually be slower. luckily, my Ti550 should be faster than ti400. guess it proves your points - speed is relative to what you do, what matters is what you need. But shouldn't Apple offer better 3D experience for you that care about it? What could be done?

[Mac Zealot]

We already are taking into account, of course, that most benchmarking programs are simply recompiled from x86 to ppc, and don't really take advantage of, or even use, the g4's special features at all!

This applies both ways.

[spb]

maybe its moto who we should be quizzing?

apple have been BADLY let down by those losers !

[slipjack]

Of Course Apple is lying (How do they get 15 Megaflops??)... but that doesn't mean that the new G4's don't rock, cuz they do.

A lame-o PC may be able to beat a Mac at some rendering tests, but the stats just posted I see as pretty positive. Macs are still faster Mhz over Mhz then PIII's, P4's, and AMDs... so what's to complain about?

Add in the fact that OS X truly rocks with dual processors on ALL tasks, not just rendering, Macs rule baby.

[dividend]

well, G4s are very good cpus, but it is not the case that one G4 is twice as fast as a P4 in the mentioned programme. G4s should be faster, and mhz do count! but they don't count for all!

[Graymalkin]

Jeez. The raytrace algorithsm LW uses do not use the G4's AltiVec unit, without that the FP performance of the G4 is pretty crappy. Apple isn't lying about the clock speed or anything, their marketing department just decided to say that their chips are little super computers because in certain operations the AltiVec processing unit increases throughput to over a GFLOP. The raytracer LW uses has little at all to do with the difference between the PPC and x86 processors, the code is largely similar if not identical. Remember you're comparing two 1GHz processors with a 2.2GHz processor. They're running agregately a similar number of cycles a second both pounding their float point processors. The G4 is a great processor but it isn't the all singing all dancing answer to all things wrong in the world.

What ought to be pointed out is the dual G4 system beat out the dual P3 system of the same core clock by 19%, which considering the G4 and P3 are similar processor generations (read: total core overhauls and relative release time) is impressive to me. The dual G4 system has only a 16% slower rendering time than a dual Athlon system with a 600MHz difference in core speed. Get some perspective.

[dividend]

Graymalkin: i don't udnerstand too much of cpus, but would it be possible to make it use altivec and would it be very usefull - is the benchmarks more a matter of benchmarking a poorly written piece of software (in relation to G4)?

[opago]

what do benchmarks matter anyway? does you mac do what it needs to do...mine sure does.

[siegzdad]

Originally posted by 11011001:
<STRONG>G4's kick major ass in RC5 cus of the altivec. </STRONG>

I completely agree. Here are the average results of process of my computers:

P-III 550 - 1,230,500 pps
Athlon XP 1800+ - 6,750,430 pps
Pismo 500 - 1,430,600 pps
867MHz Quicksilver - 8,732,650 pps

I alos agree that benchmarks don't mean a whole lot as you use the computer to do what you need to get done.

[r-0X#Zapchud]

Macs are faster in alot of things, and apple is not lying....

Macs are even faster in alot of Quake 3-tests, in those where configs are tweaked...

If the graphics are high-quality, equal on both machines of course, the PC will win. If you have a tweaked and less pretty config, the mac will win (yeah, both configs equal of course)...
ex. a Quicksilver 867 with GF2 MX equals an Athlon 1466 with GF2GTS...
(around 250 fps on both)

[Drakino]

The AMD 1.6 and 1.53 decrepency may be due to the ratings. It could have been an Athlon 1600+ MP (1.4 ghz I believe), and an Athlon 1800+ (1.53ghz)

Also, an official P4 at 2.4ghz has not been released. 2.2ghz is the fastest P4.

The code almost certainly is not taking any advantage of the real strength of the G4, the Alitivec engine.

http://macspeedzone.com/archive/5.0/...ications.shtml

That page clearly shows Lightwave as having AltiVec extensions. And I'm sure it has SSE2 extensions.

Yes the benchmarks do show that the PowerMac can somewhat keep up with PCs. But at a high cost.

Honestly, owning a Mac for me has been more for the experience, and not for the speed. Apple will continue to do the bakeoffs with an application that shows the best beformance gap. Any processor out there could be shown the best like this. It all comes down to what applications you are going to run, and the overall OS experience you want. Take any marketing benchmark with a deer lick grain of salt. I had such fun tearing apart that megahertz myth crap the Apple engineer presented at Macworld in July. Yes, megahertz isn't a peformance rating. But a longer pipeline chip is not going to run the same as a shorter one. There are other optimizations done to prevent the problems hilighted in that video. I could go on with details, but I'll save that for another day.

The graphic cards will also make a huge difference

In a Quake 3 benchmark, yes. In a render program benchmark like the one done above, no. NVidia keeps saying their cards are powerful enough to render Toy Story graphics (or recently Final Fantasy), but there are reasons all the render companies like ILM use huge servers, and not racks of AGP slots for GeForce 4 cards.

Macs are still faster Mhz over Mhz then PIII's, P4's, and AMDs... so what's to complain about?

The fact that Apple systems aren't advancing peformance marks as fast as PCs. Sure a G4 is faster mhz for mhz, but when the G4 stalls at 500mhz for an eternity while the P3 and P4 keep going ahead, there is a problem. And slapping 2 procs in an OS 9 machine is not the answer. Apple really needs to start using a mix of IBM and Motorola Power PC chips to keep them competing against each other, similar to how many PC vendors use both AMD and Intel, and if one slips, the other can be used to fill the gap.


No content is meant as a flame, an attack against the Mac, or PC. So don't take it as such please.

