Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Rod Rod said:
PCI-e's main benefits seem to be for games and probably Core Image / Core Video, but since we're 2 1/2 weeks away from seeing whether CI/CV is choked for throughput it's probably better to reserve judgment.

There are more than performance based reasons to upgrade to PCI-E and DDR2 667 RAM, for example. Namely, any modern graphics card is based on PCI-E. The cost to produce an AGP 8x version creates additional cost. In the PC world, it's about a 33% premium. That's crazy. Oh, wait-- that's what we, as Mac users, pay anyway. My mistake. :) Perhaps this means we will pay an even higher premium unless we have bus parity with the PC world (pun intended.)

The real world performance of PCI-E for graphics over AGP 8x is currently pretty much the same... but that's not why we need it right now. We need it so we can get Radeon x850's without paying $700.

I think the bus speed is also an important factor for dual core chip designs, so I have to agree with many other posters that HT 2.0 will be a key feature. On top of that, I think DDR2 (667 hopefully) will be a key factor. Maybe they go with 533 (which doesn't have much of an advantage over DDR 400), but at least we're getting our ducks in a row for current technology.

More than anything the PowerMac needs to be a cutting edge beast. We need:

Dual 3 GHz (dual core?)
DDR2 667 Ram
HyperTransport 2.0
PCI-E Graphics
3 to 4 PCI-X Slots
3 to 4 Internal Drive Bays
2 Optical or externally accessible slots
 
Metropolis Man said:
The evening of April 15th? Is that likely given the convention starts on Monday?

Anecdotally, Apple typically does not seem to like making announcements on weekends. Also, announcing on that Friday before the convention starts would give some time for buzz to spin through the roof, as well as possibly fit better logistically if hardware is going to be on display the morning of the 18th. Certainly not certain, but certainly a possibility. :p
 
eyeluvmyimac said:
From Apple.com:

"4 250GB hard drives, the largest currently available in the Power Mac G5, will also store a terabyte of data."

However, currently only 2 250GB drives can come in the Power Mac G5....Is this just a typo? A sign of things to come? Referring to after-market modifications? You decide.

I wouldn't look too far into that. You can buy 400 Gig drives right now, and 500 gig drives are out any day now. So, in theory you could have 1.6 TB (or 60% more) internally today or 2 TB (100% more) very soon.

But I also think Apple is pushing the xServe and xSan as a storage solution over heavy local storage. It makes sense, too, because to me a 400 GB internal RAID provides local backup security and a SAN provides data sharing that doesn't tax the bandwidth and drive heads of my local machine.
 
hmmm

While some have predicted Isnt the most likely pricepoint for Apples forthcoming flagship PM going to much higher than it is currently, wouldnt it make more sense for it to be closer to current prices? And if so, would that preclude the possibility of a Daul-dual machine? Would Apple really introduce a screaming new top-of-the-line at a 5 grand pricepoint?
 
Sunrunner said:
Here is another question for everyone to ponder. What is the most likely pricepoint for Apples forthcoming flagship PM?

The same price, or maybe slightly less. I don't recollect a time when Apple have upgraded their line's and made it more expensive.
I doubt there will be dual dual-core's at NAB, WWDC imo because it is a major announcement.
 
Frobozz said:
There are more than performance based reasons to upgrade to PCI-E and DDR2 667 RAM, for example. Namely, any modern graphics card is based on PCI-E. The cost to produce an AGP 8x version creates additional cost. In the PC world, it's about a 33% premium. That's crazy. Oh, wait-- that's what we, as Mac users, pay anyway. My mistake. :) Perhaps this means we will pay an even higher premium unless we have bus parity with the PC world (pun intended.)

The real world performance of PCI-E for graphics over AGP 8x is currently pretty much the same... but that's not why we need it right now. We need it so we can get Radeon x850's without paying $700.

I think the bus speed is also an important factor for dual core chip designs, so I have to agree with many other posters that HT 2.0 will be a key feature. On top of that, I think DDR2 (667 hopefully) will be a key factor. Maybe they go with 533 (which doesn't have much of an advantage over DDR 400), but at least we're getting our ducks in a row for current technology.

