970 Motherboard and Chip Delivery?

[IVIIVI4ck3y27]

Re: Please, don't let that be true...

Originally posted by Bengt77
I surely hope you'll be proven wrong here. Of course, and I know that too, there has indeed been no announcements whatsoever (perhaps maybe that PDF MacBidouille was speaking of, that came from the Micro Processor Forum) regarding Apple's commitment to using the PPC970 in their next lines of computers. But still, every bit of information now available (or should 'bit of information' be stated as being 'rumour') actually does point in the direction of it.

Again, I really hope you'll be proven wrong. I'd hate to see the xServe get them first, and then Power Macs only next year, or something. I don't know how big this rumourmill is, but I'm sure it's not that the whole Mac community is expecting the PPC970. The majority of the community probably doesn't even know of it's very existence. Apple could care less (probably, but hopefully not) about 'us' rumour mongors. So why would Apple feel the pressure just because all the rumours say it will do this or do that? They never have. Look at the iMac. Mac rumour sites said, even like one and a half year before the very introduction of the LCD iMacs, that a new model was coming. The whole community was deeply dissapointed because some rumour turned out to be wrong. That's a shame. Why should Apple be penalised becaus there are some rumours saying they would do A, while they did B?

Some rumours are wrong. Some rumours are right. Who knows which are what? This one might prove right. And hopefully rather sooner than later. I'd love Apple to have some serious power computing ready to be introduced at the Create-Expo-or-whatever-the-new-name-was-again. But just don't bet on it.

Darn, that's exactly what you said. Okay, so I agree with you, but I just hope you're wrong...



Edit: the bold 'right' was 'wrong' until just yet. That was not what I meant to say (or write, for that matter)!

Again... there's no visible evidence nor will there be any that there's new PowerMacs using 970's until release. I mean, nothing is for certain "UNTIL" it happens... that's just a truth in life.

Yet, if IBM is bothering to make PowerPC chipsets with SIMD when they, themselves, won't use them... why? If they weren't meant for Apple... IBM would just make a further evolution of the PowerPC without SIMD and just focus it towards their own needs... which have a decisively server slant.

Also... what would Apple use?

Motorola's officially announced that their G5, as previously announced, is dead as a door nail. Kaput. Not happening...

So Apple is stuck with Motorola and the G4 forever?

Riiiiiight...

Apple moving to Intel/AMD?

Ooooookay... so Apple, so quickly after getting the OS rewritten, integrating Carbon, and trying to optimize as much as they can for PowerPC is just going to bail?

pffffffftttt ::snickering::

Ain't no way.

For those rumor mongers hoping for AMD/Intel....

Keep dreaming, in 5-10 years time you might get your wish.

But not now...

Let's face it... didn't Apple just publically say they're tied to PowerPC for the near term, but after that it could get interesting?

Count on PPC 970, it's Apple's strongest offering on the table.

With SIMD that is "BACKWARDS COMPATIBLE" with AltiVec?!?

It's "TAILOR MADE".

In fact, I'd not be surprised if Apple didn't even commission to have this made, and foot some of the R&D $ just so they could continue to compete. They're losing ground to their closest competition and most bitter rival. They rely heavily on workstation-level performance to remain competitive in the creative field.

PPC 970 gives them that... and retains compatibility with all of the applications that are in the process of being ported, or have been ported already. That includes Photoshop, Office X, Illustrator, Flash MX, Director MX, etc. etc. etc.

Even Quark 6, coming soon...

Apple bails on PowerPC... everyone has to start the porting deal "ALLLLLL" over again. Carbon becomes damn near irrelavent at this early a stage, and a ton of companies will almost immediately bail on Apple and not develop for them again. Soooo... Apple has to ween the developers off Carbon eventually by fusing the API's into Cocoa and slowly removing them one by one, much like Microsoft did with the WIN_16 API's over time.

You really want to wait 6 years for Quark? LoL

I know most graphic designers don't.

