PDA

View Full Version : PPC970 and WWDC Rumors: Part 2


MacRumors
Mar 25, 2003, 12:09 PM
MacBidouille (http://www.macbidouille.com/niouzcontenu.php?date=2003-03-25#5087) feeds the fire with more rumors about the PowerPC 970 and WWDC.

According to their most recent update, they claim that Apple's WWDC dates were indeed moved due to planning around the PowerPC 970. Configurations are reported at 1.4 GHz, 1.8 GHz, and dual 2.3 GHz with availability 6-8 weeks after the announcement.

While they will not be sharing early benchmark results, they claim they are very exciting.

MacBidouille has a variable history of accuracy with their rumor reporting. A top speed of 2.3GHz is contrary to public IBM claims on the max speeds of the new PPC 970s as well as rumors (%5C%22http://www.macrumors.com/pages/2003/03/20030312094607.shtml%5C%22) from CeBit.

dabirdwell
Mar 25, 2003, 12:12 PM
Now all we need is a 970 in a PowerBook.

How long?

Do we know the power usage figures?

PretendPCuser
Mar 25, 2003, 12:14 PM
Looking forward to it! All this waiting my not have been in vain.

First?:eek:

psxndc
Mar 25, 2003, 12:15 PM
What I'd like to know is how you pronounce MacBidouille in the first place.

I'm a dumb American that only speaks one language. Is it "Mac-Bi-dooey"? "Mac-Bi-dooley"? "Mac-Bi-day"?

-p

nakavori
Mar 25, 2003, 12:20 PM
mac bee doo yuh,

plus ou moins (more or less):)

Wes
Mar 25, 2003, 12:20 PM
Originally posted by psxndc
What I'd like to know is how you pronounce MacBidouille in the first place.

I'm a dumb American that only speaks one language. Is it "Mac-Bi-dooey"? "Mac-Bi-dooley"? "Mac-Bi-day"?

-p

I thought it was Mac-bi-dool.

Very promising rumours. Dual 2.3ghz would take a p4 4ghz.

junior
Mar 25, 2003, 12:21 PM
mac-bi-du-il?

mymemory
Mar 25, 2003, 12:22 PM
Originally posted by psxndc
What I'd like to know is how you pronounce MacBidouille in the first place.

I'm a dumb American that only speaks one language. Is it "Mac-Bi-dooey"? "Mac-Bi-dooley"? "Mac-Bi-day"?

-p

Oh man! that is so easy to pronunce:D When I talk in english for me "beach" and "bitch" sounds the same, I just pronunce something in between hoping for people to undertand what I really mean:rolleyes:

lou tsee
Mar 25, 2003, 12:31 PM
french

mac bee doo yuh,



that's the correct one. (whereas the 'yuh' is hardly more than a 'y')
anyway - I don't really buy their "we know too much" thing. sounds too cheap....

nuckinfutz
Mar 25, 2003, 12:34 PM
Now all we need is a 970 in a PowerBook.

IBM already has specs on a 1.1 Volt PPC 970@ 1.2Ghz. Dissipation is only 19watts making it more efficient than the current Powerbook 1Ghz processors used by 10watts.

cc bcc
Mar 25, 2003, 12:35 PM
I think there is to much difference in clockspeed and also the 2.3 GHz is a bit too high.
I think we'll see 1.4, 1.8, 2x1.8

And that would still rock! Aight!? :D

iloveMac
Mar 25, 2003, 12:36 PM
WOW!!! a dual 2.3ghz. I am slobbering right now.

alset
Mar 25, 2003, 12:36 PM
Doesn't sound to accurate to me. I'll keep my fingers crossed, but I won't really believe it until we see it from some more reliable sources.

Dan

eunuchs
Mar 25, 2003, 12:38 PM
AHAHAHAHAHAHA...

April Fools Day isn't until next week....

j33pd0g
Mar 25, 2003, 12:52 PM
With new PowerPC 970 machines mac users may win twice:

1. We may have faster macs available to us.
2. They may also do a hefty price drop on the current line up. To make room for the new stuff.

Mr. Anderson
Mar 25, 2003, 12:55 PM
But what I question is why just duals on the top end - that's a huge difference in performance and will it be two CPUs or 1 dual core CPU.....

looking forward to finding out.

D

Flynnstone
Mar 25, 2003, 12:59 PM
Since this is a rumor, lets hypothesize on real world performance.
I've heard that the 970 is upto 4 times faster than an equivalent G4. Lets assume 2 times. Front side bus (FSB) is approximately 5 time faster (970@6.4 vs G4@1.3). For bus constrained, 2 times is reasonable. A little apples and oranges, no apples and prunes here:D . A 1.8 GHz 970 would roughly be equivalent to a 3.6 GHz G4, about 2.9 speed up factor.
Now lets use Adobe's recently released numbers for a P4 @ 3.06 GHZ and a dual G4 1.25GHz. Some say Adobe is not optimized for dual processors, so say a single and dual G4 are the same. The Adobe number for the G4 was 1 minute 25 seconds or 85 seconds. Take the 85 seconds for the G4 @ 1.25 and divide by our speed up factor of 2.9 and we get about 29 seconds.
This is (theoretically) almost twice as fast as a P4@ 3.06GHz.
Cool! :cool:

CheekyGit
Mar 25, 2003, 01:00 PM
1.4 Ghz = cool :)
1.8 Ghz = even cooler :cool:
Dual 2.3 Ghz = Slobbering!!!! <wiping chin> :D :D

I'll save my pennies for 970 PowerBook and Dual 2.3 Ghz PowerMac. This is great news.

I'm glad I held off buying a 17" Powerbook.

CheekGit :D

minux
Mar 25, 2003, 01:00 PM
I do not think this will happen, and I for one am not holding my breath on it either. I would not xpect a new line of processors for about another year from Apple.

Mike

porovaara
Mar 25, 2003, 01:02 PM
Originally posted by Flynnstone
Since this is a rumor, lets hypothesize on real world performance.
I've heard that the 970 is upto 4 times faster than an equivalent G4.

Where have you head this?

sparkleytone
Mar 25, 2003, 01:04 PM
Originally posted by dabirdwell
Now all we need is a 970 in a PowerBook.

How long?

Do we know the power usage figures?

a long time. there is really no point at the moment. the g4 has long legs when it comes to notebook usage. the powerbook should get a good 3-5 revs before switching architecture. that translates to a good two years.

the g4 in the notebook world is the chip to beat, and has been for a good two yrs now.

t^3
Mar 25, 2003, 01:08 PM
Does anybody know what prevented the PowerBook from getting a G4 until a year after the G4 was released? I'm just wondering if were making too many assumptions about a 970 PowerBook being released soon after the chip itself is released...

unclepain
Mar 25, 2003, 01:18 PM
doesn't anyone read the rules around here?

Falleron
Mar 25, 2003, 01:19 PM
Originally posted by sparkleytone
a long time. there is really no point at the moment. the g4 has long legs when it comes to notebook usage. the powerbook should get a good 3-5 revs before switching architecture. that translates to a good two years.

the g4 in the notebook world is the chip to beat, and has been for a good two yrs now.
I really do believe that if the IBM chip gets a release this summer then the powerbook will get a slower version in the autumn. After all, its the year of the laptop. Just remember, the 1.2Ghz 970 uses less power than the G4, less heat which = a very good laptop. Not to mention the performance increase. Also, this would allow Apple to move the iBook to the G4 which will be better for OSX (being G4 optimised).

reyesmac
Mar 25, 2003, 01:20 PM
If they dont come out or announce new powermacs but the Panther demo runs really fast and smooth, then they have the G5 running the demo but not ready to announce yet.
I think they will just announce the new chip even if it takes them a long time to ship like the powerbook did. They can't compete anymore without it.

mania
Mar 25, 2003, 01:21 PM
Originally posted by sparkleytone
the g4 in the notebook world is the chip to beat, and has been for a good two yrs now.

I have to completely disagree. the g4 is way past its prime with its fake ddr and limited clock speeds etc. assuming a 970 ends up in the desktop in june, i bet a 970 will be in the powerbook by january, october if we are lucky. you know once people start experiencing the 970 they will really want it in the powerbook, even if it were at slower speeds.

Jackb602
Mar 25, 2003, 01:21 PM
Here is my best attempt at a translation of the MacBidouille posting. I started with Babelfish, and made corrections for clarity. Hope this helps,

Jack

"If you do not want to read rumours, move right along. But you will miss an opportunity to feel an excitement that the mac world has not felt in years!

Apple did indeed shift the dates of the WWDC because of the PPC 970. The new machines will be presented there and will even be available for preorders at that time. The machines will also be available on the AppleStore as of their announcement, but with delivery in 6 to 8 weeks.

There will be 3 configurations available:
- single 1.4 GHz
- single 1.8 GHz
- dual 2.3 GHz

For the latter, the motherboard is not yet completely finalized but should be during the first week of April. The contractor who wins the request for bids on the single processor motherboard will also manufacture the dual processor motherboard. Therefore, note that there will be two versions of motherboards, which is to be expected since the processor is not on a daughter card but rather is mounted on the motherboard via a ZIF socket.

Benchmarks were made on the new machines. I admit being extremely excited by all these rumors. Take note! We will finally have something to compete with the Wintel world. Because let's not kid ourselves. The fate of Apple is at the center of all our concerns!

[ Update ] This update concerns the benchmarks. After discussions with the members of the Macbidouille team, we decided to censor ourselves regarding the benchmark results. For the first time, we had the feeling we knew too much. In our opinion, it is far too early to publish them. We hope that you will understand our decision.

P.S. It will be useless to e-mail us asking for more information, because we have already firmly decided to remain silent."

nuckinfutz
Mar 25, 2003, 01:21 PM
I do not think this will happen, and I for one am not holding my breath on it either. I would not xpect a new line of processors for about another year from Apple.

Mike



I don't understand your reasoning. If this was the case it would be because Apple refused to use the processors. The 970s are being sampled now and enter Volume production 2H of this year. I'm all for reality checks but blind pessimism is never a good policy


At any rate Apple has no choice but to increase speed in a hurry.

nuckinfutz
Mar 25, 2003, 01:27 PM
I have to completely disagree. the g4 is way past its prime with its fake ddr and limited clock speeds etc. assuming a 970 ends up in the desktop in june, i bet a 970 will be in the powerbook by january, october if we are lucky. you know once people start experiencing the 970 they will really want it in the powerbook, even if it were at slower speeds.

There is no such thing as "fake DDR" You're just making yourself look like a fool. The DDR is not implemented in a fashion that maximizes it's performance for the processor but it is far from being "fake"

trilogic
Mar 25, 2003, 01:35 PM
I guess the g4 cpu with 1.42 ghz would still be great if it could support faster bus speed and ram. but there must be some restriction.

anyway

what is the frontside bus speed of a 970 cpu?

oh my, if this is true ... I can't wait ...

I'm working with a 400MHz G4, I'm in the market for a new machine. oh yeah baby and a cinema display, yeah baby yeah. :D :D :D

abdul
Mar 25, 2003, 01:36 PM
Originally posted by unclepain
edited What you said was childish and has no place here. Please grow up!!

mikeyredk
Mar 25, 2003, 01:39 PM
lets keep the politics down or arn will shut this thread down like he did with the other one on gore

bidge
Mar 25, 2003, 01:49 PM
Originally posted by Flynnstone
Since this is a rumor, lets hypothesize on real world performance.
I've heard that the 970 is upto 4 times faster than an equivalent G4. Lets assume 2 times. Front side bus (FSB) is approximately 5 time faster (970@6.4 vs G4@1.3). For bus constrained, 2 times is reasonable. A little apples and oranges, no apples and prunes here:D . A 1.8 GHz 970 would roughly be equivalent to a 3.6 GHz G4, about 2.9 speed up factor.
Now lets use Adobe's recently released numbers for a P4 @ 3.06 GHZ and a dual G4 1.25GHz. Some say Adobe is not optimized for dual processors, so say a single and dual G4 are the same. The Adobe number for the G4 was 1 minute 25 seconds or 85 seconds. Take the 85 seconds for the G4 @ 1.25 and divide by our speed up factor of 2.9 and we get about 29 seconds.
This is (theoretically) almost twice as fast as a P4@ 3.06GHz.
Cool! :cool:


Ahhhhh if it is up to 4 times faster that would mean that the Dual 2.3 would be comaparable to a 10Ghz G4, that is way too cool.....

dongmin
Mar 25, 2003, 01:50 PM
There is no way this report can be accurate: too much gap between the middle and top configs. More likely the top will be either a dual 2 ghz or a single 2.3 ghz.