More than anything the PowerMac needs to be a cutting edge beast. We need:

Dual 3 GHz (dual core?)
DDR2 667 Ram
HyperTransport 2.0
PCI-E Graphics
3 to 4 PCI-X Slots
3 to 4 Internal Drive Bays
2 Optical or externally accessible slots
Thank you for that great explanation. Hopefully (now with the demise of ADC and rise of DVI) the day won't be far off where Mac video cards won't cost way more than their PC counterparts.

Sunrunner said:
Anecdotally, Apple typically does not seem to like making announcements on weekends. Also, announcing on that Friday before the convention starts would give some time for buzz to spin through the roof, as well as possibly fit better logistically if hardware is going to be on display the morning of the 18th. Certainly not certain, but certainly a possibility. :p
Apple has introduced a new Final Cut Pro on each Sunday morning of each NAB of each of the last 5 years (1.0, 2.0, 3.0, 4.0 and 4.5). This coming Sunday is ripe for FCP 5.

As for the "buzz" you're talking about, well not everyone out there is as impressed with Apple as we are. Besides, the show starts on the 16th, not the 18th. Apple's exhibit space and their demos will likely be packed (as they were last year) but Apple excitement isn't nearly universal.
 
what about HTX ??

Frobozz said:
Dual 3 GHz (dual core?)
DDR2 667 Ram
HyperTransport 2.0
PCI-E Graphics
3 to 4 PCI-X Slots
3 to 4 Internal Drive Bays
2 Optical or externally accessible slots

Replace some of the PCI-X (*old* tech) with

- x8 PCI-e I/O slots
- HTX slots

While HTX would be marginally useful for the PM - the cluster guys would love it. Since Apple would probably share the chipset with the XServe, HTX would be a logical addition. (http://www.pathscale.com/infinipath.html)
 
Rod Rod said:
Apple has introduced a new Final Cut Pro on each Sunday morning of each NAB of each of the last 5 years (1.0, 2.0, 3.0, 4.0 and 4.5). This coming Sunday is ripe for FCP 5.

As for the "buzz" you're talking about, well not everyone out there is as impressed with Apple as we are. Besides, the show starts on the 16th, not the 18th. Apple's exhibit space and their demos will likely be packed (as they were last year) but Apple excitement isn't nearly universal.

Final Cut Pro, yes... though thats not really on the same level as multiple new product announcements. Sunday announcements are certainly a possibility though. While the show starts on the 16th, dont the exhibition spaces open on the 18th?
 
Bring on the Mac Mini Media Center! I'm sick of hearing how great XP Media Center is. Blech
 
Rod Rod said:
...Besides, the show starts on the 16th, not the 18th. Apple's exhibit space and their demos will likely be packed (as they were last year) but Apple excitement isn't nearly universal.

But according to the NAB website exhibitor booths don't open until the 18th, so we'll all find out in a few days if Apple's new hardware and software debuts then or earlier.
 
Sunrunner said:
Final Cut Pro, yes... though thats not really on the same level as multiple new product announcements. Sunday announcements are certainly a possibility though. While the show starts on the 16th, dont the exhibition spaces open on the 18th?
FCP 5 will be a new product. Last year they announced multiple new products (in addition to FCP 4.5, Motion and DVDSP 3 came out). I suppose you mean hardware products.

NevadaJack said:
The convention starts on the 16th with the Exhibits opening on the 18th.
Thanks, as Sunrunner mentioned. I registered but since I didn't get a plane ticket to Las Vegas I guess I didn't look into the exhibition space schedule as I probably should have anyway.

Metropolis Man said:
But according to the NAB website exhibitor booths don't open until the 18th, so we'll all find out in a few days if Apple's new hardware and software debuts then or earlier.
Thanks, as NevadaJack pointed out.

Sunrunner said:
There it is, thats when those products will be the announced
I'm pretty sure I posted that on page 2 or 3 of this thread . . . Yup, it was post #33, on of this thread.
 
fpnc said:
~loserman~, I think you should read more carefully, I was referring to the Athlon dual-core (and I clearly said so). I've said nothing about the AMD Opteron and I've never mentioned the Opteron until this very message. For that matter, I've never mentioned the Xeon either. As I noted earlier (to daveL), I'm certain that everyone else here is talking about the 970MP and the competition for the 970MP is the dual-core Pentium and dual-core Athlon. It is not some high-end server based upon the Opteron, POWER, or Sparc processor line.

