Dual-Core PowerMacs, New G4 PowerBook, and G5 Mini?

[AidenShaw]

From more than twice as fast as a dual to not quite as fast as a single

how much more powerful would a dual dualcore 2.5 be than the current dual 2.7?

It would be very application dependant.

For a scalable parallel application like LINPACKD (used to rank the top 500 supercomputers) it should be more than twice as fast - doubling the size of the cache should make up for the 8% loss in CPU MHz. Most embarrassingly parallel applications would hit the 4x mark as well.

For a purely memory-bound application (like the STREAMS benchmark), it would probably be about the same speed - unless the memory controller changes to use faster DIMMs.

For applications that fit well in the 1 MiB L2 cache of the 970MP, but which spill out of the 512 KiB on the 970FX, it could be much, much faster than 4x (but the useful applications that fit this profile are probably few to none).

The question really becomes interesting when you ask "how much faster is application XYZ". Here it depends on the nature of the application (can the algorithm be decomposed into 4 or more threads?) and the skill of the engineers (did they find every possible place to exploit parallelism?).

Some applications run the same speed on a dual as they do on a single. Moving those to a quad won't help at all. Others run somewhat faster on a dual - those might improve very slightly on a quad.

Only the applications that run nearly twice as fast on the dual are good candidates for a quad - and only if those were written to use more than two threads.

A lot of the "MP aware" apps are really "dual aware" - they were written to decompose the algorithm into two threads only.
___________

All in all, the quads will be most beneficial for farming applications (embarrasingly parallel), and a few true SMP-aware apps.

For the rest of us, only the power users who want run several heavy apps at once (rip CDs while burning DVDs while rendering in FCP) will get much benefit.

 

[StinktOldC]

so do you think it would be noticable in an app like logic pro 7?All those AU fx and instrument plugs are very hungry!I've only had my dual 2.7 six weeks

 

[Freg3000]

I don't think anyone ever claimed that it was a lock but you can logically deduce that it probably will happen. Would Apple introduce the 17" PowerBook anywhere but MW or WWDC? Or what about the new iMac? Or the Mac Mini or the revamped 15" PowerBook? Yes I was wrong about the iPod being released at Mac Expo but in both of those cases it was delayed for a reason: Cancer causing jobs not to be there for the expo and 9|11 causing the expo to be canned. *blows whistle* I call extenuating circumstances in those cases.

I don't think its a stretch to say that Jobs will introduce new x86 hardware at MW with a shipping date anywhere from that day to a couple months later. They did it before with the first G5 PowerMac. They did it with the 17" PowerBook when that came out too. What was it? 3 months later or something?
Apple wants the biggiest audience to see their wares. There isn't going to be anything bigger then MW '06 and any hardware that would be introduced a couple months later at a "special event" is already up and running by the time MW comes around. If they are still stuck in development in January '06 Apple is in deep ****. So anything that could be introduced a few months later could be introduced at MW.


And Again I will quote Jobs:

keynote transcripts

What part of this isn't clear that Apple plans to be shipping x86 systems BY WWDC '06. Which means introing it sometime between NOW and 6.5.06. Do people really think Apple is going to let a major device out the door without an event? This is Jobs we are talking about. The guy gets off on standing infront of everyone wowing them. MW '06 guys. Mark my words, bookmark this page. Whatever.

Extenuating circumstances. He he.

My overriding point in all of these discussions isn't about the logic that one can use to determine a MWSF '06 launch. I'll admit, it makes a good deal of sense, even though I am not nearly as convinced as you are. My beef with this thread are the people who are going to advance your well-thought out theory as fact. It is the spread of misinformation that drives me crazy.

I don't mean to make this a personal attack, but look a few posts up before this one:

That's true, didn't Steve Jobs say at WWDC this past year that the consumer line will go Intel first in 2006 and then the pro line in 2007?

NO! He didn't! I don't understand how this got to be so commonly accepted that people just casually throw it around. The keynote is available on Apple's website. Watch it...Steve doesn't say that the consumer line will go first.