Like I've said on here before... it's not "IF" we'll be on PPC 970...

It's When...

 

[maxvamp]

Me thinks??

Hmmm.... You knock off XP Home for not supporting dual CPU, yet you don't position the 970 against dual P4 (Xeon, of course).

Methinks a dual 3.06GHz Xeon with Hyperthreading probably wouldn't be "eaten alive"!

A dual Xeon would not compare in price. Will XP home ( who has no concept of more than one processor )be able to handle a chip (system) that says it actually has two processor?

I doubt with HyperThreading, that a single P4 3.2 GHz would be able to compete with the dual 1.6 PM. If you want to compare a dual Xeon with HT, build a quad 970 @ 1.6 and compare. That would be a fair comparison.

Max

 

[rog]

When the G4 733 came out after 18 months stuck at 500 MHz. They wen't from duals to a single processor, even on the high end. It wouldn't surprise me at all not even to ship a dually at first, or only on the top model at an outrageous price to drive buyers to the most expensive model.

Apple has made it clear they aren't interested in being like the Wintel world and go for low prices. They are perfectly happy in their own little sub 3% market share world.

They are never going to get to even 5% market share unless they slash prices. Why do they currently have the lowly Radeon 9000 on their top tower as the standard card? Why is there a $300 premium for the outdated GF4 Ti?

when the 970 comes out, they aren't going to suddenly be kind to everyone and drop prices.

I'll be shocked to see a DP 970 under 3 grand.

 

[rjstanford]

Re: 64 bit, hmmmmmm...

Originally posted by copperpipe
Well, most people here have been saying that 64 bits doesn't really make any difference except in encryption, which is to say that the difference they will make is miniscule at best. Now the people saying this definitely know more than I do, because I can't even understand what they are talking about in some cases. But, I just don't understand why Intel, AMD, and IBM are all investing billions of dollars in the race for 64 bit? It doesn't make sense to me, for such a tiny gain. Maybe we should go back to 16 bit? or 8? did they make a big difference? The POWER 4 uses 64, and it descimates every processor out there, but I suppose that it could be 32 and would descimate every processor, except when it's decoding an encryption. I dunno, maybe everyone here is right about 64 bit not making any noticable difference, but I'm having a hard time believing it.

One thing to keep in mind ... the bit-length refers to nothing more or less than the size of number that the processor can "think about" easily. Processors can handle larger numbers, but at a surprisingly high cost for their size (ie: an 8 bit processor can do 64 bit math, but its very expensive). Combine this with the fact that some operations (say, checking memory) need to be very fast indeed, and you'll see where I'm heading.

Those number sizes?

2^08 == 256
2^16 == 65,536 (or 64Kb)
2^32 == 4,294,967,296 (or 4Gb)
2^64 == 18,446,744,073,709,551,616 (or around 17 million Tb)

The main reason that most 32-bit processors are faster than the 16-bit processors they replaced is that they were made later. In other words, they were faster because they were bigger and better, not just because of the word length.

One of the main reasons that computers jumped to 32-bits so eagerly though was for that fast memory access. 32-bits allows the processor to address a memory location with one piece of information. 16-bits required at least two -- traditionally, one to refer to a "page" of memory, and the other to position within that page.

Now, there isn't much that requires more than 2^32 indexes in the personal computer realm. Today, that's pretty much limited to disk filesystems (where the physical disk access is unbelivably slower than the extra computing time that's needed to locate something). Tomorrow? Who knows what people will come up with. Memory sizes above 4GB are a good start, although intel, et al, are using 36bit memory keys in their newer processors to help with that very problem.

Now, in the enterprise realm, things are very different. Its easy to want to track more than 4 billion different business events, for example. Or if you're doing medical modeling. Or serious 3D raytracing. And on, and on. This is why the enterprise community is excited about true 64-bit processors, with the OSs to take advantage of them (such as AIX 5, Linux, etc).

Also, I hope that this helps to explain why the move to 64-bits isn't as big a deal as the move to 32 was.