This is where we differ. The 970MP does compete against the Xeon/Dual Core Xeon. It also competes against the Opteron and Dual Core Opteron.
The reason is the PPC970/PPC970FX have always been capable as being used as dual processor configurations and Apple sells them as such. The Athlon64 and Pentium 4 have not been. When comparing platforms the PowerMac has always competed against Workstation class machines in the X86 line. Even Apple makes the same processor comparisons on their own site. I never compared the PPC970 and its derivatives against "exotic server" CPUs such as the Sparc, Power 4/5 or Itanium. Even though those comparisons can still be made the price points are too far out.
The Athlon64 FX IS an Opteron regardless of how it is being sold. The internals of the CPU are almost identical(except for its use of non ecc memory and a different pin out). The real funny part is the Single Processor version of the Opteron actually sells for less than the Athlon64 FX. The Dual Processor Opterons sell for the same price as their Athlon 64 FX cousins. Unless something that I am unaware of changes the Pentium 4 and Athlon 64(non FX version) dual-cores will still only be used in Single processor motherboards.

Back on Topic:
I am revising my estimate of when we will see dual core PowerMacs to earliest 4th quarter 2005 but more likely 1st quarter 2006 due to a conversation with IBM and their availability of Dual Core CPUs for their JS20 Blade servers. IBM claims they will release dual core JS20 blades early 1st quarter 2006.
They also claim they will upgrade the clock hertz to 2.5 Ghz on their JS20 blades in August/September Time Frame.
This bodes well for a 2.8 to 3Ghz PowerMac soon since IBM sells the faster CPUs to Apple. Traditionally IBM has provided Apple with CPUs clocked 300mhz faster than the ones they use themselves. They have also traditionally done their updates about 2 to 3 months later than when Apple does.
 
Fredstar said:
The same price, or maybe slightly less. I don't recollect a time when Apple have upgraded their line's and made it more expensive.
I doubt there will be dual dual-core's at NAB, WWDC imo because it is a major announcement.

I agree with your last sentence. However, I disagree with the same price or slightly less on the new upgrades. This would only happen if all Apple introduces is a 2.8 system without much change inside. Any new and much faster technology will increase the price of the PM. I can see Apple charging $3500 for a dual core system running at 2.5 ghz.

I have changed my mind and no longer believe that Apple will introduce PM any time soon. There have been no breakthroughs that I have seen in the cooling issues and faster more reliable chips. The iPod, the iMac, and Tiger will keep Apple afloat without new Power Macs for a few more months.
 
AidenShaw said:
Claim: G3 will rule - RISC is king, CISC is dead
Truth: After a strong early start by RISC, CISC won the race

That's entirely wrong. CISC is dead. Intel hasn't made a CISC chip for PC's since the Pentium. Penitum Pro was the first RISC CPU from Intel but it trickled down to the Penium II, III, V, Xeon, etc.

Don't confuse instruction sets and register counts with chip design techniques.
 
rockthecasbah said:
If the Power Macs have similar types of breakdown issues, this could really provide a lethal blow to Apple's craftsmenship in the eyes of the customer. :eek:

If it becomes a problem perhaps Apple will have air-cooled modules based on SSOI available in time to replace them. In this industry it's better to ship and recall than to stagnate.

I haven't fondled one of these modules myself - are they not sealed inside the radiator? I would assume in the case of a leak the G5 module case could contain the liquid and the firmware could detect the high-temp CPU and shut it down. Even a semi-sealed design would put any potential leak points at the top and let gravity handle the overflow vector. Then the OS says, "One of your CPU's has failed. Your computer will continue to run, albeit more slowly. Please call 1-800-APLCARE to schedule a replacement. Please do not turn your computer upside down until it has been serviced."

I mean, everything made by man does fail eventually, and if you make a million of anything something invariably gets by as a dud, but if you're playing in the watercooled world it's not unreasonable to have precautions as outlined above to keep the customer informed.

Being made in Mexico shouldn't have a bearing on this - their QA process is what's important.
 