I don't know if it's worth speculating about 'real world performaces' until we get some sort of specs list about the motherboard, and not just the CPU. If my memory severs me right, the SPEC benchmarks show the 970 to be about twice as fast as the G4 at the same clockspeed. With faster bus and RAM, the real world performace gap should be even bigger. But it's pure guesswork at this point until we see some real specs.

As for laptops, you probably will have to wait until next Spring, or January at the earliest, to see the 970 trickle down to the Powerbooks. This past January brought us the 12" and 17" Powerbooks, which were major revisions. I would guess that Apple wants to get at least a year out of these redesigns before they make another major revision. Putting in the 970 will require a pretty heavy redesign, I'd venture to guess.

strider42
Mar 25, 2003, 01:53 PM
Originally posted by dukestreet
But what I question is why just duals on the top end - that's a huge difference in performance and will it be two CPUs or 1 dual core CPU.....

looking forward to finding out.

D

Why "just" duals on the top end? Because quad processors would be very, very expensive, and probably wouldn't bring enough extra performance to justify the cost (diminishing returns)

And I absolutely promise it will be two CPU's and not a single dual core chip (assuming they are duals at all). Why you ask? because the 970 is a single core chip according to everything that I've read. The Power4 is a dual core chip, but the 970 is NOT the power4.

jholzner
Mar 25, 2003, 01:58 PM
Originally posted by strider42
Why "just" duals on the top end? Because quad processors would be very, very expensive, and probably wouldn't bring enough extra performance to justify the cost (diminishing returns)

And I absolutely promise it will be two CPU's and not a single dual core chip (assuming they are duals at all). Why you ask? because the 970 is a single core chip according to everything that I've read. The Power4 is a dual core chip, but the 970 is NOT the power4.

I think he's asking why the 1.4 and 1.8 models are not also duals...why just the top end model?

gothamac
Mar 25, 2003, 02:05 PM
please refrain from these types of posts against each other!!!

trilogic
Mar 25, 2003, 02:15 PM
Originally posted by dongmin
There is no way this report can be accurate: too much gap between the middle and top configs. More likely the top will be either a dual 2 ghz or a single 2.3 ghz.

I think there IS a market for really expensive and powerfull workstations, like SGI used to make. There is software around where you litterly can have enough speed. like Shake, maya, video editing and stuff. If this rumor is true, I wouldn't be surprised to see a highend super mac > 9999$. and I'm sure there are people who just wait for a machine like that.

GeneR
Mar 25, 2003, 02:16 PM
I really want a time machine right now. Go ahead about, oh, a year, grab the newest, "bestest" Powerbook with the newest, "bestest" OSX and come back to the present, and... and... well, have more fun.

:D

GeneR
Mar 25, 2003, 02:18 PM
If the 970 did come out at the show and there's a 6-8 week lag, when should we expect the first 970's to arrive in our hot little hands?

Q3? Q4?

:D

GPTurismo
Mar 25, 2003, 02:27 PM
I can't wait. I believe a single 970 2.3 by itself would beat a p4 4.0 in raw out power due to it being on a faster bus, better device communication system and most importantly 64 bit. lets hope those lazy programmers get off their bum bums and optimize their software :D

mathiasr
Mar 25, 2003, 02:31 PM
Originally posted by trilogic
...

what is the frontside bus speed of a 970 cpu?



The FSB is split in two point-to-point 32 bits unidirectional busses, running at up to 450 MHz DDR (that is 900 MHz bit rate).

This leads me to think that the rumored frequences are more likely to be 1.35 (3 x 450), 1.8 (4 x 450) and 2.25 (5 x 450) GHz.

Steamboatwillie
Mar 25, 2003, 02:31 PM
Hey, with all this talk about the 970 has there been any mention about the rumor of Apple using some new AMD chip? I thought I read something about that somewhere. I appologize if I am brain farting on this... :confused:

dongmin
Mar 25, 2003, 02:32 PM
Originally posted by GeneR
If the 970 did come out at the show and there's a 6-8 week lag, when should we expect the first 970's to arrive in our hot little hands?

Q3? Q4?

:D

WWDC is on June 23-27.

6-8 weeks puts delivery in early August. Perfect for going back to school. Perfect for me to take out more school loans and get one of these suckers.


Dual 2.3 ghz 970
900 mhz FSB
ATI 9700
2 gigs of RAM
Superdrive
FW800
Bluetooth
$3000
(the going rate for the 'Ultimate' config via the student ADC discount)


Dare to dream...



edit for formatting

ktlx
Mar 25, 2003, 02:40 PM
Originally posted by mathiasr
The FSB is split in two point-to-point 32 bits unidirectional busses, running at up to 450 MHz DDR (that is 900 MHz bit rate).

This leads me to think that the rumored frequences are more likely to be 1.35 (3 x 450), 1.8 (4 x 450) and 2,25 (5 x 450) GHz.

I don't know if it is true or not, but some have claimed that the PowerPC 970's FSB runs at half the processor speed. So a 1.2Ghz would have a 600Mhz bus, a 1.4Ghz would have a 700Mhz bus and a 1.8Ghz would have a 900Mhz bus.

I have absolutely no idea if this is correct or not and the IBM PDFs that I have seen don't allow you to conclude one way or another.

wms121
Mar 25, 2003, 02:42 PM
what software will Steve announce..after the 6-8 weeks are up?

Java J2EE ver. 1.4 will be released this summer..and 64 bit Java for IBM products is already available. Will we get "64 bit J2EE 1.4" by Christmas?

Can IBM duplicate Moto's '7540,7550,7560' series using the 970 ppc as a "Book F" template?

Can I get a Dick Tracy satellite wrist watch with a 970 built in that gets HDTV coverage of the new Korean war starting in 2004?

(Sorry ..the last one was a question for Al Gore..)

WW

Mr. Anderson
Mar 25, 2003, 02:45 PM
Originally posted by dongmin
WWDC is on June 23-27.


Dual 2.3 ghz 970
900 mhz FSB
ATI 9700
2 gigs of RAM
Superdrive
FW800
Bluetooth
$3000
(the current going price for the 'Ultimate' config via the student ADC discount)


Dare to dream...


ah, that would be so very nice. And do you know that the 6-8 weeks wouldn't be for the top of the line.....not if Apple's track record continues. They've even mentioned in the article that the dual mobo wasn't finished yet. So add another month on that time line and you'll get closer to the delivery date.

But you bette order it the first day, cause they'll probably sell tons of them and the backlog is going to be huge, further delaying the delivery.

I'm thinking of getting one, as well, but I'll wait and see. Might wait for the 2nd version and see if the bugs have been worked out since its going to be a machine I'll be working on for 3+ years.

D

hacurio1
Mar 25, 2003, 02:48 PM
Originally posted by GPTurismo
I can't wait. I believe a single 970 2.3 by itself would beat a p4 4.0 in raw out power due to it being on a faster bus, better device communication system and most importantly 64 bit. lets hope those lazy programmers get off their bum bums and optimize their software :D

Don't Forget the Vector Unit

trebblekicked
Mar 25, 2003, 02:50 PM
Originally posted by dongmin
WWDC is on June 23-27...puts delivery in early august

so that meshes with the PM's product cycle. still, panther doesn't make a logical choice for a code name if the OS is to be 64 bit. cheetah would be more apropriate. in addition, how long would it take for 64bit optimized apps? i'll say we see 970/Panther at WWDC, but panther is 32bit. mathiasr makes a good point about frequencies, too. we'll see 64bit 10.4 cheetah with a slew of 64 bit programs released come MWSF '04.

eric_n_dfw
Mar 25, 2003, 02:57 PM
I believe Cheetah was the code name for 10.0

Code names have nothing to do with the product anyway - Apple using Jaguar on released product marketing was a first, as far as I know.

trebblekicked
Mar 25, 2003, 02:59 PM
Originally posted by eric_n_dfw
I believe Cheetah was the code name for 10.0

-'doh. stand corrected. well there goes that. oh what the hell. 64 bit OS's for everyone then!

jettredmont
Mar 25, 2003, 03:04 PM
Originally posted by porovaara
Where have you head this?

That's from SPEC numbers. The 1.8 theoretical SPEC is about 4x the 1GHz G4. Note that this isn't on a per-clock basis; that brings it closer to the 2x faster figure the original post went with.

On the other hand CPU performance boost != system performance boost. you still have to read stuff from the HD and write it back out, etc. I am looking for a 1.8GHz 970 to pretty much match Intel's 3GHz P4 in overall performance, although of course I'd accept a surprise ... :)

jettredmont
Mar 25, 2003, 03:07 PM
Originally posted by t^3
Does anybody know what prevented the PowerBook from getting a G4 until a year after the G4 was released? I'm just wondering if were making too many assumptions about a 970 PowerBook being released soon after the chip itself is released...

Supply of G4s is what kept it out of the notebook form factor for so long. Apple traditionally tries to make sure demand is being met in their pro desktop lines before trickling down new processors to the notebook and consumer desktop lines.

Granted, such a policy is always subject to change, and of course IBM in general has a better record with meeting demand than Motorola, so you never know when the 970 will show up in notebooks.

strider42
Mar 25, 2003, 03:09 PM
Originally posted by jholzner
I think he's asking why the 1.4 and 1.8 models are not also duals...why just the top end model?

Ooops, I think you're right. Still comes down to the same thing in the end: money. More profit margin on a single processor machine on the low end than a dual, and its enough to make someone want to upgrade to to the upper end instead of just buying the low end. Product differentiation at its finest. Kind of like how when I bought my car, to get the automatic trans, you needed to go up a model (not a perfect analogy, but you get the idea, especially since I prefer a manual trans anyway)

jettredmont
Mar 25, 2003, 03:18 PM
Originally posted by ktlx
I don't know if it is true or not, but some have claimed that the PowerPC 970's FSB runs at half the processor speed. So a 1.2Ghz would have a 600Mhz bus, a 1.4Ghz would have a 700Mhz bus and a 1.8Ghz would have a 900Mhz bus.

I have absolutely no idea if this is correct or not and the IBM PDFs that I have seen don't allow you to conclude one way or another.

As has been discussed before, the "half the processor speed" comment appears to only apply to the 1.8GHz unit. It would be downright foolish to produce a chip in which each speed bump operates at a radically different FSB rate from the last, as changing the FSB requires modifying the chipset and possibly overall motherboard design instead of just plopping in the newer faster chip (more or less :) )

I expect the 900MHz bus will stay constant for the time being at least (also, note that the bus has an effective bandwidth of an 800MHz bus after overhead ... still a lot faster than anything else out there for the Mac and at the moment faster than Inte's 533MHz bus too).

edenwaith
Mar 25, 2003, 03:22 PM
Originally posted by psxndc
What I'd like to know is how you pronounce MacBidouille in the first place.

I'm a dumb American that only speaks one language. Is it "Mac-Bi-dooey"? "Mac-Bi-dooley"? "Mac-Bi-day"?

-p

Well, since it's French, you just pronounce the first half of the word and drop the rest. So one could pronounce it either "Mac" or "MacBi" or "MacBid".

I'm just kidding around. I just feel threatened by any language which tries to be more messed up than my own native language (Americanized English).

Anyway, if the RUMOR of a 2.3 GHz is true, then that would be sweet. I think that would put the Mac back into the race. What is the top speed AMD? The last I saw it was a 2.2GHz, but it could compete pretty well with a 3GHz P4.

seamuskrat
Mar 25, 2003, 03:38 PM
First- any news on the processor front is good news. I think even the Mac loyal will concur that we need a new birght spot in the future. That said, on to my next statement.
Correct me if i am mistaken on this as well...

1. The 970 will NOT be dual core. Never was meant to be. its based on a server design called the POWER4 which IS multicore, but the derived chip called the 970 will NOT be multicore- AND will include an 'alti-vec' style vector unit.