It is so strange for me to be arguing like a crazy person over this, since I actually agree with this theory. How ironic. But, as I said before, I fear the days before MWSF when nearly everyone thinks Intel Macs are coming because "Steve said so" and then the negative response following MWSF when they all start crying after the keynote comes and goes without any announcement.

And about the 17" PowerBook...it was supposed to be ready by February '03. We all know what happened to that (ahem April). I'm not sure Apple wants to introduce a radically new product that is not yet done with a planned multi-month lead time. Can you imagine the riots if they were to miss the ship date?

We are all clinically insane you know. And I wouldn't want it any other way.

P.S. Can you imagine Microsoft addicts "arguing" (using the term loosely) over the release date of the latest version of FrontPage? That is what makes us Mac users.

 

[brianus]

"BY" June 2006, to me, means, "hopefully we won't screw it up and be late with it." It does not mean they'll miraculously have it out 6 months before that. You folks need to stop looking at the words for what they technically could allow ("they could be here TOMORROW!") and what they actually imply, based on experience. I don't think it was entirely coincidental that Jobs set that tentative deadline in the exact same speech where he reminded everyone of the unfulfilled promises of 3GHz and a G5 Powerbook.

And have any of you who so forcefully propounded the MWSF theory thought that maybe, just maybe, they would schedule another, separate event for this, when they actually are available? Why does it have to be the same, yearly thing? All evidence suggests that, though MacIntel had been a backup plan for them for years, they didn't actually decide to take the plunge until very recently. Which means they didn't have it all pefectly planned and figured out 2 months ago, and even now are not necessarily at the point where they can say firmly that they will meet this or that deadline -- they are still working on it. Feverishly, no doubt, but as much as their marketing people might want a January launch, it simply may not be possible.

 

[Fabio_gsilva]

It would be very application dependant.

For a scalable parallel application like LINPACKD (used to rank the top 500 supercomputers) it should be more than twice as fast - doubling the size of the cache should make up for the 8% loss in CPU MHz. Most embarrassingly parallel applications would hit the 4x mark as well.

For a purely memory-bound application (like the STREAMS benchmark), it would probably be about the same speed - unless the memory controller changes to use faster DIMMs.

For applications that fit well in the 1 MiB L2 cache of the 970MP, but which spill out of the 512 KiB on the 970FX, it could be much, much faster than 4x (but the useful applications that fit this profile are probably few to none).

The question really becomes interesting when you ask "how much faster is application XYZ". Here it depends on the nature of the application (can the algorithm be decomposed into 4 or more threads?) and the skill of the engineers (did they find every possible place to exploit parallelism?).

Some applications run the same speed on a dual as they do on a single. Moving those to a quad won't help at all. Others run somewhat faster on a dual - those might improve very slightly on a quad.

Only the applications that run nearly twice as fast on the dual are good candidates for a quad - and only if those were written to use more than two threads.

A lot of the "MP aware" apps are really "dual aware" - they were written to decompose the algorithm into two threads only.
___________

All in all, the quads will be most beneficial for farming applications (embarrasingly parallel), and a few true SMP-aware apps.

For the rest of us, only the power users who want run several heavy apps at once (rip CDs while burning DVDs while rendering in FCP) will get much benefit.

Thank you very much! Good post!

 

[beatle888]

It would be very application dependant.

For a scalable parallel application like LINPACKD (used to rank the top 500 supercomputers) it should be more than twice as fast - doubling the size of the cache should make up for the 8% loss in CPU MHz. Most embarrassingly parallel applications would hit the 4x mark as well.

For a purely memory-bound application (like the STREAMS benchmark), it would probably be about the same speed - unless the memory controller changes to use faster DIMMs.

For applications that fit well in the 1 MiB L2 cache of the 970MP, but which spill out of the 512 KiB on the 970FX, it could be much, much faster than 4x (but the useful applications that fit this profile are probably few to none).

The question really becomes interesting when you ask "how much faster is application XYZ". Here it depends on the nature of the application (can the algorithm be decomposed into 4 or more threads?) and the skill of the engineers (did they find every possible place to exploit parallelism?).

Some applications run the same speed on a dual as they do on a single. Moving those to a quad won't help at all. Others run somewhat faster on a dual - those might improve very slightly on a quad.