[nana2]

Originally posted by r-0X#Zapchud:
<STRONG>Macs are faster in alot of things, and apple is not lying....

Macs are even faster in alot of Quake 3-tests, in those where configs are tweaked...

If the graphics are high-quality, equal on both machines of course, the PC will win. If you have a tweaked and less pretty config, the mac will win (yeah, both configs equal of course)...
ex. a Quicksilver 867 with GF2 MX equals an Athlon 1466 with GF2GTS...
(around 250 fps on both)</STRONG>

umm, what resolution would that be at? At anything over 1024*768 (if not 800*600) with all the pretty stuff turned on, a GF2 GTS will rape a gf2MX.

But i guess there are some of us who run at 640*480 and below

[OverclockedHomoSapien]

I agree with those who say that considering price, Macs should wipe the floor with Wintels and then some. We pay a premium for Mac hardware and for the price an OS X Powermac G4 should be superior to Wintels in every way possible, not just the OS, but also the performance.

Another perspective of this current GHz gap puts things in a more favorable light for Apple. Consider this: The Pentium 4 is a relatively recent CPU design, and it just now has hit its prime and Intel is scaling it quickly and efficiently. There's plenty of life left to the Pentium 4 and the headroom is there.

In contrast, the G4 is an old chip design, at the end of its life span. For ages Motorola couldn't even scale the G4, but now its getting speed bumps at reasonable intervals, and in fact if Motorola had been scaling the G4 all along the way they are now, there wouldn't be any GHz gap. Thus it is nearly certain that Motorola has a new PPC chip nearly ready to secede the G4 as their premium desktop CPU, and Motorola will have applied all that they learned from the G4, particularly what they learned about scaling the seemingly unscalable 500 MHz G4. This new chip, call it the G5 if you like, will debut at around 1.5 GHz, and will hopefully have plenty of headroom for regular, frequent speed bumps.

But the key part is this: Apple has recently been making acquisitions suggestive of a thrust into the 3d animation market. Apple is clearly well suited for this market, with OS X and Apple's experience with digital video editing Apple could own this market in a few years...except for one small problem: Mac hardware.

It is no secret that Powermac G4s, even the dual 1 GHz beast, totally suck at 3d animation compared to Wintels. 3d professionals would like OS X, but they NEED raw hardware performance to get their work done, since 3d animation is one of the most taxing tasks for a computer. Without superior performance, Apple's chances in the 3d market are dubious, and we all know that except for a few brain farts like the Cube, Apple is managed by incredibly savvy computer guys who know what is needed to penetrate an artistic field like 3d animation.

So finally my long-winded post arrives at it's conclusion:
1. Apple is targeting 3d animation and design as their next area of domination.
2. Current Mac hardware is not up to the task of satiate professional animator's need for raw performance.
3. Therefore, Apple MUST have plans to introduce new hardware that dovetails with their entry into the 3d animation market.

What does this mean? Aside from a new motherboard with a screaming bus and DDR RAM, it means that a new PPC chip is probably set to be released, powering new Powermacs into a performance lead over Wintels.

I have no proof, only circumstantial evidence, but Steve Jobs is not stupid and if Apple attempted to enter the 3d animation market with their current Powermac hardware, it would be a monumental failure. Jobs knows that and it means that Apple is planning accordingly.

[nana2]

Originally posted by OverclockedHomoSapien:
<STRONG>I have no proof, only circumstantial evidence, but Steve Jobs is not stupid and if Apple attempted to enter the 3d animation market with their current Powermac hardware, it would be a monumental failure. Jobs knows that and it means that Apple is planning accordingly.</STRONG>

I will have to save this quote for later use

[spb]

2 stay in 2d?

[dandbj]

Just currious. If the Mac is a Supercomputer,and according to Apple, it is, and since it trails in performance to Wintell WinAMD boxes, Wouldn't it be fair to say that the average PC is also a Supercomputer as well?

Apple's marketting division has got to rethink its strategy.

As far as the original question goes, if Apple intends to seggest that Macs are generally faster than PC's (by a factor of 3, no less), then of course the're lying. They no better. They don't even believe that themselves.

The problem is Apple can't admit under any circumstances that their systems have less performance than PC's. to do so would be suicide for them. So the further behind they get in technology and speed, the bigger the lie they have to tell. Right now, the lie is so big, it has spiralled out of control. They are trapped by their own marketting claims and cannot now back down. Nor can they acknowledge the PC's progress and performance lead.

Apple has got to get out of this self imposed trap somehow. They simply can't keep it up forever. Soon, even the most die hard Mac fans are going to start questioning Apples claims as the original poster has already done. Apple could get away with it when comparing 333MH G3 to 400MH PII. But those days are long past.

How can they get out of this comparison trap? Any thoughts?

[opago]

Apple needs to rethink their strategy and be self intensive.


We compared Final Cut Pro 3 running on a Dual Gig power mac and Final cut pro running on a 2.2 ghz intel northword and benchmarks prove that the power mac is approxiamtely 1,000,000 times faster since the intel machine can not run Final Cut Pro 3.

Think Realistic.

[Stoopid]

Don't forget that the average mac user's mind is 20% faster than that of a pc user.

[nana2]

It doesn't matter if the Mac has the speed of a Pentium 133 because OS X doesn't get in the way of the user like Win2k does. Things are so smoooooth on OS X

[�]

Pc's and Macs are just things to be sold. You read what you they want you to read. They all have their specific advantage. Apple is lucky with a clientele that is hunting for facts that proof their assumptions. Apple is famous for its rumours. You dont see that in peeceeworld.