No apparent speed gains from new powermacs READ!

[OKComputer]

http://www.barefeats.com/pmddr.html


this link came from macnn

[billiam0878]

Well that's discouraging...:(

Thanks for the link though,

Bill

[topicolo]

cyberfunk is gonna have a anneurism when he reads this. heheh :)

[edesignuk]

Kinda makes me feel better that I didn't wait all those months and just got my "old" DP1Ghz when I did, it doesn't look like owners of the new ones will benifit greatly over me and my DP1Ghz QS, better lookin' too!

[King Cobra]

Don't forget, though, a bigger L3 Cache can do a bigger job. The older dual GHz Macs (original dual GHz PowerMac, Xserve) had a 2MB L3 Cache, while the new dual has a small 1MB L3 Cache (per chip in each case). The dual 1.25GHz has the 4MB total. But, for a price cut of $500 you really don't lose THAT much power.

Besides, you just can't ignore the 18 Gigaflops the fastest Mac can produce now. :)

[Mr Jobs]

oh well...:( still dual 867Mhz for only £1300 isn't bad

[ponyboy]

ouch

[firewire2001]

i guess the ddr ram isnt actually fully ddr and is just a gimic and costs more money to buy more of.. >.<

[Megaquad]

I like the screeming part...so,lets screem!

[Chryx]

Notice how all the tests were relatively small dataset tests...

could someone try a photoshop comparison on a 900MB tiff between the two, the faster bus speed on the new one should give it a clear lead over the old one, even though it has less L3 cache.

(I think photoshop is horrible as a cross platform benchmark, but it's good for like-with-like :)

[topicolo]

good job steve jobs! Stupid overclocked powermacs...:o

[Mr Jobs]

Originally posted by Megaquad
I like the screeming part...so,lets screem!

lets not get all worked up yet after all this is just one view (benchmark) others may find different results. after all benchmarks does seem to vary depending on who's doing them.

[drastik]

hmm, that's too bad, I'd love to see a speed increase. Oh well, at least its $400 dollars cheaper now, price is more in line with performance.

[Chryx]

Originally posted by firewire2001
i guess the ddr ram isnt actually fully ddr and is just a gimic and costs more money to buy more of.. >.<

The ram is fully DDR, the processor <-> memory controller bus isn't though, other hardware can however utilise the bandwidth that the processor can't.

These tests seem to be small dataset things that would seem to mostly operate within the L3 cache, or be CPU bound..

I say we wait for bigger/meaner tests that are more stressing on the overall system.

(Though I'm not entirely familiar with the photoshop test they ran, anyone care to elaborate on it?)

[Mr. Anderson]

I'm thinking this *new* machine is just a stop gap until they can get a new chip - the Power4 - in the mix. Its unfortunate that they're limited in throughput, but the new machines can be expanded beyond the older ones. So its not a total loss.

I'm going to wait around till the new generation comes out.

D

[ftaok]

So is he implying that the 1.25ghz PMac will use the 7470?

[OSeXy!]

I had thought the DDR PMac seemed more evolutionary than revolutionary, but those Barefeats numbers even call that into question.

I guess this rev. should be seen as a price-drop rather than a real power-increase (but maybe the 1.25 with its bigger L3 will give a significant boost?). Here's hoping Steve has something really big stuffed up his black jersey sleeve next MW... The natives are getting restless.

[ShaolinMiddleFinger]

Originally posted by drastik
hmm, that's too bad, I'd love to see a speed increase. Oh well, at least its $400 dollars cheaper now, price is more in line with performance.

I totally agree with you.

It looks like Photoshop and Bryce are slower, too....what gives? Now I have to re-think about buying the old one instead of the new Powermac.....

[ffakr]

Originally posted by dukestreet
I'm thinking this *new* machine is just a stop gap until they can get a new chip - the Power4 - in the mix. Its unfortunate that they're limited in throughput, but the new machines can be expanded beyond the older ones. So its not a total loss.
D

The new chip IS NOT A POWER4! Geez.
The blurb that everyone is hanging their hopes on explicitly says that IBM is showing a 64 bit POWERPC chip that was influenced by the Power4 line. If you buy a crappy 1980's Camaro Berneletta (sp?) that DOESN'T mean you are getting an italian sports car.

The Power4 uses a {slightly} different instruction set from the PowerPC. We don't want a processor from the Power line, we want a better PowerPC.