 

[AidenShaw]

Re: Me thinks??

Originally posted by maxvamp
A dual Xeon would not compare in price.

How do you know what a dual or quad 970 would cost?

A dual 3.06GHz 1GB Xeon (Dell PWS450) is $3426, dual 2.8GHz only $2976.

Will XP home ( who has no concept of more than one processor )be able to handle a chip (system) that says it actually has two processor?

Who cares? XP Pro is about a $60 option even on low end systems, and the $60 is lost in the noise of a loaded dual or even a high end single. XP Home and XP Pro are identical, except for the restrictions placed on Home. Home vs. Pro is a silly non-issue, unless you're making your decision based on a $60 price difference.

I doubt with HyperThreading, that a single P4 3.2 GHz would be able to compete with the dual 1.6 PM. If you want to compare a dual Xeon with HT, build a quad 970 @ 1.6 and compare. That would be a fair comparison.

No, I'll build a quad Xeon with HT and big caches, and we'll check on the "eaten alive".

And, incidentally, I *did* just buy a quad Xeon.... (x440 2GHz/2MB). Where's your quad 970 ???????????

 

[mcs37]

I am crossing my fingers for Rev B of 12" PowerBook to be a 1.4 GHz SP PPC 970.

 

[IVIIVI4ck3y27]

Re: Re: Re: 64 bit, hmmmmmm...

Originally posted by ktlx
I have been thinking about this and while I generally agree with you, I see one situation where IBM might be interested in SIMD.

IBM is going to put PPC 970s in blade servers. One of the things blade servers do well is handle Web servers. Currently high volume transaction sites use add-on hardware for HTTP/S acceleration. I could see IBM pursuing using the SIMD unit as a cheaper way to provide the same functionality.

Of course it would still not perform as well as the dedicated hardware accelerators, but I wonder how many customers might say good enough when comparing a Linux Web server running on a PPC 970 blade to a Linux Web server running on an Intel Pentium 4 blade with an HTTP/S hardware accelerator given the cost savings.

It's possible, if IBM wanted to go through that trouble. But for IBM, I'm sure they probably have some of the hardware accelerators lying around. As much as they develop stuff... it'd likely be better to compete with the Pentium 4 blade with the hardware accelerator. After all... IBM machines in this segment tend to be overpriced... if it can't perform faster for the $, it's not going to be a value, and damn straight IBM's not going to make sales. Their services, which are a huuuuuuuuge part of Enterprise (and where the real profits at) are good... but you can almost hire your own IT staff to slap a custom blade server together and add their own version of Linux, and have a permanent on-site staff member for what it'd cost for the IBM machine and support and services for a year. So IBM would most likely offer the same hardware accelerator... and I doubt they'd go through extensive trouble to support SIMD. They might... but I seriously doubt it, especially *NOT* in AIX. Linux maybe... but that's still a big ?

 

[mcs37]

Originally posted by mathiasr
The PowerPC 970 is not supposed to have power saving features, you'd rather not expect to see it in a PowerBook.

A 1.4 GHz PPC 970 only draws like 23 W, I thought? That's not so bad!

 

[mathiasr]

Re: 64 bit, hmmmmmm...

Originally posted by ktlx
Also, IBM's POWER4/4+ does not decimate every other processor out there. For 32-bit integer math, it is beaten soundly by the Pentium 4 which is the reigning king for 32-bit integer math. For floating point, it is beaten soundly by the Itanium and probably by the Alpha.