Only the applications that run nearly twice as fast on the dual are good candidates for a quad - and only if those were written to use more than two threads.

A lot of the "MP aware" apps are really "dual aware" - they were written to decompose the algorithm into two threads only.
___________

All in all, the quads will be most beneficial for farming applications (embarrasingly parallel), and a few true SMP-aware apps.

For the rest of us, only the power users who want run several heavy apps at once (rip CDs while burning DVDs while rendering in FCP) will get much benefit.

FINALLY, a post from AidenShaw that isnt abrasive. or at least that i didnt find offensive.

thanks for the info. but can you put that into laymen's terms? like what about if you run graphics applications...or video editing apps? what type do these fall under as far as being "memory-bound" ....etc.

 

[SiliconAddict]

Extenuating circumstances. He he.

My overriding point in all of these discussions isn't about the logic that one can use to determine a MWSF '06 launch. I'll admit, it makes a good deal of sense, even though I am not nearly as convinced as you are. My beef with this thread are the people who are going to advance your well-thought out theory as fact. It is the spread of misinformation that drives me crazy.

I don't mean to make this a personal attack, but look a few posts up before this one:



NO! He didn't! I don't understand how this got to be so commonly accepted that people just casually throw it around. The keynote is available on Apple's website. Watch it...Steve doesn't say that the consumer line will go first.

It is so strange for me to be arguing like a crazy person over this, since I actually agree with this theory. How ironic. But, as I said before, I fear the days before MWSF when nearly everyone thinks Intel Macs are coming because "Steve said so" and then the negative response following MWSF when they all start crying after the keynote comes and goes without any announcement.

And about the 17" PowerBook...it was supposed to be ready by February '03. We all know what happened to that (ahem April). I'm not sure Apple wants to introduce a radically new product that is not yet done with a planned multi-month lead time. Can you imagine the riots if they were to miss the ship date?

We are all clinically insane you know. And I wouldn't want it any other way.

P.S. Can you imagine Microsoft addicts "arguing" (using the term loosely) over the release date of the latest version of FrontPage? That is what makes us Mac users.

True. At the end of the day this is ALL speculation. My quote above from WWDC. *shrugs* Its Jobs. He can change his mind at any time. People keep missing the point that this is MacRUMORS not MacFACTS. If all we did was talk about facts here this would be one quiet little website. I should revamp my signature with something that says if I come out as a strong proponent for something that doesn't mean I beleive with 100% certainty that it will happen. I just umm get passionate when I argue. Even logic fails when you deal with any company as secretive as Apple. Maybe it would be best to say maybe a 60% chance of an announcement in January? Or maybe its closer to 50/50%?


PS- Hmmm who would ever argue over MS software?!! Arguing over MS software is like arguing over a steaming pile of fresh pig feces. At the end of the day you are still arguing over crap. Now you will excuse me. I have to continue the patchfest in our office. @*#(*@# dang MS @$*$#)* patches.

 

[DIXIE]

Apple with Intel is a PROFIT BLEEDER!

As earlier noted by a FEW realistic Mac users, the inevitable update to the PPC PowerBook will be substantial even though it might be the last of the series and it will be designed to carry the Pro user well into 2007 AND BEYOND! If the eager beavers think the first, second, or even third Mactels will be perfect just recall some very recent history: the iMacG5 (version A), Tiger still needs 10.4.3 or better, and the 23" Cinema Displays. I do not believe Apple thought the hackers would get into the MacTel Developer Kits and start spreading OSX onto PC boxes .... that had to be a wake-up call that Apple Computer Hardware and OSX must remain reliable, durable, and secure well past the "transition" period. Regardless of the iPod, Apple is not going to risk losing M-O-N-E-Y from the absence of desirable and useful high profit products. If things don't work out with the new bride, a quickie divorce by Steverino is not impossible. Therefor, Apple must have a tried and true fallback plan to carry the load during rough and uncharted waters! Amen.