[Chryx]

Originally posted by ffakr


The new chip IS NOT A POWER4! Geez.
The blurb that everyone is hanging their hopes on explicitly says that IBM is showing a 64 bit POWERPC chip that was influenced by the Power4 line. If you buy a crappy 1980's Camaro Berneletta (sp?) that DOESN'T mean you are getting an italian sports car.

The Power4 uses a {slightly} different instruction set from the PowerPC. We don't want a processor from the Power line, we want a better PowerPC.

POWER4 = PowerPC-64 implementation (http://www.digit-life.com/articles/ibmpower4/)

All the material I've seen points to the POWER4 being a PPC-64 chip already, not a POWER chip (aside from the branding)

Including an IBM datasheet than I'm now struggling to find.

[OSeXy!]

Originally posted by ftaok
So is he implying that the 1.25ghz PMac will use the 7470?
No, I think he's saying (hoping?) the 7455 will be short-lived in the new PMac and that the next revision will use the same MoBo, but silently replaces the 7455 with the less restricting 7470. Sort of the opposite of the Yikes! routine (where it was 'right chip, wrong motherboard').

[elensil]

Well no "APPARENT SPEED GAIN" due to the new BUS is upsetting...

But I concider it GOOD NEWS for me....
The only reason i was considering buying a 1ghz over the base model was the new bus. Now i khow its completely useless, so i can just pay $1597 (with a student discount) and fell great about myself.

Thank you Apple for making no "APPARENT IMPROVEMENTS" in architecture since the last year.

[King Cobra]

All right. I'm not too sure if even half of you guys even paid attention to the article and the little "**" beside the L3 cache, but that can do a lot of help. Reread my post, if you are still worried.

ffakr, READ duke's DAMN POST MORE CAREFULLY before you go and slam him out the damn door. [Settling down...] He stated that the current new chip in the current new PowerMac is supposed to be, basically, a hold over until the newer chip, the Power4, becomes implanted in upcoming PowerMacs.

OSeXy, well put. I'm glad someone understands this whole damn ordeal. :)

elensil, you are a complete moron. Not to make this sound like a flame post, but...

Apple would have had put in faster chips and kept the old design if Motorola could produce the faster 7470 chips in both fast speeds and in suffecient quantity. But the [censored] there could not perform this duty. The result: Apple had no choice, but to overclock their current chips in order to get faster PowerMacs out the door.

Plain and simple: Apple is in the tightest spot you can think of with these PowerMacs. Once Apple gets the supposed Power4 going you probably won't have to worry about any of these "Yikes!" issues.

[elensil]

My post meant to bring up the spirits and offer a different point of view on the whole ordeal.

Ps I wonder why Queeny takes things so personaly:)

PPS I do not claim to be a hardware expert so corrections are welcome.

[King Cobra]

elensil, flat out and simple, this below statement sounded like you were blaming Apple.

>Thank you Apple for making no "APPARENT IMPROVEMENTS" in architecture since the last year.

If I sounded overwhelming, it was because Apple is not the one to blame. Remember, Motorola has been the supplier for Apple's chips and had to be aided once before. Now, Motorola just cannot pull off the faster chips. Apple didn't have a choice, but to increase their chip speeds.

Again, Apple is not to blame, but those :eek::eek:s at Motorola. :)

[Sun Baked]

It would be interesting to have these single action tests COMBINED with another test via scripts to see what happens.

I really am wondering what the differences in speed would be there when testing (Photoshop & Quake framerates) or (Photoshop & Storage) to better test the different system buses.

---

These single action tests were fine under OS 9, when single actions were the norm. But OS X is a multi-threaded system - let's see tests that better reflect this.

[barkmonster]

I'd like see some benchmarks that really show how the DDR makes a difference. Gaming has got to be one, although I want to see what barefeats come up with. As an AGP graphics card can share the system RAM, that's 2.7Gb/s for the graphics card to play with. Also, audio applications rely on ram to buffer the audio streams, the DDR and ATA100 will definately show a boost to hard disk recording, it should also take more strain off the cpu meaning if you're using a track limited DAW like Logic Audio Silver or Protools LE you should get more plug-ins running on at once.

The 900Mb photoshop file idea is also a really good one to try.

I think crunch tests that are bound to mainly the L3 cache arn't too relevant, I want to see something that's pushing 100s of Mbs of RAM around and taking advantage of the faster ATA100 controller.

Saying that, the entry level is the fastest, cheapest, most highly specified powermac I've ever seen, strip out the modem and add a DVD-R/CDRW combo drive and you've got a pretty cool system for about £1,500!