I would not say that the POWER4+ is an under dog, here are some SPECfp2000 figures:

SPECfp2000 (base/peak)
1124/1482 : Alpha 21364@1150 MHz HP AlphaServer GS1280 7/1150
1391/1404 : Intel Itanium2@1000 MHz Bull NovaScale 4040
1202/1266 : POWER4@1300 MHz IBM pSeries 690 Turbo
1004/1228 : SPARC64 V@1350 MHz Fujitsu PRIMEPOWER900
1122/1219 : AMD Opteron144@1800 MHz Einux A4800
1097/1158 : POWER4+@1450 MHz IBM pSeries 630 Model 6C4
945/1106 : UltraSPARC IIICu@1200 MHz Sun Netra 20
1092/1103 : Intel Pentium4@3066 MHz Dell Precision WorkStation 350
1053/1063 : Intel Xeon@3066 MHz Dell Precision WorkStation 650
776/ 869 : AMD AthlonXP3000+@2167 MHz ASUS A7N8X Deluxe
602/ 650 : AMD AthlonMP2600+@2133 MHz MSI K7D Master
499/ 529 : MIPS R14000@600 MHz SGI Origin 3200
437/ 456 : Intel PentiumIII@1400 MHz Dell PowerEdge 1500SC

IBM has announced the POWER4+ running at 1.5 and 1.7 GHz not covered here, and the Itanium is expected at 1.5 GHz anytime soon; notice that an IBM p690 server build around 32 POWER4+ 1.7 GHz processors has outperformed a 64 Itanium 1.5 GHz HP Superdome server in the TPC-C benchmark:
http://www.theregister.co.uk/content/61/30642.html

 

[maxvamp]

Your right!!

--------------------------------------------------------------------------------
Originally posted by maxvamp
A dual Xeon would not compare in price.
--------------------------------------------------------------------------------



How do you know what a dual or quad 970 would cost?

A dual 3.06GHz 1GB Xeon (Dell PWS450) is $3426, dual 2.8GHz only $2976.

I guess $3426 == $2200 . I know I probably have an extra $1200 in my couch cushions alone.

This original thread tried to start out comparing $2200 machines.

Based on Apple's general price steps, there will pobably be a 1.6 in the low to mid $2k range.

If you want to derail this train try making a valid argument with the dual 970 1.8 or 2.0 against the Xeon 3.06. I believe that would be closer to a valid comparison, that is if you are into fair comparisons. IBM has already published some "Conservative" numbers on this. BTW: the dual and Quad PPC 970s are coming. They will be labled IBM for sure. Maybe Mac ( anyone think there will be a Quad Xserve? ).

In all honesty, you sound like you are just trying to start a 'Why my PC is better' fight. I think you would have a better chance trying this crap over at http://www.amdzone.com . There are many reasons why people prefer Macs.

Speed of processor != Productivity.


Good Day,

Max

 

[copperpipe]

thanks for the info

Thanks for more info on the 64 bit discussion, Ktlx and IVIIVI4ck3y27 especially. Your well thought out replies have helped me to understand things a little bit more.

 

[macdong]

Originally posted by greg6028
I heard of talk about a dual processor in a PowerBook.
Will this happen now with the 970?

Ultimate PowerBook with Dual 970
Fastest PowerBook with 970
Faster PowerBook with Dual G4
Fast PowerBook G4

I have a feeling this will happen in winter of 2003. The desktops will be first. Are not the desktop to oldest in the Apple line now?

You are way over your head, dude.
I don't want a steaming fried pan on my laps.

 

[AidenShaw]

Originally posted by maxvamp
In all honesty, you sound like you are just trying to start a 'Why my PC is better' fight.

No, just trying to point out the inconsistency in the comparison of single and dual CPUs, and the information which shows that while the POWER4 (and estimates for the PPC970) are certainly very good, being "eaten alive" by a PPC970 is a bit of a stretch.

Look at it another way - if you listen to me (and other calm voices), you'll most likely be pleased when the PPC970 shows up. If you listen to the wild statements, you'll be quite disappointed when it arrives....

 

[mathiasr]

Re: Re: 64 bit, hmmmmmm...

Originally posted by IVIIVI4ck3y27
Why is it so apparent that IBM won't use SIMD?

Simple. AltiVec is a vector-processing SIMD chipset. It's designed for handling more graphical tasks. It's exactly why Apple was so adamant on using it in the G4, and exactly why IBM and Motorola split in their roadmaps, with IBM continuing to evolve the 32-bit & 64-bit versions of the G3 rather than go Motorola's route in developing the G4 with a SIMD unit.