2. A 64 bit OS will be no faster in most tasks than a 32 bit OS. The bits have to do with how memory is addressed and such. So, Photoshop will not have to be re-written to be '64 bit clean'. etc. That said, some apps WILL benefit from having more RAM available to them. 64 bit systems and OS will allow for a higher RAM total than is currently available.
2a) Certain tasks can be designed and coded in a way to take advantage of the 64 bit processor. Aspects of OS X and other apps could have compenents that will benefit in performance on this. However, IMO, not a dramatic performance improvement.
2b) Because the 970 runs 32 bit PERFECTLY fine, 64 bit OS and/or software IS NOT A PREREQUISITE for release.
2c)There is no reason why certain aspects of a package cannot be installed for 64 bit vs 32 bit machines. A compiler update can allow for binaries to be compiled in 64 bit vs 32 bit. In a package the installer could check and install the appropriate version. So, vendors would not have to 'make separate' versions of software. They could compile binaries in each variety and have it be transparent. Again, specific advantages to using 64 bit clean code will NOT apply to a 32 bit system, such as memory addressing.

3. Lets assume that we are not all Apple fanatacists. From a corporate stategy, Apple wants to re-coop development costs on its current designs. So, taking this track, Apple will want to maximize life out of the current 17 and 12 powerbooks. That would indicate that laptops 970 would be less likely early in the product line. But since we are talking business, we know that Apple needs some help. The 970 SEEMS to be the help that is needed. I fear that laptop sales would plummet if desktop 970's were released and laptop updates were not forthcoming. So, it makes sound business sense to have 970s ina desktop and laptop modality as soon as technically possible.
3a) Consumer machines. We know that Moto has a few more breaths of life in the G4 series. Its possible that the newest G4 series could prolong the consumer line while the 'pros' get the 970 for the first year of production. What I do not know is the cost factor. What does the current G4 and upcoming models cost and what will the 970 cost per chip. This would go a long way to 'roadmap' a consumer vs pro future. However, if the consumer line would get 1.25 to 1.8 G4 with L2 and L3 cache with DDR (maximied or as is) at the current (or lower) prices and the pro line gets 970s then Apple would have a period of 9 to 12 months to breathe, and retool the consumer line for 970s. In that 12 months, I think we can feel confident IBM will take the 1.8 and allow it to scale to 2.8. So, lets assume we have this. Consumer lines with 1.25 to 1.42 G4 with L2/L3. Pro line with 1.4 to 1.8 970 for year 1. Year 2 consumer gets 1.8 to 2.x 970 and pro gets 2.5+ 970 )with dual options). Sounds reasonable from a tech standpoint and buisiness model. If the SPEC scores standup it makes the proline compeative, and the consumer lines very nice (afterall- they would be the same or slightly btter than todays PRO line).

As I said, I may have misinterpreted some of the technical facts, but I am nearly sure I have managed to explain it all correctly. I hope we can all say that come summer things will be very different for the platfrom.

shadowfax
Mar 25, 2003, 03:51 PM
Originally posted by seamuskrat
1. The 970 will NOT be dual core. Never was meant to be. its based on a server design called the POWER4 which IS multicore, but the derived chip called the 970 will NOT be multicore- AND will include an 'alti-vec' style vector unit.

I don't think the chip has anything on it to preclude dual processing. the system controller or something like that should decide about DP. the 970 may not be DP on new PMacs when they arrive, but that doesn't mean that there is no way to DP 970s.

edenwaith
Mar 25, 2003, 03:57 PM
Originally posted by dongmin

6-8 weeks puts delivery in early August. Perfect for going back to school. Perfect for me to take out more school loans and get one of these suckers.


One thing which I don't quite get is why Apple will announce a product 1.5 to 2 months before it is scheduled to ship...it seems like that would hurt their sales, or at least their revenue, especially on a quarter system. Such as when Apple has showcased new items, yet they aren't going to be released for several months. I can understand how the MacWorld expos might have pushed Apple into wowing the audiences again and again, even if the products weren't QUITE ready.

wrylachlan
Mar 25, 2003, 04:00 PM
Originally posted by Shadowfax
I don't think the chip has anything on it to preclude dual processing. the system controller or something like that should decide about DP. the 970 may not be DP on new PMacs when they arrive, but that doesn't mean that there is no way to DP 970s.


Dual Core and Dual Processor are not the same thing. Dual Core means you have two processors in the same piece of silicon.

Dual Processor means two separate pieces of silicon working together.

The 970 will definitely be DP capable, but it is not Dual core.

pgwalsh
Mar 25, 2003, 04:01 PM
Originally posted by dongmin
WWDC is on June 23-27.


Dual 2.3 ghz 970
900 mhz FSB
ATI 9700
2 gigs of RAM
Superdrive
FW800
Bluetooth
[B]$3000
(the going rate for the 'Ultimate' config via the student ADC discount) Okay, this looks pretty good, but I'm shooting for the ATI 9800...and standard two hd raid.

steve53e
Mar 25, 2003, 04:03 PM
Rumor or not.... I don't care if it takes a year or more, but I'm not giving Apple another dime for a computer until they release a box that clocks out at 2.5 ghz or better. Until then, 800 mhz will have to do and I'll gladly spend my money elsewhere.

Frobozz
Mar 25, 2003, 04:07 PM
Originally posted by iloveMac
WOW!!! a dual 2.3ghz. I am slobbering right now.

yeah, I think that day-- should these live up to the rumors, be one of the best days in apple's history. seriously. That means somewhere between the middle of July to the middle of August we will have dual 2.3 Ghz 970's. That's a HUGE performance boost. It ain't like getting dual 2.3 Ghz G4's... this is much closer to 4 or 6 times faster.


YES! YES! YES!

Frobozz
Mar 25, 2003, 04:10 PM
Originally posted by steve53e
Rumor or not.... I don't care if it takes a year or more, but I'm not giving Apple another dime for a computer until they release a box that clocks out at 2.5 ghz or better. Until then, 800 mhz will have to do and I'll gladly spend my money elsewhere.

WOW. You missed the point, didn't you? :-)

A 2.3 Ghz 970 is roughly equivalent to a G4 at twice that speed. How is 4 Ghz for ya? Just as you can not translate a P4 to a G4 in Mhz, you can't translate a G4 to a 970 easily. (though I tried)

Frobozz
Mar 25, 2003, 04:13 PM
Originally posted by edenwaith
One thing which I don't quite get is why Apple will announce a product 1.5 to 2 months before it is scheduled to ship...it seems like that would hurt their sales, or at least their revenue, especially on a quarter system. Such as when Apple has showcased new items, yet they aren't going to be released for several months. I can understand how the MacWorld expos might have pushed Apple into wowing the audiences again and again, even if the products weren't QUITE ready.

They don't care about selling G4 Towers at MWNY, because the focus will be changing. They won't have new hardware. Their existing sales are pretty rough around edges, and are nothing to write home about. They really aren't doing anything but showing the world they are still relevent in the raw-computing arena. If sales slipped to practically zero, it's not a big impact on their overall performance because their current numbers are so low.

jgracia
Mar 25, 2003, 04:16 PM
Why single 1.4, 1.8 an dual 2.3 and not a single model with 2.3 Mhz processor???

Dont Hurt Me
Mar 25, 2003, 04:19 PM
How does this french site know whats happening in cupertino, that is the question. and how the heck do they come up with a 2.3 ghz 970? i would think a 1.4,1.8 and dual 1.8 or dual 2.2 but not a 2.3. a single 1.4 or 1.8 will do me just fine. Some food for thought. Apple wont just throw this into a minor redo of the powermac. They will want a brand new looking case to show off this brand new 970. Any ideas?How about black with chrome?

Nipsy
Mar 25, 2003, 04:21 PM
Originally posted by GPTurismo
I can't wait. I believe a single 970 2.3 by itself would beat a p4 4.0 in raw out power due to it being on a faster bus, better device communication system and most importantly 64 bit.

What 64 bit aware apps have you got?

IJ Reilly
Mar 25, 2003, 04:21 PM
Originally posted by edenwaith
One thing which I don't quite get is why Apple will announce a product 1.5 to 2 months before it is scheduled to ship...it seems like that would hurt their sales, or at least their revenue, especially on a quarter system. Such as when Apple has showcased new items, yet they aren't going to be released for several months. I can understand how the MacWorld expos might have pushed Apple into wowing the audiences again and again, even if the products weren't QUITE ready.

True, and Apple has to be finally getting that point after the G4 iMac debacle of last year and the 17" PowerBook of this year. Maybe -- just maybe -- that's why WWDC was put back a month. Sure, I know what wishful thinking gets you, but I've got to be optimistic about this move coming off very soon, or accept that Apple has fallen down and can't get up.

nubero
Mar 25, 2003, 04:30 PM
Look at

http://www-3.ibm.com/chips/techlib/techlib.nsf/techdocs/A1387A29AC1C2AE087256C5200611780/$file/PPC970_MPF2002.pdf

See the Page with the SPECint2000 Benchmarks?

Now Look at

http://www.ideasinternational.com/benchmark/spec/specint2000.html

Will somebody now please explain to me, why the PPC 970 is supposed to be any faster than the P4? They can be gratefull if they come CLOSE to the P4...

Or am I overlooking something here?

nighthawk
Mar 25, 2003, 04:30 PM
Originally posted by IJ Reilly
True, and Apple has to be finally getting that point after the G4 iMac debacle of last year and the 17" PowerBook of this year. Maybe -- just maybe -- that's why WWDC was put back a month. Sure, I know what wishful thinking gets you, but I've got to be optimistic about this move coming off very soon, or accept that Apple has fallen down and can't get up.

But the 970 is a whole different tamale... it would make the most sense to announce the 970 ProPPC Mac at WWDC, where the developers would need to know the most about it... even if the wait was 2 months, it would be official (and that is what Apple needs now more than ever).

seamuskrat
Mar 25, 2003, 04:42 PM
Originally posted by Shadowfax
I don't think the chip has anything on it to preclude dual processing. the system controller or something like that should decide about DP. the 970 may not be DP on new PMacs when they arrive, but that doesn't mean that there is no way to DP 970s.

I apologize for not being more clear.
The 970 is NOT dual core, but IS dual Processor capable.

There is a significant difference. In some cases, a dual processor system will outperform a dual core, in other cases vice versa. What it means to us, is that we can be reasonable sure that some configuration will include dual processors.

What is fun ti imagine, is that the hypothetical design of the POWER5 derititive (maybe called the 980?!?) could perhaps be dual core. So, in 2 to 3 years we could see dual core/dual processor macs. Confused? heehee

Please for the sake of accuracy refer to the 970 as a DP capable CPU not a dual core CPU.
Thanks!

nighthawk
Mar 25, 2003, 04:46 PM
Originally posted by seamuskrat
...Lets assume that we are not all Apple fanatacists. From a corporate stategy, Apple wants to re-coop development costs on its current designs. So, taking this track, Apple will want to maximize life out of the current 17 and 12 powerbooks. That would indicate that laptops 970 would be less likely early in the product line. But since we are talking business, we know that Apple needs some help. The 970 SEEMS to be the help that is needed. I fear that laptop sales would plummet if desktop 970's were released and laptop updates were not forthcoming. So, it makes sound business sense to have 970s ina desktop and laptop modality as soon as technically possible.


The way I see it, the Powerbook 12" is the path to replace the iBook line with the G4. Meanwhile, the 15" ALu and 17" will sport the PowerPC 970.

Although supplies were low for the G4 at the beginning -- the main reason why the PowerbookG4 was so much later than the desktop was Power consumption and heat. A couple of revisions were needed to reduce both of these, and still the Powerbook was difficult to engineer.

Now the current 970 1.0ghz only uses 10watts which is less than the current 1.0ghz G4 processor in the Powerbook now. (I have no idea how the 1.0 is configured with bus/clock-mult., but those are the figures on the IBM website). Possibly we will see the 1.35ghz in the Powerbook this year also, but it depends on cost/availablility, and whether or not there is some sort of speed-stepping technology that would work with the 970 (which would make a lot of sense if IBM wants to create blades with it).

seamuskrat
Mar 25, 2003, 05:03 PM
Although we hate 'vaporware' products and products that are announced as 'here' and take months to ship, Apple has had a few of these. Here is MY PERSONAL brief analysis on this aspect of the company.

Up until rather recently, Apple has choosen to focus on MacWorld expos, every 6 months to announce new machines. Yes, I know there have been exceptions, but it has become rather customary. That means to maximize the press, focus and bang for your buck you announce when half the world's mac population is watching with baited anticipation of new machines.

With the situation is it stands today, Apple is in a bit of a pickle. Its got a great OS, with a loyal userbase and what we all know has the vast potential yet unrealized to be superior. Yet the current line-up although impressive, is not the blazing speed busting machines we dream of. We all would love to say to our Wintel buddies tha we were RIGHT all along and we do have the better/faster machines.