[giovanni]

come on come on come on - sorry dudes but my tolerance for stupidity has hereby ended !

what idiots !!!!

have these Powermac written Motorola or Apple on them ????
These are "APPLE COMPUTER, Inc." computers and APPLE COMPUTER ONLY AND UNIQUELY IS RESPONSIBLE for the squallid and pathetic performance improvement.

I hope I don' need to say more on this. Just stop this Motorola crap.

Am I disappointed with APPLE ? yes. Am I pissed at Apple ? no. But some of you just don't want to blame mother god Apple so now you blame it on Motorola. Get a life ! may be ?

for me I will simply wait until they come up with a decent high end top of the line desktop, possibly up to date with current rather than historical technologyl. Meanwhile I will buy a 17" iMac.

[cyks]

I'm sorry King, but I'm going to have to join elensil on this one... it's Apple's fault.

Granted, Moto is the one holding Apple back- but Apple is letting them. Moto isn't the only chip maker out there...and the no speed gains a year later only hurts Apple. Not Moto.

The general public doesn't know or care who makes the processors...and if you were to ask them, I'm sure they'd think Apple made them themselves.

Apple is the one in the public's eye...they're the one that are trying to sell computers... they're the ones that chose Moto to make thier current chips. If Moto can't keep up with demands then Apple should have dumped them long ago.

[Chryx]

Originally posted by giovanni

have these Powermac written Motorola or Apple on them ????
These are "APPLE COMPUTER, Inc." computers and APPLE COMPUTER ONLY AND UNIQUELY IS RESPONSIBLE for the squallid and pathetic performance improvement.

I hope I don' need to say more on this. Just stop this Motorola crap.


Apple can only build a computer as fast as the parts they have available to them...

Motorola are currently the snag upon which Apple's performance is stuck.

(Fully paid up member of "SWITCH TO IBM YOU FOOLS" here)

[giovanni]

yes yes yes, but it still remains and forever will be Apple's problem and responsibility. Assuming (ASSUMING) it is indeed Mot that is slowing down progress (sorry what did Steve Jobs call it, "innovation" !?!?!?! ahahahahahahha) then it may be the case that Apple simply committed a huge strategic mistake. For instance noone better than Apple would have realized such difficulties in the "evolution" of the PPC, so Apple should have hedged itself by having other suppliers ready (business 101).

I guess Steve Jobs mentioned "innovation" without specifying he was referring to software only. there has been no hardware innovation whatsoever in ages at Apple Inc.

I would like to say, I still love Apple, but I am ready to spend on truly innovative hardware.

[bousozoku]

Originally posted by drastik
hmm, that's too bad, I'd love to see a speed increase. Oh well, at least its $400 dollars cheaper now, price is more in line with performance.

Exactly! It may not be faster than the old one, for what we've seen so far, but it's cheaper than the old one and it performs better than the old middle machine.

Either way, that's positive!

[elensil]

I admit that Flame wars are the last thing we need at these forums. I didn't mean to start one.

I know it's Motorola's fault and Apple DID a lot to improve the low end performance.
What i really want to see is some benchmarks (read test of any sort) to see the performance of the 867 dual.

[Hemingray]

Originally posted by elensil
I admit that Flame wars are the last thing we need at these forums. I didn't mean to start one.

It's okay, flamewars are bound to happen in a thread that people care so much about. This is very important to us, because we've been waiting for DDR for so long now, only to find out that it's another half-baked Yikes situation.

I'm not worried, I know Apple will update the chips as soon as Motorola gets their lousy act together. Apple is indeed a tight spot as always, and we can most likely thank Motorola once again for putting them there.

[rice_web]

That BareFeats test doesn't even show the specs of each machine used! How much RAM was present in each machine?

Overall, I'd say that the new dual-1GHz machine is faster, and will seem significantly faster than the old dual-1GHz machine once Jaguar is shown (yeah, the Barefeats test used Jaguar). I want to see games in these benchmarks, not iTunes tests. Who cares if iTunes is a bit faster? Unless one has a CD to import and burn before heading to work, I don't see the use (I can't imagine any corporations using iTunes).

And for god's sake, let's wait for the dual-1.25GHz benchmarks before we're too harsh.

And... if you can't stand these machines, I'd wait until November, when I expect the 7470 at 1.33GHz.

[Gelfin]

Why would you test the performance benefit of a new memory architecture using four processor-bound tasks?