[snip]

IBM doesn't use PowerPC's in workstations... because the only OS's they use on PowerPC are AIX and Linux, both more geared towards server than workstation because neither is optimized with Assembly calls, there's no vector processing calls, and there's not likely to be any anytime soon. Why? Linux and Unix are designed to be portable. Any application or part of the system is written to be ported easily through some mild alterrations and a recompile. The minute you specialize the code for MMX, or 3DNow, or AltiVec... is the minute you have a lot more porting work ahead of you to gut out, remove, and reoptimize said items.

IBM 64 bits CPUs are not based on the G3, they are related to the RS64 or POWER3 chips:
http://www.research.ibm.com/journal/rd/461/tendler.html

IBM offers some assembly hand tuned libs like MASS (Mathematical Acceleration Subsystem):
http://www.rs6000.ibm.com/resource/technology/MASS
and give you some hints to speedup your apps:
http://www.redbooks.ibm.com/pubs/pdfs/redbooks/sg247041.pdf

 

[mcs37]

Re: Re: Wild speculation

Originally posted by seamuskrat
64 bit OS will NOT make it run FASTER. A misconception at work here. A 64 bit Os will allow for more memory and will allow for some efficiency with large memoroy intesnsive files. In SOME rare cases, it would actually slow a process down if it was re-written to be 64 bit.

True but enhancements will come once enough software takes advantage of the 64 bits. Number-crunching intense applications like Final Cut Pro will most definitely be much faster with that wider pipeline.

 

[ktlx]

Re: Re: 64 bit, hmmmmmm...

Originally posted by mathiasr
I would not say that the POWER4+ is an under dog, here are some SPECfp2000 figures:

A much more expensive processor loosing to a cheaper processor by almost 15% is getting soundly beaten in my book. I can buy more Alphas or Itaniums than POWER4/4+ for the same amount of money and get a faster processor for floating point.

IBM has announced the POWER4+ running at 1.5 and 1.7 GHz not covered here, and the Itanium is expected at 1.5 GHz anytime soon; notice that an IBM p690 server build around 32 POWER4+ 1.7 GHz processors has outperformed a 64 Itanium 1.5 GHz HP Superdome server in the TPC-C benchmark:
http://www.theregister.co.uk/content/61/30642.html

That is true but TPC-C has absolutely nothing to do with the SPECfp marks of a processor. It has much more to do with L3 cache sizes, I/O bandwidth and all of those things that I mentioned in the paragraph you left out of your quote.

As I said, on a per processor basis, there isn't anything out there that competes with the POWER4/4+ line for databases (except maybe IBM mainframes ). The only negative is the high cost per processor on these servers.

The link points out the problem I mentioned. The 64 processor Itanium 2 machine was $600K cheaper than the 32 processor IBM. IBM had to pay 7% more money to get 3% more performance.

 

[mcs37]

Originally posted by IVIIVI4ck3y27

The odds of dual processor Powerbooks isn't that good... for one that's an extra processor chewing up juice (desktops it makes sense, laptops... unless they can get dual G4's or 9xx's running at 7w like the current iBook, you're talking even more heat than a current laptop G4 which is already hot to a lot of people's chagrin), generating more heat, and eating up more battery life. Yet if it does happen, it'll likely be some hybridization of G3/G4... and that's just not logical at this stage. [/B]

I disagree here. First, it would be worth's Apple's time to attempt to release the PowerPC 970 into their PowerBook line as close the PowerMac line as possible since many users would simply wait 6 months or longer until they were available, resulting in lower sales for Apple until they fill the gap.

Second, dual processors in a PowerBook is a fantastic idea. But you are right about power consumption. Solution? The OS shuts down the second processor when the computer is running on battery only. Should be a pretty straightforward feature that would give PB users all the power of a desktop when plugged in and still a powerful box when on the plane.