With Pentiums reaching past 3 gig and the G4 at a mere 1.42 ist getting harder to brush away the gap. With M$ focusing on its own switch campaign and the folks at Adobe pushing Intel chips, it places Apple in a position where it does NOT set the rules or agenda.

Apple can no longer dictate the terms it announces products. It must now react to the market conditions.

By announcing/previewing 970s and 10.3 at WWDC they can get some vitally needed sales recorded on the books before quarter end. The reality of today is that Powermac sales are meager. The cost/performance ratio is not there for 'switchers' and many power users are aware of SOMETHING better in the pipeline. Of course, we have all too often been dissappointed by MOTOs latest efforts.....

So, that menas Apple MUST do something fairly drastic soon. WWDC seems like a great opportunity. Get the developer crowd excited. The delay was in part due to current politics and safety concerns, and mostly due to the chance to announce at WWDC, get early sales for the quarter and to some extent confirm though action that MacWorld July is DOA for Apple this year.

Whether Apple announces AMD fab chips, IBM 970s, Intel chips, or even another gen G4, they have to do it realtively soon or they will be forever placed in the margins of computing. I doubt Apple COULD go under, as throughout its history it has made blunders that would have sent many a company to its grave, but Apple could still be relgated to the footnotes of history, so IMO they need to act NOW.

Logic dictates that the IBM 970 will be the way to go. Its available. Its PPC compatible. Its scalable upwards for the foreseeable future. Its NOT Moto. Contrary to folks like Dvorak at PCMag, I doubt Intel is the way to go. Another transition akin to 9 to X would be harmful overall. And a Mhz to Mhz comparison of OS X vs Win on a X86 platform would still prove problematic.

It seems like Apple has fallen back to punt as they say. I suspect we will hear some great software news at NAB, and see some more spectacular stuff in the next weeks leading up to WWDC. At WWDC we will know for sure.

Keep in mind Apple is a public company out to make money. So many of their choices must be dictated by this directive. I think (hope) that we will see more product announcements at non-MacWorld events over the next 18 months and it will all be fun.

scem0
Mar 25, 2003, 05:04 PM
dual 2.3! They will sell a lot of those. How can you justify spending
~$2,000 for a single 1.8 when for a couple hundred more you
can get two, much faster processors?

seamuskrat
Mar 25, 2003, 05:09 PM
Originally posted by nighthawk
The way I see it, the Powerbook 12" is the path to replace the iBook line with the G4. Meanwhile, the 15" ALu and 17" will sport the PowerPC 970.

Although supplies were low for the G4 at the beginning -- the main reason why the PowerbookG4 was so much later than the desktop was Power consumption and heat. A couple of revisions were needed to reduce both of these, and still the Powerbook was difficult to engineer.

Now the current 970 1.0ghz only uses 10watts which is less than the current 1.0ghz G4 processor in the Powerbook now. (I have no idea how the 1.0 is configured with bus/clock-mult., but those are the figures on the IBM website). Possibly we will see the 1.35ghz in the Powerbook this year also, but it depends on cost/availablility, and whether or not there is some sort of speed-stepping technology that would work with the 970 (which would make a lot of sense if IBM wants to create blades with it).

I tend to aggree. I think that to remain competative they HAVE to have laptops with new chips in the pro line. We KNOW thorugh the 12 PB that a G4 can be made small and fit in a small package. We can assume that the 970 will be usable in a laptop config but until we see usage specs and such we cannot know fur sure. I have seen some HUGE heat sinks on some chips shown in the literature. Cost of course is another issue.

But Steve did say 2003 was the year of the laptop, so we can only hope that 970 will be seen in laptops before 2004. But logic still dictates that the justnow shipping G4 17 will have to be around for a few (6?) months before going 970.

nighthawk
Mar 25, 2003, 05:25 PM
Originally posted by seamuskrat
I tend to aggree. I think that to remain competative they HAVE to have laptops with new chips in the pro line. We KNOW thorugh the 12 PB that a G4 can be made small and fit in a small package. We can assume that the 970 will be usable in a laptop config but until we see usage specs and such we cannot know fur sure. I have seen some HUGE heat sinks on some chips shown in the literature. Cost of course is another issue.

But Steve did say 2003 was the year of the laptop, so we can only hope that 970 will be seen in laptops before 2004. But logic still dictates that the justnow shipping G4 17 will have to be around for a few (6?) months before going 970.

Not necessarily, we have no idea how many Apple has made or plans to sell... it was a risk to begin with with such a large screen. A risk that seems to be paying off, especially with HP and one other company coming out with their own 17" versions later this year.

I think the biggest issue reason to switch to the 970 ASAP is because Apple is a "Company of firsts", and the title of the "First 64bit Notebook Computer" will be one of the greatest selling points for Apple... "free" publicity, even though it costs more in R&D and overhead.

As soon as it is technically possible, Apple will come out with the 970 Powerbook. But this does not mean that they will annouce/release it at WWDC. Just annoucing the 970 Pro would meet the requirements of the developers conference, being that it wouldn't matter to the developers if the 970-based machine is a desktop or powerbook. The 970 Powerbook will, therefore, be a surprize public announcement (most likely in late summer), with shipments in the 2-3 week range.

WWDC is still 3 months away... so we have to wait until at least then.

VIREBEL661
Mar 25, 2003, 05:27 PM
[QUOTE]Originally posted by strider42
Why "just" duals on the top end?... [/
QUOTE]

Remember the Daystar Genesis??? A very good point you make - make the top end REALLLLLY the TOP END... I mean, since we have a powerful Unix variant for an os - EXPLOIT IT.... Look at IBM's high end Unix workstations... Why shouldn't Apple compete on this level???

Just MHO!!!!

utilizer
Mar 25, 2003, 05:31 PM
Originally posted by mania
I have to completely disagree. the g4 is way past its prime with its fake ddr and limited clock speeds etc. assuming a 970 ends up in the desktop in june, i bet a 970 will be in the powerbook by january, october if we are lucky. you know once people start experiencing the 970 they will really want it in the powerbook, even if it were at slower speeds.

I'm gonna have to side with you mania. My house mate's 2 Ghz P4 smokes a 1 ghz TiBook, big time in Photoshop. It's an HP and it has the desktop processor in it before any of you ask; no it's not mobile! The machine still gives him hell 'cause it's Windows XP but I've convinced him to switch down to NT/2000 for the added stability.
It's completely ludicrous to me that Apple would even be considering to continue to place the G4 chip in their portables!
This is why I haven't even considered buying any of Apple's product at this time. I'm still on my Pismo 500 mhz G4 and it still is going along alright, nothing stellar.
But hey, when I buy the best, I want to know that that's what I'm getting. It sure does sound like it with these new 970s rolling out. I would trust this source, as I was very skeptical before when Apple didn't drastically change the casing, going from Quicksilver to Wind Tunnels. I was absolutely saddened when it was true and the guy that gave them the blueprints did indeed get caught! So yes, I trust them, even though they're a French site!

utilizer
Mar 25, 2003, 05:37 PM
But Steve did say 2003 was the year of the laptop, so we can only hope that 970 will be seen in laptops before 2004. But logic still dictates that the justnow shipping G4 17 will have to be around for a few (6?) months before going 970. [/B]

Logic told us that new PowerBooks wouldn't be released at MWSF since they were updated in November.
It's a different game plan now for Apple.

type_r503
Mar 25, 2003, 05:38 PM
Originally posted by nighthawk
Not necessarily, we have no idea how many Apple has made or plans to sell... it was a risk to begin with with such a large screen. A risk that seems to be paying off, especially with HP and one other company coming out with their own 17" versions later this year.

I think the biggest issue reason to switch to the 970 ASAP is because Apple is a "Company of firsts", and the title of the "First 64bit Notebook Computer" will be one of the greatest selling points for Apple... "free" publicity, even though it costs more in R&D and overhead.

As soon as it is technically possible, Apple will come out with the 970 Powerbook. But this does not mean that they will annouce/release it at WWDC. Just annoucing the 970 Pro would meet the requirements of the developers conference, being that it wouldn't matter to the developers if the 970-based machine is a desktop or powerbook. The 970 Powerbook will, therefore, be a surprize public announcement (most likely in late summer), with shipments in the 2-3 week range.

WWDC is still 3 months away... so we have to wait until at least then.

you can get sparc based laptops.

jettredmont
Mar 25, 2003, 05:48 PM
Originally posted by Shadowfax
I don't think the chip has anything on it to preclude dual processing. the system controller or something like that should decide about DP. the 970 may not be DP on new PMacs when they arrive, but that doesn't mean that there is no way to DP 970s.

"Dual processing" and "dual core" are completely different animals. "Dual core" means a single processor with two sets of logical units, two pipelines, etc (but of course shared cache and FSB between the cores).

In October, IBM described the 970 as being a single-core desktop variant of the Power4. As others have noted on these boards, there's more to a dual-core chip than just adding another core; there's all the support circuitry and logic which also needs to be put in place. Keeping the 970 single-core drastically reduces its complexity and hence reduces production costs.

Until IBM says otherwise, don't expect dual-core 970 chips.

Dual-processor (two separate chips working together) is of course a possibility.

nuckinfutz
Mar 25, 2003, 05:48 PM
My house mate's 2 Ghz P4 smokes a 1 ghz TiBook, big time in Photoshop

:rolleyes:

Ok so you actually want to compare a Desktop to a Portable? Of course the Desktop's going to win....duh...faster HD and 60watt processor.

Now compare a P4 Mobile to a 1Ghz Powerbook and come talk to us about that. ;)

utilizer
Mar 25, 2003, 06:28 PM
Originally posted by nuckinfutz
:rolleyes:

Ok so you actually want to compare a Desktop to a Portable? Of course the Desktop's going to win....duh...faster HD and 60watt processor.

Now compare a P4 Mobile to a 1Ghz Powerbook and come talk to us about that. ;)

Apparently I wasn't clear enough.
The HP is a laptop with a desktop processor in it. HP does this consistently with their machines on the higher end, which of course, he has.
So in the words of Donald Rumsfeld, boo-yah! :cool:
j/k :p

Kid Red
Mar 25, 2003, 06:31 PM
Originally posted by Flynnstone
Since this is a rumor, lets hypothesize on real world performance.
I've heard that the 970 is upto 4 times faster than an equivalent G4. Lets assume 2 times. Front side bus (FSB) is approximately 5 time faster (970@6.4 vs G4@1.3). For bus constrained, 2 times is reasonable. A little apples and oranges, no apples and prunes here:D . A 1.8 GHz 970 would roughly be equivalent to a 3.6 GHz G4, about 2.9 speed up factor.
Now lets use Adobe's recently released numbers for a P4 @ 3.06 GHZ and a dual G4 1.25GHz. Some say Adobe is not optimized for dual processors, so say a single and dual G4 are the same. The Adobe number for the G4 was 1 minute 25 seconds or 85 seconds. Take the 85 seconds for the G4 @ 1.25 and divide by our speed up factor of 2.9 and we get about 29 seconds.
This is (theoretically) almost twice as fast as a P4@ 3.06GHz.
Cool! :cool:

Actually, the 970 is 1.5-2x faster then a G4 at the same speed and the 970 1.8ghz is comparable to a 2.8ghz P4. So a dual 2.3ghz would trash anything current and future :)

Hawthorne
Mar 25, 2003, 06:52 PM
Originally posted by utilizer
Apparently I wasn't clear enough.
The HP is a laptop with a desktop processor in it. HP does this consistently with their machines on the higher end, which of course, he has.
So in the words of Donald Rumsfeld, boo-yah! :cool:
j/k :p

As I the sole iBook user in a school full of Wintel laptops, I am the subject of scorn for my choice of such an "inferior" product. The other students with laptop make pains to show off the power of their machines...

Once they lug their 8 lb monster over, that is, and wait the five minutes for it to wake up, then find a power plug because their battery just died. And then they wait for their machine to log onto the network drive wirelessly. Only then do they show me how fast they work on their machine. :D

In the mean time, my little "useless" 700 mhz iBook keeps puttering along, waking up from sleep instantly and working for hours on end without a recharge. I can get to the network drives thru the wireless faster and easier using Samba in the finder than they can using Win2000. And with all the apps I have in common with my iBook and the school's desktops (Photoshop, Illustrator, Dreamweaver MX, etc), I can see no discernable difference.

There's more to working fast than just processor speed. Having an iBook means I work when I want to, where I want to, not where the nearest power outlet is.

darndog
Mar 25, 2003, 06:53 PM
This sound's right, Apple is desperate for speedy chips and the 970's been obviously destined for macs since the inclusion of altivec, even the fab dates tie in.