[tjwett]

i knew it. overclocked crapola. still waiting for something good. i think the best thing to do now is find an "old" dual 1ghz for cheap. these tests are enough proof for me. i trust barefeats.

[topicolo]

Originally posted by Chryx
Originally posted by giovanni

have these Powermac written Motorola or Apple on them ????
These are "APPLE COMPUTER, Inc." computers and APPLE COMPUTER ONLY AND UNIQUELY IS RESPONSIBLE for the squallid and pathetic performance improvement.

I hope I don' need to say more on this. Just stop this Motorola crap.


Apple can only build a computer as fast as the parts they have available to them...

Motorola are currently the snag upon which Apple's performance is stuck.

(Fully paid up member of "SWITCH TO IBM YOU FOOLS" here)

Totally agree. Motorola has f-ed up apple one time too many. Time for apple to give moto the finger and stick it to 'em :eek: Apple needs new processors and they need them soon. Let's hope IBM can get good yields on their new chips and that they can ramp up production soon.
Either way, nothing's going to save apple this quarter :(

[rice_web]

Oh, and don't forget about overclocking! I bet the dual-1GHz machine could go to 1.33GHz or so. That cooling system that the new PowerMacs use is, I dare say, wicked.

[King Cobra]

Guys, I started a pure debating thread for who really is to blame for this, in case you are interested in debating there.

http://forums.macrumors.com/showthread.php?s=&threadid=9625

[Sun Baked]

Remember people these are typical OS 9 style testing procedures, these tests may use SMP capabilities better than OS 9 - but it's still a single test - single action.

OS X allows you to do MORE at the same time, let's combine these tests into something real world for OS X.

The new machine is supposed to allow better response when you combine heavy HD access with number crunching and QE.

They (Apple) hinted at better system response, let's test it.

[Rower_CPU]

Originally posted by rice_web
That BareFeats test doesn't even show the specs of each machine used! How much RAM was present in each machine?

Overall, I'd say that the new dual-1GHz machine is faster, and will seem significantly faster than the old dual-1GHz machine once Jaguar is shown (yeah, the Barefeats test used Jaguar). I want to see games in these benchmarks, not iTunes tests. Who cares if iTunes is a bit faster? Unless one has a CD to import and burn before heading to work, I don't see the use (I can't imagine any corporations using iTunes).

And for god's sake, let's wait for the dual-1.25GHz benchmarks before we're too harsh.

And... if you can't stand these machines, I'd wait until November, when I expect the 7470 at 1.33GHz.

Why would they release another new PM in November?

When has Apple ever revved a product 3 months after a new one was released?
Hell, the 1.25 GHz doesn't ship for another month or so...you think they'll say: "We just started shipping the 1.25 GHz model to all of you who wanted the fastest PowerMac available, and we sure appreciate your $3000+...but we decided to release a newer, faster PowerMac now." :confused:

[xelterran]

as soon as apple dump moto the better i think a 64bit Power4 chip would look VERY nice in the powermacs - hopefully they release it sometime soon (january?)

[rice_web]

Why would they release another new PM in November?

Because they have to. :D

But seriously, Intel releases new processors all the time. AMD releases processors all the time. Both don't give a damn about the customers that just bought their CPUs last week.

Apple could do the following (and quietly):

dual-900MHz
dual-1.1GHz
dual-1.33GHz

This would be such a seemingly small update, but with the 7470, the increase in speed would be very significant at the same clock speed. I wouldn't be surprised if we saw this in November, with something to the tune of a G5 in March (the G5 may still be a year or so away, but I'm betting on March right now).

[Sun Baked]

Originally posted by Rower_CPU
When has Apple ever revved a product 3 months after a new one was released?

Can you say Yikes!

The G4 400 lived 1999.08.31 to 1999.10.13
The G4 350 lived 1999.10.13 to 1999.12.02

They liked it so much they did it twice and back-to-back.

[ftaok]

Seriously,

The new dual-1ghz machine is definately not an overclocked 7455. The new dual-1.25ghz machine is not available yet, so how does anyone know if it's an overclocked chip???

Maybe Motorola has gotten the 7455 up to 1.25ghz, but not yet at the yields that Apple needs.

Or, maybe the dual-1.25ghz will feature the 7470 (with full DDR support - assumed).

Either way, it wouldn't be an overclocked 1ghz 7455. Apple would be stupid to sell machines like that. If they did, they'd be looking at a lot of warranty claims when anything processor related breaks.