This clip illustrates that the Intel "friendship" could just be the "beginning of a beautiful disaster....." no matter what the Pentium lovers rant and rave. PROFIT is the life blood of any business and all the juvenile esoteric technobabble that "mine is faster than yours" is meaningless. Apple can not afford to create an environment which permits massive piracy of its OSX, since the OPERATING SYSTEM sells the HARDWARE! You guys are forgetting that Apple is primarily a HARDWARE company, whereas Microsoft (ugh!) originally did a better SALES job pushing crummy software to the gullible corporations. You don't abandon the PPC ship so quickly, if the sharks are in the water!

 

[thehydra]

Hi all. I'm new...

Okay, I was interested in a Mini, just cuase I really want a PowerMAC when they come out with intel chips, so the Mini would serve as a affordable solution for the next year or so.

But now that they may ship with G5's, I may wait...

Since I'm a switcher, what would your recommendations be?

I'm gonna spend about $600-$1,000. Should I wait for a G5 Mini, or do you think an suped up iBook would be better?

I mainly want this for photo editing/management, music and light video editing.

Thanks in advance.

 

[oskar]

It's officially "Macintel" by the way.

I think you should've said: It's "officially" Macintel by the way. Quotation marks on "officially" because that's what many on MacRumors are calling them, but that has NOT been stated officially by any Apple authority.

I would hope that Apple gets the MacTel iMacs and Powerbooks out first.
BTW, "MacIntel" is the term that children use.

MacTel is an improper term or nickname altogether because "Tel" could be anything that begins with those 3 letter. Wintel which makes a little more sense includes the whole word "intel" and the beginning of windows.

I might as well drop this that I've already said many posts. They should be called x86 Macs to differentiate them from PPC Macs.

 

[generik]

I think you should've said: It's "officially" Macintel by the way. Quotation marks on "officially" because that's what many on MacRumors are calling them, but that has NOT been stated officially by any Apple authority.

You can say that if you are one who takes everything Apples says as gospel truth, but hello? Smell the coffee, Apple owns the trademark Mactel now, what do you think it stands for? Do you think Apple is going to be selling imitation Martel barbie dolls?

 

[Freg3000]

True. At the end of the day this is ALL speculation. My quote above from WWDC. *shrugs* Its Jobs. He can change his mind at any time. People keep missing the point that this is MacRUMORS not MacFACTS. If all we did was talk about facts here this would be one quiet little website. I should revamp my signature with something that says if I come out as a strong proponent for something that doesn't mean I beleive with 100% certainty that it will happen. I just umm get passionate when I argue. Even logic fails when you deal with any company as secretive as Apple. Maybe it would be best to say maybe a 60% chance of an announcement in January? Or maybe its closer to 50/50%?


PS- Hmmm who would ever argue over MS software?!! Arguing over MS software is like arguing over a steaming pile of fresh pig feces. At the end of the day you are still arguing over crap. Now you will excuse me. I have to continue the patchfest in our office. @*#(*@# dang MS @$*$#)* patches.

Sure I see the emphasis on Rumors and not Facts. But I'd like to think that this forum should be used to discuss rumors without stating theory as fact. It doesn't help anyone's thought process to be fed misinformation. Speculation is great. I love speculation. That is why I am here. But how can anyone do a good job speculating if they are constantly reading purported facts that are nothing more than speculation by someone else.

Even if it makes a world of sense (50, 60%+ ) that is no excuse to present theory as fact.

Things that are good for this forum:

• I like Apple.
• I wish Apple wasn't going to using Intel chips.
• I think Apple should get Intel Macs out to the public ASAP.
• I think Steve will announce Intel Macs at MWSF '06.
• Since Steve said they'd have Intel Macs shipping by WWDC '06, I think it's logicial to assume he will announce them at MWSF '06.
• It would make most sense for Apple to convert the consumer lines to Intel before the pro machines.
• Apple is going to take over the world and here is how...

Things that are bad for this forum:

• Steve said Intel Macs will be released at MWSF '06.
• I think that the iBook will be the first Intel Mac since Steve said the consumer line was going to Intel first.

I think it is clear where I believe the line should be drawn: Facts are good. Speculation is good. Speculation presented as fact is not good.