Back in the day Apple used to have a top machine that was way in front of the rest of the pro range, with price tags to match. If these clock speeds are correct I would not expect the dual 2.3's to be anywhere close to the 1.8's, i'm thinking more like xServe prices.

I also think the G4 'handle' cases are long overdue for a redesign, the consumer lines (iMac, eMac, iBook) are all ice white, the pro stuff (Powerbooks, xServe & Raid) are all metal, the current G4 tower designs don't fit in anywhere. my guess is a vertical xServe, tall and thin with an aluminium (powerbook finish) case.

Jobs says this is the year of the laptop, which means either 'we still don't have a decent chip for our desktops' or 'by the end of the year our laptops will be faster than your desktops', lets hope its the latter.

darndog

shadowfax
Mar 25, 2003, 06:53 PM
Originally posted by utilizer
Apparently I wasn't clear enough.
The HP is a laptop with a desktop processor in it. HP does this consistently with their machines on the higher end, which of course, he has.
So in the words of Donald Rumsfeld, boo-yah! :cool:
j/k :p

Only if by "higher end" you mean "heavy" and "burn your legs/3rd leg hot," lol. we usually don't count 8-10 lb laptops that are 1.5-2 + inches "thin" as comparable to apple portables, as apple doesn't have a portable in that class (because it believes those aren't truly portables in the spirit of portability).:D ;) :p

but let's not start this discussion again ;)

barkmonster
Mar 25, 2003, 07:28 PM
I believe Cheetah was the code name for 10.0

Code names have nothing to do with the product anyway - Apple using Jaguar on released product marketing was a first, as far as I know.

I know apple's naming scheme for the next few years worth of OS X updates :D

Jaguar - Atari's 64bit console

Panther - Atari's scrapped 32bit console

Lynx - Atari's colour handheld gaming system

it all sort of goes wrong after that with silly names like, Falcon, TT, STe and 2600 that don't make a lot of sense.

Arcady
Mar 25, 2003, 07:28 PM
Originally posted by GeneR
I really want a time machine right now. Go ahead about, oh, a year, grab the newest, "bestest" Powerbook with the newest, "bestest" OSX and come back to the present, and... and... well, have more fun.

:D

Why not go 6 years into the future, buy a 5 year old PowerBook (still 1 year away from now) for like $200, and then come back? If there are plenty of them on eBay in 6 years, then you will also know the thing was built properly...

mathiasr
Mar 25, 2003, 07:30 PM
Originally posted by seamuskrat
What is fun ti imagine, is that the hypothetical design of the POWER5 derititive (maybe called the 980?!?) could perhaps be dual core. So, in 2 to 3 years we could see dual core/dual processor macs. Confused? heehee

The PowerPC 980 is rumored to be avaible during Q3 2004, in fact Apple could use the POWER5 itself.

NWFusion article about the POWER5 (http://www.nwfusion.com/news/2003/0204ibmplots.html)

Some quotes :

"The Power5 chip will replace the Power4 across the board, but improvements to the chip design and more attention to heat issues will help the new chip scale further down IBM's server line."

"Amezcua said that each team ultimately does what is right for the company, but that the pSeries team hopes to relegate Itanium to a niche in high-performance computing or better yet exterminate the processor altogether.

Always thickening the plot, IBM_may also throw another lower-end system into the mix by adding a server based on Advanced Micro Devices' 64-bit Opteron processor, Arimilli said. Arimilli claims that once Power5 arrives, it will be IBM's answer for low-end 64-bit computing."

djcobb44
Mar 25, 2003, 07:46 PM
So 2.3Ghz is not out of the question. Lets face it...Motorola had little to gain with their desktop cpu's, not only does IBM use these cpu's in their workstations, but it will compete with the wintel world and Unix/Linux servers too. So R&D with this family of chips will be greatly improved compared to moto(slow)a's half rear efforts.

rice_web
Mar 25, 2003, 07:53 PM
http://www6.tomshardware.com/howto/20030224/asetek_vapochill-13.html

Not to sound like the "PC Weenie," but the new batch of P4s, due any week now, will boast an impressive 800MHz cache (the 2.2GHz P4 with the 800MHz system bus slaughters the 3.0GHz P4 with a 533MHz bus).

But....

More importantly, it houses 1MB L2 cache, as opposed to the 512K in the current P4s.

So, the prospective benchmarks are quite worthless. However, I will add that SPEC is virtually the worst way to measure a processor (and a 4x jump in SPEC scores does not mean a 4x jump in performance; chips can be built to "cheat" at SPEC).

nuckinfutz
Mar 25, 2003, 08:06 PM
Rice

Daunting indeed. Prescott has the potential to be pretty stout. I think it could be the processor to send AMD back to the outhouse. That Toms Hardware article shows us exactly why these new PPC 970(boy am I optimistic) need to come sooner rather than later. Apple can beat Microsoft as far as OS innovation but the hardware must stay close.

Boy 2004 is going to have some wicked computers.

Apple suprise me....please suprise me!

rice_web
Mar 25, 2003, 08:24 PM
Oh wow, I'm getting to close to having an avatar....

Anyhow, Intel will release this update to the P4 very shortly, while the 970 will be in PowerMacs (optimistically, in my opinion) in late August, and there is no guarantee of a victory in the speed race.

I think it's unfortunate; these WWDC-announced processors seem unlikely to me, and even they may not suffice.

AidenShaw
Mar 25, 2003, 08:38 PM
Originally posted by nighthawk
and the title of the "First 64bit Notebook Computer" will be one of the greatest selling points for Apple...

Tadpole (http://hw.tadpole.com/html/) has been selling 64-bit SPARC and Alpha laptops for years.

shadowfax
Mar 25, 2003, 08:42 PM
Originally posted by AidenShaw
Tadpole (http://hw.tadpole.com/html/) has been selling 64-bit SPARC and Alpha laptops for years.

WOW. 650 MHz maybe, but jeez, 4 GB of RAM and 120 GB of HDD space. that's a workhorse!

Frobozz
Mar 25, 2003, 08:46 PM
Originally posted by rice_web
Anyhow, Intel will release this update to the P4 very shortly, while the 970 will be in PowerMacs (optimistically, in my opinion) in late August, and there is no guarantee of a victory in the speed race.

Well, I don't know about you... but I'd rather be (initially) 10% behind instead of 50% behind. The 970 and 980 will ramp up in speed very rapidly. I think the future's so bright, I'll have to wear shades.

eric_n_dfw
Mar 25, 2003, 08:57 PM
Originally posted by rice_web
http://www6.tomshardware.com/howto/20030224/asetek_vapochill-13.html

Not to sound like the "PC Weenie," but the new batch of P4s, due any week now, will boast an impressive 800MHz cache (the 2.2GHz P4 with the 800MHz system bus slaughters the 3.0GHz P4 with a 533MHz bus).

But....

More importantly, it houses 1MB L2 cache, as opposed to the 512K in the current P4s.

So, the prospective benchmarks are quite worthless. However, I will add that SPEC is virtually the worst way to measure a processor (and a 4x jump in SPEC scores does not mean a 4x jump in performance; chips can be built to "cheat" at SPEC).
What speed are the current G4's L3 cache's at?
EveryMac.com says 1/2 the processor seed, but has a "?" next to that.

If that's true, and they use DDR on the cache then that would be, effectively, 1.42 GHz cache memory bus speed!?!?! (I find that hard to beleive)

Does anyone have hard numbers for this?

Apple's pages are strangely silent on this, except to say that they provide 4GBs.

steve53e
Mar 25, 2003, 09:00 PM
Originally posted by Frobozz
WOW. You missed the point, didn't you? :-)

A 2.3 Ghz 970 is roughly equivalent to a G4 at twice that speed. How is 4 Ghz for ya? Just as you can not translate a P4 to a G4 in Mhz, you can't translate a G4 to a 970 easily. (though I tried)

Nope. Didn't miss the point at all. I'm well aware of everything you're saying. I'm just waiting for a machine that meets MY desired specifications before I buy another one. I waited for Apple to hit my price point on the 23" HDCD. Now I'm waiting for them to hit my spec point for a computer. That's all.

IJ Reilly
Mar 25, 2003, 09:10 PM
Originally posted by nighthawk
But the 970 is a whole different tamale... it would make the most sense to announce the 970 ProPPC Mac at WWDC, where the developers would need to know the most about it... even if the wait was 2 months, it would be official (and that is what Apple needs now more than ever).

Right. This is the current source of my optimism. I only hope I'm not looking for another in three months.

NavyIntel007
Mar 25, 2003, 09:11 PM
Originally posted by steve53e
Rumor or not.... I don't care if it takes a year or more, but I'm not giving Apple another dime for a computer until they release a box that clocks out at 2.5 ghz or better. Until then, 800 mhz will have to do and I'll gladly spend my money elsewhere.

MHz doesn't matter. Even intel will tell you that when their Centrino processor at 1.6 Ghz out performs the Pentium 4-M at 2.4 Ghz

NavyIntel007
Mar 25, 2003, 09:24 PM
Originally posted by seamuskrat
I tend to aggree. I think that to remain competative they HAVE to have laptops with new chips in the pro line. We KNOW thorugh the 12 PB that a G4 can be made small and fit in a small package. We can assume that the 970 will be usable in a laptop config but until we see usage specs and such we cannot know fur sure. I have seen some HUGE heat sinks on some chips shown in the literature. Cost of course is another issue.

But Steve did say 2003 was the year of the laptop, so we can only hope that 970 will be seen in laptops before 2004. But logic still dictates that the justnow shipping G4 17 will have to be around for a few (6?) months before going 970.

Not true. The 15" powerbooks had a minor redesign before the 12" and 17" came out. People were pissed about it, but Dell does it all the time. I think Apple will start moving in that direction, updating when things are available.

Stelliform
Mar 25, 2003, 09:28 PM
Originally posted by NavyIntel007
Not true. The 15" powerbooks had a minor redesign before the 12" and 17" came out. People were pissed about it, but Dell does it all the time. I think Apple will start moving in that direction, updating when things are available.

I would agree with that, but I think that apple is limited to this release schedule due to their size. They have to clear out old laptops before new ones can come in.

bobindashadows
Mar 25, 2003, 09:38 PM
Originally posted by utilizer
Logic told us that new PowerBooks wouldn't be released at MWSF since they were updated in November.
It's a different game plan now for Apple.

The powerbook, as we knew it the day before MWSF, was not updated at MWSF. Two new models were introduced. The 15"(15.2"?) Powerbook Titanium has not yet been updated, as follows from the product cycle.

porovaara
Mar 25, 2003, 09:46 PM
Originally posted by AidenShaw
Tadpole (http://hw.tadpole.com/html/) has been selling 64-bit SPARC and Alpha laptops for years.

ahh you haven't learned to think like Apple marketing yet:

The first consumer 64 bit notebook super computer!

NavyIntel007
Mar 25, 2003, 10:02 PM
Originally posted by Stelliform
I would agree with that, but I think that apple is limited to this release schedule due to their size. They have to clear out old laptops before new ones can come in.

The general gist of things are that if Apple went ahead with a mobile 970 they would make more money than if they waited to clear out inventory.

However, If we see 15" updates any time soon, there probably won't be 970's in the powerbooks by June.

Rincewind42
Mar 25, 2003, 10:03 PM
Originally posted by eric_n_dfw
What speed are the current G4's L3 cache's at?
EveryMac.com says 1/2 the processor seed, but has a "?" next to that.

If that's true, and they use DDR on the cache then that would be, effectively, 1.42 GHz cache memory bus speed!?!?! (I find that hard to beleive)

Does anyone have hard numbers for this?

Apple's pages are strangely silent on this, except to say that they provide 4GBs.

Actually, it's not hard at all to find this information, it's on Apple's Hardware Developer page here: http://developer.apple.com/techpubs/hardware/hardware.html. Under the section for the PowerMac G4 it lists the 3 models as follows:

CPU L3 Ratio
1 GHz 250 4:1
1.25 GHz 250 5:1
1.42 GHz 236 6:1

So the Dual 1.42 PowerMacs clock their L3 caches at 236Mhz DDR (effectively 472 Mhz). The L2 cache of the MPC7455 is 256k clocked at the processor speed.

The PPC 970 doesn't have connections for an L3 cache, but does feature a 512k L2 cache, and of course the now famous 900 Mhz bus.