Think about it.

[Nipsy]

Howzabout a new owner and an old owner run a set of Apple's benchmarking tools, so we can have some results not from a guy running macros with a stop watch.

ftp://ftp.apple.com/developer/Tool_...rformance_tools

It would be really nice if we could rally up a New 1GHz DP under 6c115, and an old 1GHz DP under 6c115...

[Sun Baked]

http://arstechnica.infopop.net/OpenTopic/page?a=tpc&s=50009562&f=8300945231&m=9600944035

[Wry Cooter]

For the Barefeats tests to mean anything (And he should seriously retest), he should be running more than one of these homebrewed benchmarks simultaneously.. they are processor bound tasks! Of course they aren't going to show any difference (Maybe one or two of them can access multiprocessing at all, but I don't think so, maybe the bryce render, I'm not so sure).

Where the speed of the new machines will show up is in multithreaded tasks. Ripping MP3s *WHILE* simultaneously doing Photoshop flipflops.

So run more than one of the benchmarks simultaneously on the earlier Dual and the New Dual, both running Jaguar, then we will see ....

In any case, I agree that these particular machines are a stopgap, a stepping stone, a small speedbump in the road for something that can use a front side bus later. But this is typical for towers, if you look at the Blue and White>Yikes>Sawtooth>Digital Audio>SuperDrive moves.

Marketing and readiness of parts will probably keep another bump from showing until next summer though, although really, January or May would be better.

[demars]

...that such a big deal is being made over these particular benchmarks. After thinking about the nature of these particular benchmarks, I could have predicted the results. And the guys at Bare Feats should know this too, if they are going to be in the business of comparing performance they should be aware of what aspect of the system is being tested by any particular benchmark.

First of all, a special note to all of you seem to think that these benchmarks prove the DDR implementation is worthless; remember, independent of the DDR, the new machine does have a 25% faster system bus, so it can definitely talk to the RAM 25% faster than the old machine, and as the Bare Feats article mentioned this didn't seem to make a difference either. Does this clue you in that there is something about these benchmarks that makes them less than ideal to measuer the effective RAM bandwidth?

Let's take the Altivec Carbon Fractal benchmark first, that yielded identical times across the board. I'm the author of a fractal program myself and I know what's involved. This doesn't surprise me at all.

What a fractal program does, in simple terms, is walk across a set of points, an image if you will. for each pixel in the image, it repeats a simple set of arithemetic calculations over and over again; for some pixels this can be repeated hundreds of times. So, for instance, you might have something like:

newx = x*x - y*y + cx
newy = 2xy + cy
x = newx
y = newy

in a loop repeated many times for each point.

For something like this 99.999% of the time is spent by the CPU doing calculations, not accessing memory; in fact the code and a large part of the image will reside in the cache throughout the calculation, and I'm not even talking about the L3 cache, most of the time it is accessing the registers and the on-chip cache!

Well, you can see why this would be a very good test of CPU power and floating point speed in particular. You couldn't come up with something that was more CPU intensive. Or, less memory bandwidth intensive. This kind of algorithm spends a miniscule amount of the total time left reading from or writing to RAM anyway; if you made the memory bandwidth infinite it would barely have made a difference.

The other benchmarks suffer from a similar criticism. These things are all quite CPU intensive, although not quite as single-mindedly as fractal calculation, and they spend a small amount of time moving data to the internal cache compared to the time they spend crunching the data once it is in the cache. Apple likes to show off CPU-intensive software like PhotoShop filters, at least if it has been Altivec-optimized, because it shows off the speed of the Altivec unit. But for that very reason, these tasks don't really test the performance of the rest of the system, but JUST the CPU. The only thing that would speed them up is increasing the clock speed of the CPU.

So why bother with anything other than speeding up the CPU clock speed? Because we don't spend our entire life running fractal programs and PhotoShop filters. It is great to have these things speeded up, because they are so time consuming, but most of our time we are doing other things too, like web browsing, text editing, etc.

So where would we see the 25% system bus increase? Well, if you are a PhotoShop user, maybe something as simple as scrolling through a very large image would seem faster. That is a task that requires a lot of data being shifted through memory and relatively little CPU execution time. That's the sort of thing that might not seem much faster if you had a 4GHz CPU but would seem faster with a higher memory bandwidth.