And another thing that just popped in to my mind is that as I remember correctly Steve Jobs proclaimed the 2005 to be the year of HD video, and the president of Sony wasnt there on stage because he had nothing better to do and considering they didnt announce any collabortive product back then but it was rather "just" a speech I am wondering if any or all of it has to do with Powermac development (CELL processor/coprocessor) and whose to say Apple hasnt been on it secretly from the beginning cuz of their nature of total secrecy... Jobs already had one fiasco with the 3ghz promise and I doubt that he would be crazy enough to pop in some more words without backing it really up in the test labs...the year of hd video, what does it mean, to me it definately means more then just a release of a mac mini or some kind of media center device, first thing that pops in to my mind is seriously powerful video editing equipment with software and we all know that even the most powerful personal computers (dual 2.5ghz macs included), when it comes down to decoding/encoding pretty much are slow compared with other computer tasks we are accustomed to... I dont think that January speech was there so Apple would release just speed bumps because it definately needs more then that... Of course I may be wrong but it all makes sense to me and I also know Apple wants to redeem itself for the 3ghz fiasco, I know that some things really are not in their hands and pretty much they depend on other companies but there is too much competition going around which also helps Apple to deal with their partners...
 
ClimbingTheLog said:
That's entirely wrong. CISC is dead. Intel hasn't made a CISC chip for PC's since the Pentium. Penitum Pro was the first RISC CPU from Intel but it trickled down to the Penium II, III, V, Xeon, etc.

Don't confuse instruction sets and register counts with chip design techniques.

Indeed.. look at the IBM AS/400 line which used to be CISC and is now well and truely RISC.

Doesn't the "IT" Industry just love its acroynms.

>> "IT" << see thats the problem when you work in the industry.
 
ClimbingTheLog said:
That's entirely wrong. CISC is dead. Intel hasn't made a CISC chip for PC's since the Pentium. Penitum Pro was the first RISC CPU from Intel but it trickled down to the Penium II, III, V, Xeon, etc.

Don't confuse instruction sets and register counts with chip design techniques.

This is mostly true.

The Pentium Pro and its derivatives have taken the RISC based approach, even though it continues to use the x86-32 instruction set. These processors breaks up the fairly large, CISC based x86 instructions - into what is called micro-ops, tiny code fragments that the processor can digest faster (ala RISC).

AS to RISC
I would also add that RISC has also migrated back to be more like CISC.
In the last ten years almost every single RISC architecture has gradually restored the very same features that were removed during the RISC "purge" of the 1980s. Unaligned memory access, bit-wise addressing, media instructions, DSP instructions, variable-length instructions, code compression, multi-cycle functions, multiply-accumulate... etc., have all come back to RISC chips. It seems that RISC might be "reduced" in name only.
 
Fredstar said:
I doubt there will be dual dual-core's at NAB, WWDC imo because it is a major announcement.
It would be ridiculous to delay a product release by one or two months for the sake of WWDC or any other event. It makes much more sense to release new products when they're ready, no matter which trade show, conference or event is on the calendar.

As blitzkrieg79 said, even today's top of the line Power Mac G5s feel slow when working with high definition video. The sooner that faster Macs come out, the better.
 
AidenShaw said:
Replace some of the PCI-X (*old* tech) with

- x8 PCI-e I/O slots
- HTX slots

While HTX would be marginally useful for the PM - the cluster guys would love it. Since Apple would probably share the chipset with the XServe, HTX would be a logical addition. (http://www.pathscale.com/infinipath.html)

I haven't heard of HTX. Do tell? Does it offer advantages to PCI-X or PCI-E ? (I'm curious since I figure "faster" is obvious ...)
 
pgre said:
Indeed.. look at the IBM AS/400 line which used to be CISC and is now well and truely RISC.

Doesn't the "IT" Industry just love its acroynms.

>> "IT" << see thats the problem when you work in the industry.
Ahhh...


You haven't considered the latest architecture:

Tiny Instruction Set Computer-- TISC (pronounced "Tsk")

The only instructions are:

00 NOP
01x Load/Store
10x Add/Sub
11x Test/Branch

x represents a modifier bit that refines the basic op codes.

The advantage is that all ops can be reperesented on one hand (with the modifier selected by the thumb)

That's all the instructions any computer needs
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.