And it's not like I think you shouldn't be able to advocate your position so strongly that it seems as though you think of it as fact, I just don't think that blunt statements said so casually (as if they were facts carved in stone) are helpful for anyone.

No matter what your opinion is on this topic, or any other, we can all agree that the spread of misinformation is harmful to this forum. Can't we?

 

[runninmac]

Hi all. I'm new...

Okay, I was interested in a Mini, just cuase I really want a PowerMAC when they come out with intel chips, so the Mini would serve as a affordable solution for the next year or so.

But now that they may ship with G5's, I may wait...

Since I'm a switcher, what would your recommendations be?

Hey welcome to MacRumors... with what you said I dont know what you should do becuase the G5 mini rumors are semi-scetchey (sp?) I know Im thinking about getting one but I wont have enough money untill then so It doesnt matter... I would wait if money is tight. (sorry if that doesnt help at all or makes no sense.)

 

[thehydra]

Hey welcome to MacRumors... with what you said I dont know what you should do becuase the G5 mini rumors are semi-scetchey (sp?) I know Im thinking about getting one but I wont have enough money untill then so It doesnt matter... I would wait if money is tight. (sorry if that doesnt help at all or makes no sense.)

Thanks for the reply.

Okay, how bout this one.

What would the usual Mac user opt for? An iBook, or a Mini?

 

[sambo.]

I would sell my current PB an buy a dual core PB G4 in heart beat. The thing is only 3 months old too!

steve jobs loves people like you.

he hates people like me who still use their G3/333 PB's..... (six years old, not a single dead pixel)

 

[madmaxmedia]

I think Apple is really shooting to intro Mactel stuff at MW 06. It really makes too much sense to not go for that.

I think timewise they should be okay, if they have devoted enough resources and not underestimated the project.

 

[Squire]

Question: When was the last time Apple introduced new or updated computers at a "special media event" (i.e. NOT a MacWorld or Developer's conference)?

Anyone know? We know they've introduced software/services (the iTunes Music Store) at other events and, of course, iPods. How about actual Macs?

If the answer is "never" then the case for intel-based products at MWSF in January just got stronger. (But still not a certainty by any means.)

Squire

 

[powerbook911]

Thanks for the reply.

Okay, how bout this one.

What would the usual Mac user opt for? An iBook, or a Mini?

thehydra,

I can almost guarantee you the Mac Mini will not go to G5. Of course, you can never count something out 100 percent, but there is not any sensible belief in me that this will happen. It does not make sense.

What would the usual mac person opt for? What do you need? Do you need portability? If so, the iBook really is a great machine. However, if you want a great desktop system, with fantastic output options (including high resolutions) I would go with the Mac Mini.

 

[ajampam]

would it even make any sense to update mac minis to g5 when none of the portables have that....also with the impending intel step?? And heres another question....what about the same heat dissipation issues that did not let the g5 PBs to come out?

 

[2Cheap2Switch]

If a G5 mini came out, I would definitely have to pick that up. In fact me and a buddy of mine have had a bet going for a while and the loser has to buy the winner a mac mini. Win or lose I'll have to pick one up. If I win however I think I'll wait a little bit for the G5 mini.

 

[Xephian]

If a G5 mini came out, I would definitely have to pick that up. In fact me and a buddy of mine have had a bet going for a while and the loser has to buy the winner a mac mini. Win or lose I'll have to pick one up. If I win however I think I'll wait a little bit for the G5 mini.

Yeah, I would get one too.

 

[AidenShaw]

Okay, I was interested in a Mini, just cuase I really want a PowerMAC when they come out with intel chips, so the Mini would serve as a affordable solution for the next year or so.

But now that they may ship with G5's, I may wait...

Since I'm a switcher, what would your recommendations be?

Hmmm....

Do you have a significant investment in Windows software that you'll need to repurchase (same or equivalent) for the Mac? Do you plan to buy much software for the Mac?

If not, then a Mini now and a PowerMacIntel later wouldn't be too bad.

If you have a lot of Windows software, I'd definitely wait. Since the MacIntel can run Intel instructions, and since a compatibility layer like Wine lets people run many Windows binaries on x86 Linux - a port of Wine to OSX86 would mean that most of your Windows software would simply run on the MacIntel - no new purchases necessary.