And where would we see the improvement from Apple's DDR implementation, since it doesn't increase bandwidhth to the CPU? Well, since it does allow other parts of the system to do DMA without necessarily stealing bandwidth from the CPU (or from other parts of the system) maybe you would see it when, say doing a big file transfer and simultaneously watching a full screen movie without seeing a glitch in the movie. (Probably you can do that on the old dual 1GHz too, but if you keep adding on real time tasks you can bring any system to its knees, perhaps the DDR implementation in the new machines would push that point further out). So the DDR may enable the system to have more things happen reliably in real time when some of them don't require intense CPU intervention.

That seems like a worthwhile improvement to me, but how do you benchmark it? There probably is a way to do it, maybe running several benchmarks simultaneously, some of which exercise the I/O; but just running CPU-intensive benchmarks is not going to do it.

(I think I might send a copy of this to Bare Feats and see what they have to say).

[rice_web]

Thanks for a nice explanation, Dennis.

[Grokgod]

Who is to blame here, well that is obvious.

APPLE is responsible for their company and what it produces and NO ONE else!

Moto is a pain in the buttocks to be sure, but Apple is not their company.

And Moto is NOT telling Apple the manner in which to HYPE machines that will SO obviously be uncovered by those technically ableminded.

Its sad that Apple is in a TECH HOLE, whatever the reasons!

Yet, it's Apple's responsibility to get OUT of it and back into the computer game that they make such HUGE claims of being the best at!

I think that the new HUGE heat sink and fan speaks for itself.
The slow FSB speaks for itself!
The prices increases for no speed inprovements speaks for itself!
The slowwitted case with addon cd doors and vents for overclocked chips...
Speaks for itself!

There isnt any confusion here , this is Apple at its worst, for whatever reasons.
Be it Moto or Jobs or the newest big breasted secretary that has reduced Jobs mind to mush, it doesnt matter.
The realities are still there.

So I will say what I have said before many times...

We are not stupid, Apple.
We will not buy OLD hardware at premium prices.
Get it together and bring out the GOOD stuff.

Although, I think that for those of us that "Refuse to Lose" and move to the Pcheese world.

There is a decent hole here and thats the pricing on the old Gigers!
If you can find an older 1 giger for maybe 1200 or something like that.
I say thats the BUY!

[ImAlwaysRight]

Originally posted by Rower_CPU
When has Apple ever revved a product 3 months after a new one was released?


Originally posted by Sun Baked


Can you say Yikes!

The G4 400 lived 1999.08.31 to 1999.10.13
The G4 350 lived 1999.10.13 to 1999.12.02

They liked it so much they did it twice and back-to-back.
Umm, the 1.25 GHz won't be shipping until the last week of Sept, first week of Oct. Then you've got to fill the backlog, so it'll be mid/late Oct. before Apple gets caught up. And you'd like a 1.33MHz bump in Nov? Sorry, don't think so. We're stuck with these speeds until at least the remainder of 2002. January/February next year will hopefully bring our next speed bump.

[dongmin]

Originally posted by Grokgod
Who is to blame here, well that is obvious.

APPLE is responsible for their company and what it produces and NO ONE else!


Dude, this guy just got through explaining how the Barefeats benchmarks measure mostly just the CPU performance and you go off ranting about how Apple is to blame for everything. Apple does not design or manufacture the CPU, Moto does.

Of course, that doesn't mean that Apple is free from blame. But let's be a little more rational here. Let's withold judgement about the new tech Apple has put into the new PowerMacs UNTIL someone does some decent benchmarks that measures system-wide performance. Certain benchmarks, like the ones done by Xinet, show some substantial performance advantages of the Xserve architecture.

I think the biggest problem with Apple is that it has essentially one supplier for its chips. I fault them for not having developed enough of a contingency plan. We need a healthy competition like you see between AMD and Intel. Hopefully, IBM with its new PowerPC chip will give Apple some more options. But who knows when that'll happen...

[TheCat]

Excellent!! this is exactly wot i was saying yesterday, but you've written it out far better than i could :)

[jadam]

yeah really good job explaining all of that :D much better than i could.

[peterh]

Originally posted by dongmin


Dude, this guy just got through explaining how the Barefeats benchmarks measure mostly just the CPU performance and you go off ranting about how Apple is to blame for everything. Apple does not design or manufacture the CPU, Moto does.

Of course, that doesn't mean that Apple is free from blame. But let's be a little more rational here. Let's withold judgement about the new tech Apple has put into the new PowerMacs UNTIL someone does some decent benchmarks that measures system-wide performance. Certain benchmarks, like the ones done by Xinet, show some substantial performance advantages of the Xserve architecture.