Rocketman
Mar 25, 2003, 11:00 PM
Originally posted by dabirdwell
Now all we need is a 970 in a PowerBook.

How long?

Do we know the power usage figures?

The year of the portable?

Understatement?

Rocketman

http://v-serv.com/-upload/avatar.jpg

Get high.
And fast.

Rocketman
Mar 25, 2003, 11:05 PM
Originally posted by mikeyredk
lets keep the politics down or arn will shut this thread down like he did with the other one on gore

That thread was purely over the top.

Rocketman

Stelliform
Mar 25, 2003, 11:10 PM
Originally posted by NavyIntel007
The general gist of things are that if Apple went ahead with a mobile 970 they would make more money than if they waited to clear out inventory.

However, If we see 15" updates any time soon, there probably won't be 970's in the powerbooks by June.

The 15" is ripe for an update, but I hope they postpone the update for the 970's. But then the 17's wouldn't be the top portable. My bet is that they update the 15" and then they update the 17" with the 970's 4th quarter this year at the earliest.

NavyIntel007
Mar 25, 2003, 11:51 PM
Originally posted by Stelliform
The 15" is ripe for an update, but I hope they postpone the update for the 970's. But then the 17's wouldn't be the top portable. My bet is that they update the 15" and then they update the 17" with the 970's 4th quarter this year at the earliest.

Nope, they're going to update the 15" and the 17" at the same time. It doesn't make much sense to have them updated separately. That's like updating one powermac model but no the other two. Too bad for all those 17" buyers... suckers :p

NavyIntel007
Mar 26, 2003, 12:25 AM
Originally posted by moofy
What exactly are the odds that this site has it right 6 months in advance when lately it's been an anomoly if ANY rumors site gets it right a week before dates of known announcements like at the macworlds? Let's not get out hopes up.

Sadly, you have a good point. :(

It doesn't matter because I'm buying a 15" powerbook in June no matter what. The 970 would sweeten the deal.

Sun Baked
Mar 26, 2003, 01:52 AM
First week of April, strikes me as way too close to things coming together April 1, 2003...

:eek:

Though that is the best date for confirming the existance of the next "big" thing from Apple.

barkmonster
Mar 26, 2003, 06:01 AM
Not to sound like the "PC Weenie," but the new batch of P4s, due any week now, will boast an impressive 800MHz cache (the 2.2GHz P4 with the 800MHz system bus slaughters the 3.0GHz P4 with a 533MHz bus).

But....

More importantly, it houses 1MB L2 cache, as opposed to the 512K in the current P4s.

That test was overclocking the FSB without adjusting the multiplier. The 2.26Ghz P4 that was beating a 3.06Ghz P4 was actually running at 3.4Ghz (2.26 / 533) x 800 = 3.4

3.4Ghz with 6.4Gb/s of bandwidth is bound to be faster than 3.06Ghz with 4.2Gb/s of bandwidth. Especially when all test factors are equal (same cpu family, L2 cache size, motherboard, chipset etc...).

What's not impressive about that is the elaborate cooling used in the tests and the fact intels own roadmap (http://www.intel.com/products/roadmap/) lists the P4 as 3.2Ghz+ by the end of this year. This seems to put everything into perspective. Intel are obviously having a hard time getting any more performance out of the P4 using clockspeed alone so they're improving other areas of the design to squeeze as much life out of the design.

Here's the table off the 800Mhz FSB (http://www6.tomshardware.com/howto/20030224/index.html) article on tom's hardware.

The Shadow
Mar 26, 2003, 06:06 AM
Remember that rumour a day or 2 ago that Apple had called for tenders for a new motherboard with 970 chips?

Does anyone know anything about the tendering and/or tooling up and production processes etc?

Assuming that a tender is won at best at the end of March (or April if you want to be conservative) how long till production is in full swing, shipped to wherever PowerMacs are assembled and Macs are out the door?

(ie) Does this gel with the latest rumours that PowerMacs may be available 2 months after the WWDC? :confused:

blogo
Mar 26, 2003, 09:28 AM
Apple has tested the 970 in over a year now so I thikn it should be reasy for nest lineup.

ffakr
Mar 26, 2003, 09:29 AM
Originally posted by porovaara
ahh you haven't learned to think like Apple marketing yet:

The first consumer 64 bit notebook super computer!

Actually, Sun used to sell a portable with a 64bit SPARC in it. It was hella expensive, so I've heard.

Apple wouldn't be the first to market with a 64bit laptop.
also, there will be Athlon64 notebooks. I have a feeling that AMD will have mobile [mainstream] 64bit processors out before we see a 970 in a Powerbook.

Frobozz
Mar 26, 2003, 09:32 AM
Originally posted by steve53e
Nope. Didn't miss the point at all. I'm well aware of everything you're saying.

And yet none of that has anything to do with the GHz of the chip... you're waiting for a ficticious number. Buy what works. A price point on a monitor and a theoretical SPEC number on a CPU are two ttoally different things.

ffakr
Mar 26, 2003, 09:43 AM
Originally posted by some other guy
Not to sound like the "PC Weenie," but the new batch of P4s, due any week now, will boast an impressive 800MHz cache (the 2.2GHz P4 with the 800MHz system bus slaughters the 3.0GHz P4 with a 533MHz bus).

Too Late! you do sound like a PC Weenie ;-)
it isn't an 800MHz Cache, it is an 800MHz bus.

Originally posted by barkmonster
That test was overclocking the FSB without adjusting the multiplier. The 2.26Ghz P4 that was beating a 3.06Ghz P4 was actually running at 3.4Ghz (2.26 / 533) x 800 = 3.4

3.4Ghz with 6.4Gb/s of bandwidth is bound to be faster than 3.06Ghz with 4.2Gb/s of bandwidth. Especially when all test factors are equal (same cpu family, L2 cache size, motherboard, chipset etc...).


You also might want to take a minute and look at the benchmarks between the 400MHz bus P4s and the 533MHz bus P4s. The scores below are based on a 2.5GHz part.

Though the bus is 33% faster,
DX8 benchmarks are only 2.3% faster on the 533MHz bus.
DX7 benchmarks are 1.9% faster.
MP3 Maker Platinum is 1% faster.
Mpeg4 encoding is 3.9% faster.
SiSoft Sandra is FASTER on the SLOWER bus!
PC Mark 2002 is .6% faster on the 533MHz bus.

Apparently there is more holding back the P4 than just system bus. benchmarks available at Tom's Hardware (http://www6.tomshardware.com/cpu/20020506/p4b-05.html#benchmarks_under_windows_xp)

Intel will have to do more than increase the bus to 800MHz to make a huge difference. The L2 cache will help, but it will also prevent the cost of the parts from coming down since the die will contain many more transistors. Even on a smaller process, the chips will likely yield about the same quantity per wafer.

TMay
Mar 26, 2003, 12:59 PM
Now that IBM will be fabbing Nvidia FX, what are the possibilities of Apple licensing the FX core to drop onto 970 silicon to create 64 bit tablet chip. More than that, what would be the technical hurdles to overcome to accomplish this?

Rocketman
Mar 26, 2003, 01:13 PM
Originally posted by TMay
Now that IBM will be fabbing Nvidia FX, what are the possibilities of Apple licensing the FX core to drop onto 970 silicon to create 64 bit tablet chip. More than that, what would be the technical hurdles to overcome to accomplish this?

Stop with the tablet talk.

It ain't gonna happen and that is a technical statement.

HOWEVER.

Some phone variant could. Think of the Nokia 9000 or whatever it is called that is a rectangular phone box that opens up to show a keyboard and a display. Now imagine a variant of that with a blueberry keytboard and a clamshell display with displays on both faces of the clamshell to double the screen real estate.

Now imagine it as tri or quad band cellular including internet, bluetooth, possibly even 802.11g. Then imagine it with OSX portable and rondevoux.

Imagine it with voice recording features, a convenience camera and a video capability, all CMOS.

Now imagine it selling for under $1200, possibly as low as $499 with service contracts attached.

Stop imagining. Several companies are coming out with very similar devices really soon. I would vastly prefer to buy mine from Apple Computer no matter what wacky guys are on the board.

Rocketman

nighthawk
Mar 26, 2003, 01:30 PM
Originally posted by Rocketman
Stop with the tablet talk.

It ain't gonna happen and that is a technical statement.


Personally, I don't think that Tablet PCs (I think Digital Assistant) are not going to take off in their present form either. But what I do expect to see in the nearish future is some sort of optical (glasses, whatever) screen to interact with instead with some glove technology...

Yes, I know that sounds like VR, but think of this... clear glasses that interpose 2D pixel data on-top of real world... displaying a virtual keyboard for you to type midair, and controlling virtual movement in the similar to the way Black and White (the game) does with it's own hand.

mathiasr
Mar 26, 2003, 02:07 PM
Originally posted by ffakr
Apparently there is more holding back the P4 than just system bus. benchmarks available at Tom's Hardware (http://www6.tomshardware.com/cpu/20020506/p4b-05.html#benchmarks_under_windows_xp)

Intel will have to do more than increase the bus to 800MHz to make a huge difference. The L2 cache will help, but it will also prevent the cost of the parts from coming down since the die will contain many more transistors. Even on a smaller process, the chips will likely yield about the same quantity per wafer.

The 3.2 GHz P4 is expected next quarter, they will also bring new chipsets to the market : Canterwood and SpringDale with dual DDR 400 support and a new technology called PAT (Performance Acceleration Technology) to speed up memory access.

Details about Canterwood (http://endian.net/details.asp?ItemNo=3773)

Pentium 4 Intel chipsets (http://www.geek.com/procspec/chipsets/p7intel.htm)

eric_n_dfw
Mar 26, 2003, 03:10 PM
Originally posted by nighthawk
Personally, I don't think that Tablet PCs (I think Digital Assistant) are not going to take off in their present form either. But what I do expect to see in the nearish future is some sort of optical (glasses, whatever) screen to interact with instead with some glove technology...

Yes, I know that sounds like VR, but think of this... clear glasses that interpose 2D pixel data on-top of real world... displaying a virtual keyboard for you to type midair, and controlling virtual movement in the similar to the way Black and White (the game) does with it's own hand. TechTV recently had a review on some kind of headset like you are speaking of - gave the reviewer headaches big time and made walking with it on very dangerous. It would have to be something that ws able to adjust to stay in focus - I do not know how that could be done. (Can you determine an eye's field of vision by looking at it?) Maybe if it projected images directly on the retina, but the lens would still distort things I would think.

NavyIntel007
Mar 26, 2003, 04:09 PM
Exactly. People all over town would be running into mail boxes and light posts. :D

ffakr
Mar 26, 2003, 04:57 PM
Originally posted by mathiasr
The 3.2 GHz P4 is expected next quarter, they will also bring new chipsets to the market : Canterwood and SpringDale with dual DDR 400 support and a new technology called PAT (Performance Acceleration Technology) to speed up memory access.

Details about Canterwood (http://endian.net/details.asp?ItemNo=3773)

Pentium 4 Intel chipsets (http://www.geek.com/procspec/chipsets/p7intel.htm)

As a mac AND pc user, I hope that "Performance Acceleration Technology" is more impressive than "Quantaspeed Architecture", AMDs fancy sounding name for the return of the performance rating. ;-)
For that matter, I hope it is more impressive than HT, which has shown a -10% to 20% speed boost so far (while doubling some software licensing costs) ;-)

ffakr
Mar 26, 2003, 05:04 PM
Originally posted by TMay
Now that IBM will be fabbing Nvidia FX, what are the possibilities of Apple licensing the FX core to drop onto 970 silicon to create 64 bit tablet chip. More than that, what would be the technical hurdles to overcome to accomplish this?

IBM makes chips for a LOT of people...
They have zero intellectual property over what they produce. This will in no way affect whether or not Apple has special access to Nvidia chips or technologies.

I don't know why you'd want to put a the FX core on the die with the 970 anyway.
Would it be because you want a 150 Million gate processor that ran at 400MHz and gave off 100 watts of heat?
Or, maybe because you want the traces from the 970 to wrap around the FX core, so the signals are delayed...
Or, maybe you want to move the graphics chip as far away from the video ram as possible.

TMay
Mar 26, 2003, 06:30 PM
Originally posted by ffakr
IBM makes chips for a LOT of people...
They have zero intellectual property over what they produce. This will in no way affect whether or not Apple has special access to Nvidia chips or technologies.