If you plan to buy a lot of Mac software, then this could be a foolish time to switch unless you verify that all the software will get free upgrades to fat binaries that will support MacIntel. Why buy the same software twice?

 

[Freg3000]

Question: When was the last time Apple introduced new or updated computers at a "special media event" (i.e. NOT a MacWorld or Developer's conference)?

Hmmm...it certainly has never happened in my "Apple following" lifetime (2002-present). A quick look at Apple-History says this about the original iMac...

Announced in May 1998 and shipped in August, the iMac was Apple's computer for the new millennium.

Apple events 7 years ago probably looked quite different than they do today ( anyone?) but what goes on in May? Perhaps WWDC, I think it used to be happen in May, but I'm not sure.

Anyway I think you're right, Apple's Special Event computer hardware releases are few and far between. But who knows, these are strange times indeed.

And just as food for thought...since traditionally MWSF is a consumer oriented show (2005 was quite so, although 2004 featured G5 xServes and 2003 new PowerBooks) don't you think it would be strange for Steve to only pre-announce Intel Macs? Talking about a processor change, performance per watt, and level 2 cache goes over most people's heads, but Apple could get away with it at MWSF if at the end of the whole thing Steve had a pretty new Mac with all the goodies inside.

But few pushing the MWSF '06 theory suggest we'll be able to run out and buy Intel Macs immediately afterwards. So what are the chances Steve talks about the technicals and promises availability in March. It's not a developer show, it's a "I wanna go on the floor and play with the new stuff" show.

Just throwing out more ideas. (Presented as such...not as fact )

 

[beatle888]

You can say that if you are one who takes everything Apples says as gospel truth, but hello? Smell the coffee, Apple owns the trademark Mactel now, what do you think it stands for? Do you think Apple is going to be selling imitation Martel barbie dolls?

((O)) <––– chill pill


and you take apple simply trademarking a name as gospel?

 

[Squire]

would it even make any sense to update mac minis to g5 when none of the portables have that....also with the impending intel step?? And heres another question....what about the same heat dissipation issues that did not let the g5 PBs to come out?

First of all, those "same heat dissipation issues" have likely been solved with the announcement of the newer low-power G5. There are several arguments against a G5 mini but that's probably not one of them. "Well where's my Powerbook G5 then?" others may ask. In my opinion, it's not going to happen simply for, if nothing else, reasons related to Steve Jobs' ego. According to my interpretation of his keynote address, IBM's apparent inability to deliver a low-power G5 for portables was one of the prime reasons for the switch. That may or may not be the real reason. Nevertheless, it was touted as the reason at the time.

Hmmm...it certainly has never happened in my "Apple following" lifetime (2002-present)...Anyway I think you're right, Apple's Special Event computer hardware releases are few and far between. But who knows, these are strange times indeed.

It hasn't happened in my "Apple following" lifetime, either. But, as you mentioned, these are strange times. Such an introduction would definitely warrant a special event.

And just as food for thought...since traditionally MWSF is a consumer oriented show (2005 was quite so, although 2004 featured G5 xServes and 2003 new PowerBooks) don't you think it would be strange for Steve to only pre-announce Intel Macs? Talking about a processor change, performance per watt, and level 2 cache goes over most people's heads, but Apple could get away with it at MWSF if at the end of the whole thing Steve had a pretty new Mac with all the goodies inside.

But few pushing the MWSF '06 theory suggest we'll be able to run out and buy Intel Macs immediately afterwards. So what are the chances Steve talks about the technicals and promises availability in March. It's not a developer show, it's a "I wanna go on the floor and play with the new stuff" show.

Well...I think the "consumer topics vs. developer topics" argument (which always seems to surface when discussing what won't be unveiled at a certain WWDC) is grossly blown out of proportion. And I think it only works one way. For example, I think one could suggest that WWDC offerings are always slightly more limited but at MWSF, almost anything goes.

Just throwing out more ideas. (Presented as such...not as fact )

Yes, just speculation. Which is healthy, by the way.

Squire