I think the biggest problem with Apple is that it has essentially one supplier for its chips. I fault them for not having developed enough of a contingency plan. We need a healthy competition like you see between AMD and Intel. Hopefully, IBM with its new PowerPC chip will give Apple some more options. But who knows when that'll happen...

I guess Apple could be faulted for not having enough money to convince Moto to only build chips to its specifications. The fact of the matter is that most CPU architectures have very little competition, and the x86 only has it because of a mistake Intel made in the 1980s. Believe me, Intel has no desire to compete with AMD. Incidentally when everyone moves to 64bit, this competition goes by by, IA64 and x86-64 cannot be made binary compatible.

Given the trials and tribulations of the G4 CPU arch, what do you want Apple to do, in the short term? Switching to the x86 is not feasible, considering the effort it would take to make Carbon portable. Apple must look at the long term to fix this problem. Apple's only choice is to spend money on R&D, both internally and externally. To do this Apple needs good gross margins, i.e. in the 20%-30% range. Note: gross margins != net margins. This is not the case for peoples like Dell, or the PC divisions of HP & IBM, which do much less R&D. Given the lead times required, even if Apple started spending this money 2.5 years ago, you would only start seeing the stuff about now, i.e. the motherboard controller from the XServe/PMG4.

If you want to argue about these facts, you need to get a book on economics and try and understand it first. If you want cheeper kit from Apple, you have to give up the future of the Mac

[skunk]

Yeah, right. What he said.....

[MacCoaster]

Originally posted by peterh
Given the trials and tribulations of the G4 CPU arch, what do you want Apple to do, in the short term? Switching to the x86 is not feasible, considering the effort it would take to make Carbon portable. Apple must look at the long term to fix this problem. Apple's only choice is to spend money on R&D, both internally and externally. To do this Apple needs good gross margins, i.e. in the 20%-30% range. Note: gross margins != net margins. This is not the case for peoples like Dell, or the PC divisions of HP & IBM, which do much less R&D. Given the lead times required, even if Apple started spending this money 2.5 years ago, you would only start seeing the stuff about now, i.e. the motherboard controller from the XServe/PMG4.
About that Carbon portable, do you mean Cocoa? NeXT was once running on x86, and Darwin is running on x86. So it's quite possible. All Apple would need to do is recompile the OS with x86 as host processor. They might do this when Mac OS X is very mature and has a lot of Cocoa applications. These Cocoa apps would just have to be recompiled with VERY little code change. This is synonymous to Linux on x86 then later ported to the PowerPC. You can compile most x86 Linux programs on PowerPC provided you have the libraries. See, KDE runs on Linux/PPC just as fine as it does on x86.

[peterh]

Originally posted by MacCoaster

About that Carbon portable, do you mean Cocoa? NeXT was once running on x86, and Darwin is running on x86. So it's quite possible. All Apple would need to do is recompile the OS with x86 as host processor. They might do this when Mac OS X is very mature and has a lot of Cocoa applications. These Cocoa apps would just have to be recompiled with VERY little code change. This is synonymous to Linux on x86 then later ported to the PowerPC. You can compile most x86 Linux programs on PowerPC provided you have the libraries. See, KDE runs on Linux/PPC just as fine as it does on x86.

No I mean Carbon. I am talking short term, i.e. the next 6 to 18 months. While Cocoa is quite portable, some of it exists on other platforms as we speak, it is called Webobjects. A lot of carbon is tied to the PPC. It could be made portable, but it would take a fair amount of effort, and there is no guarantee that it would be a simple recompile, (That is also the case with a lot of Altivec based code). The nice thing about Cocoa is that the libraries can change and one doesn't necessarily have to rework the code. Porting MacOS X especially the Quartz layer to X86 is not a trivial undertaking, since it is designed for PPC and more specifically Altivec usage (Quartz did not exist under NexT).

My post was meant to address those who want Apple to fix the PowerMac deficiencies now, it just isn't a reasonable desire, kind of like the desire to fly across the ocean twice as quickly for half the price. While I can't guarantee that Apple is working on remedies to these problems, I assume they are.

In the end it makes little difference, if you don't want to buy a Mac, don't; however, if enough of you swear off Macs & Apple, there is no way they will improve to anywhere near what you desire.