 

[type_r503]

64 bits will increase performance after software makers learn how to use the extra bits. As an example, you can fetch 2 32 bit ints in a single clock cycle or 4 bytes, and you can fill an Altivec register in half the time. Take a look at first generation PS2 games compared to Enter the Matrix. If the tools are there they will get used.

 

[jettredmont]

Re: 64 bit, hmmmmmm...

Originally posted by copperpipe
But, I just don't understand why Intel, AMD, and IBM are all investing billions of dollars in the race for 64 bit?

First, 64-bit computing will be quite useful, and will drive an entire wave of new computing technologies (judging from the past). What you are hearing people say here is that there will be little to no immediate performance jump due to having a 64-bit processor. The things you do today work just fine in 32 bits; the only apps which would benefit from 64 bit CPU are those using 64-bit-manipulation-using-32-bit-registers kludges today.

That having been said, there are existing applications for 64-bits, but they are centered around the area where 64-bit machines have existed for a while: on the server. These include databases, file/web servers, etc.

It doesn't make sense to me, for such a tiny gain. Maybe we should go back to 16 bit? or 8? did they make a big difference? The POWER 4 uses 64, and it descimates every processor out there, but I suppose that it could be 32 and would descimate every processor, except when it's decoding an encryption. I dunno, maybe everyone here is right about 64 bit not making any noticable difference, but I'm having a hard time believing it.

Again, the Power 4 decimates other 64-bit processors, not because it is 64-bit (so is its competition) but because it is a damned fine implementation on an excellent architecture.

Note that the 8- and 16-bit comparisons are not really applicable. In both of those transitions, mainstream software existed which used kludges to manipulate higher-width integers (16 and 32 bit ints, respectively), and the limits of memory addressing were a very real and pressing concern (personally, I know no one with 2GB of memory on their desktop; everyone I knew had their 286's completely maxed out with one (then very expensive) MB of RAM and still were constantly fiddling to conserve just a little bit more memory).

This is not exactly the case today. That is not to say that soon we won't be hitting the 32-bit limitations, but we haven't really hit it en masse yet like we had with 8- and 16-bit limitations (4-bit limitation being a bit ahead of my time, except in terms of video colors ... )

As for why Intel and AMD are falling over themselves trying to be first out with a 64-bit desktop computer: well, first off Intel surrendered and took its ball (Deerfield) home; secondly the main driving force there is marketing (64-bits is obviously better) and bragging rights. Do you remember the "race" to 1GHz? Did that mean diddly squat? Of course not. Except it meant that suddenly 800 and 866MHz machines were dirt cheap. "1GHz" was a bragging rights benchmark, and so in a large part is "64-bit desktop".

That having been said, there are real and lasting benefits to having 64-bit processors on the desktop. It's just not an overnight or even one-compile-away type of benefit. It enables new concepts and new paradigms to blossom and create the 64-bit "killer app".

 

[freundt]

As for linup... they will keep it as simple as it is now.

1.4 SP Power Mac
1.6 DP Power Mac
1.8 DP Power Mac

However, I'm not so sure about the powerbook. I think we will have a 970 based powerbook in the near future. I mean, it's ok to have a 1 generation behind processor(or multiple if you look at other chipsets) in a consumer based laptop - G3 ibook -, but it is not ok to have a 1 generation behind processor in the pro laptop area - especially when that processor is behind all other chipsets in regards to speed.

I still think the 15" is the flagship laptop, and they have been holding out for a major upgrade avaiable for it before releasing a newer version. (I think the Lapzilla is a flash in the pan product that will probably not be carried on throughout future iterations - it's too impratical, and I believe it was released because Apple was/is stuck with slower processors and they needed a novel product to boost sells)

Anyways, that's my flawed thoughts on the whole mess...

 

[maxvamp]

calm

Look at it another way - if you listen to me (and other calm voices), you'll most likely be pleased when the PPC970 shows up. If you listen to the wild statements, you'll be quite disappointed when it arrives....