I don't know why you'd want to put a the FX core on the die with the 970 anyway.
Would it be because you want a 150 Million gate processor that ran at 400MHz and gave off 100 watts of heat?
Or, maybe because you want the traces from the 970 to wrap around the FX core, so the signals are delayed...
Or, maybe you want to move the graphics chip as far away from the video ram as possible.

I should have been more explicit about Apple licensing the FX from NVidia. Either way, building a system on a chip may be attractive to Apple (which is why I brought it up) and may be feasible at .09nm or .065nm, which the Fishkill facility will be able to handle. The point is, that Apple is in the position to customize the PPC with cores such as a TCP/IP network processor, (which IBM may add to the 980), DSP, and, yes, even a graphics processor core, though, I have to admit, throwing "tablet" in is in bad taste on my part.

rice_web
Mar 26, 2003, 06:59 PM
Originally posted by ffakr
Too Late! you do sound like a PC Weenie ;-)
it isn't an 800MHz Cache, it is an 800MHz bus.


Oops, typo. Anyhow, benchmarks like the SiSoft and SPEC benchmarks do little to effectively guage a processor's speed.

The Quake III benchmark showed a 10% jump from a 3.4GHz P4 with an 800MHz to a 3.0GHz P4 with a 533MHz bus; a jump nearly in line with the percentage jump in megahertz. Increases like these are exceptional. In the benchmarks where the 3.4GHz overclocked P4 is beaten by a 3.0GHz P4, remember that this chip is exactly that: overclocked.

steve53e
Mar 26, 2003, 10:42 PM
Originally posted by Frobozz
And yet none of that has anything to do with the GHz of the chip... you're waiting for a ficticious number. Buy what works. A price point on a monitor and a theoretical SPEC number on a CPU are two ttoally different things.

Do you have a learning disability or something?

I bought a monitor when it met my price point.

Now I'm waiting to buy a computer that meets my desired specifications (not SPEC !).

Fairly simple concept.... or so I thought.

rjwill246
Mar 26, 2003, 11:19 PM
What I'd like to know is how you pronounce MacBidouille in the first place.

I'm a dumb American that only speaks one language. Is it "Mac-Bi-dooey"? "Mac-Bi-dooley"? "Mac-Bi-day"?



The closest you can get in US English is:
Muc -bee -doo -wee, run together quickly

Cubeboy
Mar 27, 2003, 11:52 AM
IBM has stated that the 1.8 ghz PPC970 will offer comparable to a 2.8 ghz Pentium 4, which goes in line with the SPECmarks. Prescott is going to offer a pretty decent increase in performance to the Pentium 4 line. Here's a article about 800 mhz bus performance.

http://www.anandtech.com/chipsets/showdoc.html?i=1806

As we can see, the system bus increases performance fairly significantly, remember that the P4 using 800 mhz bus was using a unoptimized chipset and memory, and is bottlenecked by DDR that only runs at 666 mhz tops. With the new chipset and DDR400 memory, I'm expecting the complete fsb to increase performance to around 15%.

I've also made some estimations of Prescott performance on another forum listed below.

http://forums.macnn.com/showthread.php?threadid=152602

Regardless, dual processor PPC970 2.3 will be the processor for digital video/audio editing, 3d rendering, and other bandwidth intensive apps.

sparkleytone
Mar 27, 2003, 02:18 PM
Originally posted by steve53e
Do you have a learning disability or something?

I bought a monitor when it met my price point.
...
Fairly simple concept.... or so I thought.

here at macrumors we tend to dislike newbies who come in and spout trolls, flames, and overall idiocy. if you are going to be offensive while making dumb comments like 'i wont buy unless its 2.5GHz' then don't bother registering.

on a different note, i just assumed it was mac-boo-dwah but i'll ask my g/f later.

ktlx
Mar 27, 2003, 02:43 PM
Originally posted by sparkleytone
here at macrumors we tend to dislike newbies who come in and spout trolls, flames, and overall idiocy. if you are going to be offensive while making dumb comments like 'i wont buy unless its 2.5GHz' then don't bother registering.

Why is "I won't buy unless its 2.5Ghz" such a dumb comment?

And he never got offensive until Frobozz basically called him stupid for waiting to buy a machine with a certain speed processor in it. I don't recall ever seeing steve53e calling anyone stupid for buying the first PowerPC 970 machines that came out. He simply said he would wait until a certain future point.

Why is that automatically stupid?

Isn't it at least possible that he could have looked at the "published" SPEC information for the various speed PowerPC 970s and decided that something around 2.5Ghz gave him a large enough increase in speed to justify spending probably $3500 on a new machine?

The only thing I found offensive was you and Frobozz calling this guy various names just because he wants to wait until probably the second version of the PowerMac PPC 970. I don't find that all unreasonable after Apple dorked up the PowerMac MDD v1 with a loud power supply. The changes to support a PowerPC 970 will be considerably more difficult and while I don't expect Apple to make a boo-boo, I also did not expect such a loud power supply.

Frobozz
Mar 27, 2003, 11:08 PM
Originally posted by steve53e
Do you have a learning disability or something?

I bought a monitor when it met my price point.

Now I'm waiting to buy a computer that meets my desired specifications (not SPEC !).

Fairly simple concept.... or so I thought.

So you thought! You smart, smart man. ;-)

Frobozz
Mar 27, 2003, 11:30 PM
Originally posted by ktlx
Why is "I won't buy unless its 2.5Ghz" such a dumb comment?

And he never got offensive until Frobozz basically called him stupid for waiting to buy a machine with a certain speed processor in it....

...The only thing I found offensive was you and Frobozz calling this guy various names just because he wants to wait until probably the second version of the PowerMac PPC 970....

You might want to go back and read my post. I never called anyone a name. I replied to his first post with "whoa, you missed the point, didn't you?" Yeah, you could say that was agressive in tone... but was more a response to get a newbie to chill out on his previously agressive replies to innocent posts. I apologize if my tone incited a flame war... but it was sincerely meant as irony-- one outlandish comment for another. Anyone who's read more than a couple of my posts knows this is my tone and it is not slanderous or mean in any way. I should have been more thoughtful and put a smiley or something after it.

I simply think waiting for 2.5 GHz as an arbitrary number is fairly short sighted. Especially since a 2.5 GHz 970 is comparable to a 4.x GHz G4. See my point? Does he want a 2.5 GHz G4 or a 2.5 GHz 970? There's roughly a 200% difference in performance. That's a whole lot of difference, and it has zipity-do-dah to do with the GHz rating. That's my whole point. No need to get looney about it. You do have some good points in your post, though.

The only post in that thread that had some mud-slinging in it was when I was accused of having a "learning disability." LOL. Seriously, I got a good chuckle. I would have assumed he was joking about the "learning disability", except nearly all his posts in this thread have had negative or aggressive undertones... so I'll take it at face value... immature.

Oh, and sparkleytone... thanks for sticking up for me, bro. Sheesh. I need to get some flame-retardent. :-) :-)

steve53e
Mar 28, 2003, 07:30 AM
Originally posted by Frobozz
I replied to his first post with "whoa, you missed the point, didn't you?" Yeah, you could say that was agressive in tone... but was more a response to get a newbie to chill out on his previously agressive replies to innocent posts.

I simply think waiting for 2.5 GHz as an arbitrary number is fairly short sighted. Especially since a 2.5 GHz 970 is comparable to a 4.x GHz G4. See my point? Does he want a 2.5 GHz G4 or a 2.5 GHz 970? There's roughly a 200% difference in performance. That's a whole lot of difference, and it has zipity-do-dah to do with the GHz rating. That's my whole point. No need to get looney about it. You do have some good points in your post, though.

The only post in that thread that had some mud-slinging in it was when I was accused of having a "learning disability." LOL. Seriously, I got a good chuckle. I would have assumed he was joking about the "learning disability", except nearly all his posts in this thread have had negative or aggressive undertones... so I'll take it at face value... immature.

Oh, and sparkleytone... thanks for sticking up for me, bro. Sheesh. I need to get some flame-retardent. :-) :-)

Previously agressive replies to innocent posts? In my post "I'm waiting" I made a simple statement that I was waiting for Apple to offer a computer with a particular processor before I spent anymore money on a new computer. That's aggressive?

It was you who implied I didn't have a clue what I was talking about and that it was silly/short-sighted for me to do such a thing. I responded that I didn't miss the point and explained my reasoning. That's aggressive? Yet in every instance I explained myself, it's you who felt it necessary to continue with your badgering, hence, my final response.

As for being a newbie. The number of post I make to a particular forum is hardly the defining point in one's life with regard to their knowledge or use of computers. Yet to you, the fact that you've been a member of this forum a whole 2 months longer than I have is.

.... except nearly all his posts in this thread have had negative or aggressive undertones... so I'll take it at face value... immature.

Odd that it appears you're the only one who seems to have seen it that way. Well, you and sparkleytone.

Maybe some day we can sit down over a beer and talk about Dual Digital AFCS Servos, Inner Loop and Outer Loop functions and the general process of converting electrical signals from computers into mechanical motion to move the main and tail rotors on large helicopters. Or maybe we can just chat about the events we've witnessed over the past 45 years and how they've affected our daily lives. I'll leave that up to you. That is if you're old enough to drink. :)

ktlx. Thanks! Nice to know someone here understood what I was saying. With that said, I'll be departing this hostile forum.

Frobozz
Mar 28, 2003, 10:35 AM
Originally posted by steve53e

Maybe some day we can sit down over a beer and talk about Dual Digital AFCS Servos, Inner Loop and Outer Loop functions...

God, I hope not. :-)

steve53e
Mar 28, 2003, 11:42 AM
Originally posted by Frobozz
God, I hope not. :-)

Didn't think so. :-) Truce.

keithcobbett
Mar 28, 2003, 02:55 PM
This is a great story and hopefully we will see some faster processors. But, I have 1 teenie weenie issue that I would like to address. How about the bus speeds. I don't care if you are running a 3 GHz processor, if you are only communicating over a 133MHz bus speed, it is killing you. Will this be upgraded????

Wes
Mar 28, 2003, 03:21 PM
Originally posted by keithcobbett
This is a great story and hopefully we will see some faster processors. But, I have 1 teenie weenie issue that I would like to address. How about the bus speeds. I don't care if you are running a 3 GHz processor, if you are only communicating over a 133MHz bus speed, it is killing you. Will this be upgraded????

I don't think 3ghz processors have 133 bus speed. If they did, yes, that would hurt the processor. Doesn't the 970 have a 900 mhz fsb? Currect pcs have 400+. Please correct me.:eek:

Rincewind42
Mar 28, 2003, 04:58 PM
Originally posted by Wes
I don't think 3ghz processors have 133 bus speed. If they did, yes, that would hurt the processor. Doesn't the 970 have a 900 mhz fsb? Currect pcs have 400+. Please correct me.:eek:

Oddly enough, this is one of those strange issues =).

The current P4 uses a Quad-Pumped 133 Mhz bus (the previous generation used Quad-Pumped 100Mhz) that gives an effective rate of 533 (and previously 400) Mhz. Of course, this isn't as efficient as a 533Mhz Single-Pumped bus, but it does have the same peak theoretical bandwidth.

The PowerPC 970 uses a 450Mhz Double Pumped bus (for the 1.8 Ghz part at least - no solid information is available for what speed the bus would be for other speed ratings - yet). This is effectively a 900 Mhz bus.

mathiasr
Mar 28, 2003, 05:31 PM
Originally posted by Rincewind42
The PowerPC 970 uses a 450Mhz Double Pumped bus (for the 1.8 Ghz part at least - no solid information is available for what speed the bus would be for other speed ratings - yet). This is effectively a 900 Mhz bus.

More precisely it has two point-to-point unidirectional 32 bits busses. Other architectures tend to use a single bidirectional 64 bits bus.

The Shadow
Mar 29, 2003, 01:57 AM
Originally posted by Macrumors
MacBidouille (http://www.macbidouille.com/niouzcontenu.php?date=2003-03-25#5087) feeds the fire with more rumors about the PowerPC 970 and WWDC.

According to their most recent update, they claim that Apple's WWDC dates were indeed moved due to planning around the PowerPC 970. Configurations are reported at 1.4 GHz, 1.8 GHz, and dual 2.3 GHz with availability 6-8 weeks after the announcement.




After reading many of the previous posts in this thread, I personally believe that if Apple is going to restore cred. to the PowerMac lineup, they should not go with the 1.4 GHz PowerPC 970 CPU. So, if that means that only the 1.8 GHz CPU is initially available, then there should only be 2 PowerMacs in the lineup - a single and dual processor running the 1.8.