[giovanni]

dudes,
have you seen the updated benchmarcks at barefeats.com. Well, take a look and stupid dreaming.

[peterh]

Originally posted by giovanni
dudes,
have you seen the updated benchmarcks at barefeats.com. Well, take a look and stupid dreaming.

I am not sure what you are trying to say, but the results are to be expected. According to several posts both at Macrumors and Arstechnica, and more importantly, most of these tests are either CPU bound, or are memory constrained. Incidentally, I think the memory throughput graph tells it all. You will notice the the DDR main memory is actually faster than the SDR memory, this would be attributable to the faster FSB. None of these tests will show any of the benefit available to the peripherals. I would love to see the 1K windows bloom test for comparison.

As an aside you will notice the the random64 DDR scores are identical to the SDR if not a little lower. This is actually to be expected as the latency for DDR SDRAM is slightly higher than that for the SDR SDRAM.

A final comment of the BareFeats test. They could not be called statistically accurate, hell he doesn't even use enough test points to apply the mean value theorem. As it stands he only uses one data point, the best, for his observations. Any statistician would laugh him off the web. Right now the Standard deviation for all of his tests is [B] infinity[\B]. This is because all benchmarks are not perfectly repeatable. Hell even if those values were perfectly repeatable. There is ~34% chance that the XServe and SDR PMac photoshop 7 scores are identical. This is not what I would call statistically significant

[P-Worm]

Just because the dual 1 Ghz machin'e tied is not a bad thing. What that means simply is that the new mid range machine can keep up with the old high end machine. That's right the new MID RANGE can perform up there with the HIGH END. When do you see that in the PC world? Does Intel or AMD make it so a not as good computer can match their best from before? If anything we should all be VERY excited to see what this dual 1.25 Ghz can do!:cool:

[Pauls]

Does anyone know how much of a bottle neck SD RAM was with the current G4 chip? Is it true that the LSU on the G4 is limited to one load or one store per cycle? If so the major improvement of ram access would be tied to higher CPU clock speeds.

Also the debate about moving to other CPU's cannot be tied solely to performance. Apple uses the one chip architecture for its servers, desktops and laptops, and software is tied to this chip architecture. What would apple use for its notebooks if it moved to AMD for example, their chips have around twice the number of transistors as the G4 and so drawing more power. (The G4 powerbook cuts it with most other laptops in performance). Would we be prepared to pay the price to replace all our software for recompiled versions, so soon after OSX debut?
Would we be prepared to have different software for different chip architectures to accomodate everyones specialist needs or do we prefer the 'one OS for all' that we enjoy on the Mac.

What encourages me about the new line of towers is that the bus architecture has been redesigned to make the tower a real pro range. This architecture can't possibly find it's way into the iMac or eMac, they're too compact. I'm in no hurry for faster clocks, Moto or IBM have to make the grade sooner or later or else Apple is sunk, if they move to Intel they will have to make the same decisions as everyone else on that roadmap. I buy our macs on a 3-4 year cycle, so every time I upgrade, the Macs are about 3-4x faster, not just a percentage boost. I'm happy.

[amnesiac1984]

Originally posted by P-Worm
Just because the dual 1 Ghz machin'e tied is not a bad thing. What that means simply is that the new mid range machine can keep up with the old high end machine. That's right the new MID RANGE can perform up there with the HIGH END. When do you see that in the PC world? Does Intel or AMD make it so a not as good computer can match their best from before? If anything we should all be VERY excited to see what this dual 1.25 Ghz can do!:cool:

Thankyou!!! Just what I was going to say! If you want to see how much this machne has improved over it's predecessor you need to do some comparisons with the single 933 mhz, My bets are on this new machine opening several cans of whoopass on it!

[ivangough]

I dunno much about the tech jargon or the over hyped speed tests, but I do know this..I am a professional musician working with a G4 running Logic Audio, and right now my blood is boiling. I am sick and tired of watching the guys in PC land kick my audio arse with machines that cost less than half what I pay. They get loads more firepower, and when running Logic with loads of plug ins and soft synths running, they win by a LONG mile! If Jobs expects me to keep forking over the hard earned cash to buy overpriced boat anchors he's got another thing coming. Don't get me wrong, I love my G4, but like alot of my friends in the music industry I'm growing ever impatient at waiting for the kick butt macs I keep being promised. We need faster machines NOW! The only thing keeping me in a G4 is the fact that Apple bought Emagic. Sorry for the rant peeps, I'm calm now...

Ivan Gough.