I seriously doubt taking a comparison of two machines, and heavily skewing it to include a machine that cost a third more to prove that the IBM based machine is not as good as it most likely will be ( IBM has an excellent track record in this area, even based on the IBM numbers provided ) is acting in a calm way. If I listen to you, I will be buying those dells that you found nessesary to try to compare as a dollor for dollar machine.

As Judge Judy would say : Don't pee on my leg and tell me it is raining. Your attack on the comparison did not come across as calm.

Again, If you and a few others want to compare machines on a dollar level, lets talk about current and projected machines that fit that range. If you want to compare fastest, then acknowledge that there is a difference in price.

A $2200 dual 970 **WILL BEAT** a single $2200 P4, especially from Dell. Dell's dual lines, as well as server lines are built nicely, but the lower the cost of the Dell, the more the preformance drops, even with the same processor. They tend to cut back on leading technolgy that makes the machines faster, as performance costs money.

If you would like to discuss this without trying to directly compare more expensive Intels to less expensive Apples, we can, but for now, I am going back to my AMD 3000+ and my PM 733. Oddly enough I seem to get about the same amount, but different, work done on each.

Must not be the clock speed.....

Max

 

[hvfsl]

A 64bit chip on a 64bit OS is quite a bit faster than a 64bit chip on a 32bit OS. What people are forgeting here is that every clock 64bits of data are processed instead of 32bits. So twice as much data. Although in the real world there will not be a 2x performance boost, much like a dual G4 is not twice the speed of a single G4. IF you look at the new figures for the AMD 64bit on www.futuremark.com you will see a 1.4Ghz AMD is 70% faster than a 2.26Ghz P4.

 

[AidenShaw]

you can already do much better...

Originally posted by type_r503
As an example, you can fetch 2 32 bit ints in a single clock cycle or 4 bytes, and you can fill an Altivec register in half the time.

An AltiVec load or store instruction already does 128 bits per operation. (Same for SSE2 on Pentium)

The internal datapaths are 128-bit to 256-bits wide already, so the increase in width from 32-bit registers to 64-bit registers is helpful, but not a big improvement. (Besides, getting the data across the system memory bus is still slower by far than any internal data movement - even if multiple smaller internal moves are needed.)

As many before have said, 64-bits will have 2 main advantages:

- algorithms which need intensive use 64-bit integer arithmetic will be faster

- programs which need more that 2GB to 4GB of actual RAM per program will be able to run

And, by the way, 64-bit filesystems (which are basically any filesystems with UNIX semantics that can support files larger than 2GB) are always 64-bit filesystems - they don't shift modes for larger files. The system has to refer to files with 64-bit arithmetic even if they are only a few bytes in length!

This does not count as "intensive 64-bit use", though, since a filesystem has to do a lot of work beyond calculating file offsets. One would probably be unable to measure any improvement in filesystem performance due to 64-bit integers (but you might find a huge improvement with filesystem caches that are 64-bits (i.e. caches larger than 2GB)).

 

[maxvamp]

must have missed it

Aiden: In the file system stuff, For the most part you are right. I would ask what is your take on any speed impact when the processor has to manage the upper and lower set of bits of a given file. I know that the Win32 API takes care of this rather transparently. Wouldn't you see some performance on disk intensive applications?

Just curious,

Max

 

[arn]

For what it's worth, PowerJack (MacWhispers) does not believe this to be the case:

http://www.spymac.com/forums/showthread.php?s=&threadid=25573&perpage=15&pagenumber=3

Foxconn is not manufacturing either the new 970 PowerBook or the 970 PowerMac. There have actually been zero PPC 970 chips delivered to Foxconn.

Shame on LoopRumors.

I know some people have issue with MacWhispers' reliability... but LoopRumors isn't known to have reliable sources either.

The original article (LoopRumors) was posted on MacRumors as it corresponds with other information floating around (namely the MacBid info)

arn