I just believe that it is desireable the at the current price point, the low end PowerMac should equal the top end PC's in performance. Afterall, even at this price, the Apple buyer is still paying a premium over high end PC's.

Logic would suggest that if IBM did not produce any 1.4 MHz chips, it could produce more 1.8 MHz chips.

I believe that for Apple to release the new 1.4 MHz chip, given current price points, would result in embarrassment. What are the PC magazines going to say when they test Apple's latest "salvation" and find it is still not up to muster?

And as an Apple consumer, what about our value for money?

nuckinfutz
Mar 29, 2003, 03:06 AM
After reading many of the previous posts in this thread, I personally believe that if Apple is going to restore cred. to the PowerMac lineup, they should not go with the 1.4 GHz PowerPC 970 CPU. So, if that means that only the 1.8 GHz CPU is initially available, then there should only be 2 PowerMacs in the lineup - a single and dual processor running the 1.8.

Why? That would be Apple supporting the "Megahertz Myth" Apple is somewhat isolated because they're not competing with any ther Mac compatible computers. Intel has come out with Centrino Chipsets which are clocked slower but deliver greater performance than P4M. It's not that hard to market.

Logic would suggest that if IBM did not produce any 1.4 MHz chips, it could produce more 1.8 MHz chips.

Logic has nothing to do with processor yields. Some cores will not handle running reliably at 1.8Ghz so they can be run fine at a lower speed. In a perfect world all your chips would run at full speed but I doubt that will be the case here. Apple will need the full breadth of the line.


I believe that for Apple to release the new 1.4 MHz chip, given current price points, would result in embarrassment. What are the PC magazines going to say when they test Apple's latest "salvation" and find it is still not up to muster?

Once again you're assuming that all people are fooled by a simple metric. Worries about megahertz will be erased upon the first benchmark. Even at 1.4Ghz the PPC 970 is a formidable chip. Most likely twice as fast as the G4 @ 1.4Ghz

The Shadow
Mar 29, 2003, 07:35 AM
Originally posted by nuckinfutz
Why? That would be Apple supporting the "Megahertz Myth" Apple is somewhat isolated because they're not competing with any ther Mac compatible computers...

...Once again you're assuming that all people are fooled by a simple metric. Worries about megahertz will be erased upon the first benchmark. Even at 1.4Ghz the PPC 970 is a formidable chip. Most likely twice as fast as the G4 @ 1.4Ghz

I agree with what you are saying, per se, and your opinions are helpful.

Forgive me, but I thought the "meghead" myth (blame Red Dwarf) related to comparisons between chips of different architecture, not between different, for example, PPC 970 chips. But I expect a G4 1.25 GHz chip to perform better then my G4 533 MHz chip, because I generally expect the naming conventions to be internally consistent, although I also know that's not always the case.

Clearly, I'm led to believe the 1.8 GHz PPC 970 chip will be welcome, compared to high end Intel chips in 6 to 8 months, so clearly (so I thought) I don't subscribe to the alleged meghead myth.

My comments were based on previous posts, such as one quoting IBM as comparing the performance of the 1.8 GHz PPC 970 to the current Intel P4 2.8 MHz chips. If it's better than that, that will be great.

The point you made about chip production is a great point, as it relates to the number of chips of a certain spec being available. I'd forgotten that as its not my field. I'm a user only.

So all jargon aside, my concern, is that by the end of this year, the low end PowerMac chip will need to be at least as good as the current P4 2.8 chip, which by then will be going cheap and slow.

So, unless IBM has been misquoted, which is of course quite possible, Apple should ditch the PPC 1.4 MHz chip (or knock another 40% off the price of the low end PowerMac - which will never happen). Perhaps they should put the 1.4's in the iMacs - there's a sensible thought. It won't happen either.

nuckinfutz
Mar 29, 2003, 10:32 AM
So, unless IBM has been misquoted, which is of course quite possible, Apple should ditch the PPC 1.4 MHz chip (or knock another 40% off the price of the low end PowerMac - which will never happen). Perhaps they should put the 1.4's in the iMacs - there's a sensible thought. It won't happen either

You'll get no complaints from me here. If this was to happen it would mean IBM is fabbing PPC 970s and getting great yields. This sounds definitely plausible as IBM just snatched Millions of Fab work from TMSC for producing Geforce chips :)

If the rumors of 2000+ PPC 970s are true then we'd be looking pretty damn good. I'm trying to temper my enthusiasm somewhat because we've been hoping for "the next thing" for so long. At any rate I think the direction we have is going well. If Apple is agressive we just might see what you've posted here.

socokid
Mar 31, 2003, 01:17 PM
Take a look. (http://e-www.motorola.com/collateral/PPCRMAP.pdf)

Just a curious post in the Apple FAQ part.

post link (http://discussions.info.apple.com/WebX?14@22.ZU8OaqdljG3.21@.3bbff83f)

Cubeboy
Apr 4, 2003, 12:11 PM
Originally posted by The Shadow
I agree with what you are saying, per se, and your opinions are helpful.

Forgive me, but I thought the "meghead" myth (blame Red Dwarf) related to comparisons between chips of different architecture, not between different, for example, PPC 970 chips. But I expect a G4 1.25 GHz chip to perform better then my G4 533 MHz chip, because I generally expect the naming conventions to be internally consistent, although I also know that's not always the case.

Clearly, I'm led to believe the 1.8 GHz PPC 970 chip will be welcome, compared to high end Intel chips in 6 to 8 months, so clearly (so I thought) I don't subscribe to the alleged meghead myth.

My comments were based on previous posts, such as one quoting IBM as comparing the performance of the 1.8 GHz PPC 970 to the current Intel P4 2.8 MHz chips. If it's better than that, that will be great.

The point you made about chip production is a great point, as it relates to the number of chips of a certain spec being available. I'd forgotten that as its not my field. I'm a user only.

So all jargon aside, my concern, is that by the end of this year, the low end PowerMac chip will need to be at least as good as the current P4 2.8 chip, which by then will be going cheap and slow.

So, unless IBM has been misquoted, which is of course quite possible, Apple should ditch the PPC 1.4 MHz chip (or knock another 40% off the price of the low end PowerMac - which will never happen). Perhaps they should put the 1.4's in the iMacs - there's a sensible thought. It won't happen either.

I remembered the claim was made by IBM on their pdf presentation a while ago. If anyone has a copy of that pdf, they could elaborate on this.

Remember that PPC 970 chips could be used to great effect in dual configurations, and because of their fast system buses, they won't be plagued bandwidth problems like the G4. In dual configurations using well threaded programs like final cut pro and OSX, a dual 1.4 ghz PPC970 system well be more than a match for a 2.8 ghz P4.

mathiasr
Apr 4, 2003, 04:40 PM
Originally posted by socokid
Just a curious post in the Apple FAQ part.

post link (http://discussions.info.apple.com/WebX?14@22.ZU8OaqdljG3.21@.3bbff83f)

On the other hand the PowerPC is far from being a dead product line, Motorola announced the new PowerPC 5xx family aimed at extreme temperatures environments.

http://www.motorola.com/mediacenter/news/detail/0,1958,2555_2087_23,00.html

They developed a new DSP like extension to the PowerPC instruction set called SPE (Signal Processing Engine).

http://www.simdtech.org/spe

Of course these microcontrollers have no interest for Apple. Motorola has definitely focused on other markets, and IBM will probably become the sole high end PowerPC supplier.

ffakr
Apr 4, 2003, 05:11 PM
Originally posted by Rincewind42
Oddly enough, this is one of those strange issues =).

The current P4 uses a Quad-Pumped 133 Mhz bus (the previous generation used Quad-Pumped 100Mhz) that gives an effective rate of 533 (and previously 400) Mhz. Of course, this isn't as efficient as a 533Mhz Single-Pumped bus, but it does have the same peak theoretical bandwidth.

Why would you say that a quad pumped bus is less efficient than a 'single-pumped' bus (never heard that slang)??

As far as the system is concerned... all that matters is how many 'ticks' per second are available. If you have 533 Million ticks per second, you have 533 Million chances to do something since each clock tick is a signal to the system.

There are currently two ways to generate these ticks. You have a 'clock' (usually a quartz chip) that ticks off time.... tick, tick, tick. You can listen for each cycle generated and therefore get one tick per Hz, or you can listen to each cycle at different points and get multiple ticks per Hz.

On an old SDR Bus, say a 133MHz bus, the quartz clock ticks off 133 Million times per second.
On a 'quad-pumped' [QDR] bus, the quartz clock may still tick at 133MHz, but you use multiple point on each cycle to represent a 'tick'. this allows you more signals to do something. The beauty of DDR and QDR is that you don't have to push the clock speed up to make it do more things. Try running an SDR DIMM and see what happens at 533MHz... the electrons hop from one gate to another, it overheats....bad stuff.

So,.. this is how it actually works... think of a sine wave on a typical x,y graph...
The wave goes up, the wave goes down. For each cycle (Hz) it accends, peaks, decends, crosses x, decends, peaks(-), accends, crosses x... and so on, and so on.
With SDR, the 'tick' comes after a full cycle.
With DDR, the 'tick' comes when the signal crosses x=0... so, you get a 'tick' while decending to x=0 and a 'tick' while accending to x=0.

Now... how is this any less efficient than an SDR solution?
It is actually significantly MORE efficient as you can generate significantly higher effective 'clock rates' without actually raising the clock and buring out your components.

And another thing, If it isn't as efficient, how could it have [as you assert] the same bandwidth? As far as the rest of the system, the QDR subsystem is running at an effective 533MHz (in this example). The idea of doubling or quadrupling a clock rate in this manner doesn't require increased latency.

:confused: :)

Rincewind42
Apr 5, 2003, 01:47 AM
Originally posted by ffakr
Why would you say that a quad pumped bus is less efficient than a 'single-pumped' bus.

As far as the system is concerned... all that matters is how many 'ticks' per second are available. If you have 533 Million ticks per second, you have 533 Million chances to do something since each clock tick is a signal to the system.

On an old SDR Bus, say a 133MHz bus, the quartz clock ticks off 133 Million times per second.

On a 'quad-pumped' [QDR] bus, the quartz clock may still tick at 133MHz, but you use multiple point on each cycle to represent a 'tick'. this allows you more signals to do something.

Now... how is this any less efficient than an SDR solution? It is actually significantly MORE efficient as you can generate significantly higher effective 'clock rates' without actually raising the clock and buring out your components.

And another thing, If it isn't as efficient, how could it have [as you assert] the same bandwidth? As far as the rest of the system, the QDR subsystem is running at an effective 533MHz (in this example). The idea of doubling or quadrupling a clock rate in this manner doesn't require increased latency.

:confused: :)

ArsTechnica has an article that describes how all this works in current memory technology. It is here: http://arstechnica.com/paedia/b/bandwidth-latency/bandwidth-latency-1.html

The gist of the article is that 200 Mhz DDR (400 Mhz effective) is less efficient than 400 Mhz SDR due to latency. No matter what you do to the bus, a 200Mhz bus will always take twice as long to deliver a packet than a 400 Mhz bus will. A 100Mhz QDR (400 Mhz effective) will take 4 times as long as the 400Mhz SDR bus. They will all deliever the same theoretical max bandwidth because they all have 400 million events on which to transfer data, but the SDR bus can get that first transmission from Point A to Point B 4 times faster than the QDR bus with the same bandwidth. All a DDR or QDR system does is pack the ticks closer to each other, it doesn't make them travel faster.

Look at it this way. Just because you can put more runners on the track doesn't mean that you can get the first runner around the track faster. All that double/triple/quadruple data rate does is get more runners on the track.

That initial read latency is why a 400x1 bus is superior to a 200x2 which is superior to a 100x4 bus.

As for burning out components, that is really a side issue. If the component manufacturers could create components that ran at higher rates they would, if only to make double pumped memory that they could sell as PC-8500 (DDR1066). Until they can we will have to use DDR/QDR products that just aren't as efficient as equivalent SDR products.

socokid
Apr 6, 2003, 09:51 AM
Originally posted by mathiasr
[B]On the other hand the PowerPC is far from being a dead product line, Motorola announced the new PowerPC 5xx family aimed at extreme temperatures environments.


Very interesting reads. It sure seems the hard core desktop CPU's are on the wayside though.

Hmmm, the next few months will be very interesting. I can't wait...