Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

More on PowerMacs at NAB?

~loserman~ said:
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.
Click to expand...

As a bit of side commentary, I think it's common for people to loose sight of the applications running on the chip and how the OS mitigates that. Correct me if I'm wrong, but server oriented chips and operating systems won't necessarily run Photoshop better. I heard a good explanation of this once: servers run many small tasks over and over. In the case of a web site with a fairly static data set, the applications that run fastest exploit the conditions they're dealing with, as do desktop processors. This is a pretty simple optimization principle.

So, even if an Itanium or Sparc were great server CPU's they wouldn't _necessarily_ be good as desktop user applications.

I think the HyperTransport 2.0 standard and the 970mp will come hand-in-hand to the PowerMac. Hopefully, this will be soon. :)
 
Sorry, but please tell me

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.
Click to expand...


Then please tell me what the "IS" stands for in the acronyms CISC and RISC ?? :eek:



BTW, I agree that we're in the "post-RISC" era - where the original definition of RISC has been blurred, and CISC ISAs are implemented on silicon with decidedly RISC characteristics.

My "Claim:/Truth:" statements were meant to discredit simple distinctions like CISC/RISC, not to promote them.
 
Rod Rod said:
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.
Click to expand...


It is totally ridiculous except for the fact that Apple doesn't ship the day they annouce a product. If they know they can ship in June, they can announce at any time before June. I don't think that Apple would delay announcing a product that is sitting in their warehouse. At least not for longer than a few days. ROI and all the other financial metrics would suffer enourmously if Apple kept a ton of inventory for longer than a few days.

I'd also like to say that all of you expecting Bluray should prepare to be disappointed.
 
HTX is a HyperTransport IO card

Frobozz said:
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 ...)
Click to expand...

HTX brings a HyperTransport "port" to the I/O card - avoiding the HyperTransport to PCI-X bridge. Faster, much lower latency.

Yahoo for "hypertransport htx" to find out more, or look at the link that I posted for some brief info. (http://www.pathscale.com/infinipath.html)

Opterons (with the direct HyperTransport controller on the CPU chip) get great improvements for latency on InfiniBand transfers used in HPC clusters.

Since the G5 doesn't have a direct HT controller on the CPU, the improvement might not be as great. But since the PCI-X controller is a HyperTransport tunnel, you'd avoid the latency of the HT -> PCI-X conversions.
 
Well I heard from the head apple custodian that the new PM will be dual 3.0ghz dual core + cell. Which, for cooling, requires being hooked up to your refrigerator. If something like this (sans fridge) actually came out which is slightly unlikely ;) I wonder how many people/companies would switch and how much people would be willing to pay for it.

Like some have pointed out though, apple NEEDS to pull out some major hardware. Editing HD is serious computation, especially when you get into compositing and quick and solid NonLinear Editing with a good amount of Real Time action going on. And if its the Year of HD, you'd think they'd release stuff early in the year...other wise itd be more like the Late Autumn of HD.
 
Beck446I'd also like to say that all of you expecting Bluray should prepare to be disappointed.[/QUOTE said:
I'm not expecting anything in particular for the next update Rev. C of the Power Mac G5. Don't want to be disappointed. Do expect that improvements will be made. Current noted problems should be foxed.
Click to expand...
 
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.
Click to expand...

That is what differentiators a RISC from CISC computer. All x86 CPU's by definition are CISC computers as they support the original x86 instruction set. Modern CPU's (especially AMD's) employ techniques borrowed from RISC computer (super pipelining for example) but they are still CISC computers.

A RISC CPU is faster than CISC cpu in that it is designed to execute fixed-length instructions very quickly and rely on the fact that most compilers rarely make use of long complex instructions when they compile code. Chip design is not what makes a CPU RISC or CISC it is the instruction set and length of instructions (fixed or variable) that determines it.
 
Rod Rod said:
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.
Click to expand...

I agree, it seems to make more sense to release now rather than later. If iMacs, iBooks, or even Powerbooks (don't hold your breath) are to be updated at WWDC, the Power Mac being released early such as at NAB would increase its publicity. Spacing them out decreasing the chance of one product overshadowing another. Also, this may be a perfect point to both show High Definition Video software and products working on a new Power Mac ;) .
 
Frobozz said:
As a bit of side commentary, I think it's common for people to loose sight of the applications running on the chip and how the OS mitigates that. Correct me if I'm wrong, but server oriented chips and operating systems won't necessarily run Photoshop better. I heard a good explanation of this once: servers run many small tasks over and over. In the case of a web site with a fairly static data set, the applications that run fastest exploit the conditions they're dealing with, as do desktop processors. This is a pretty simple optimization principle.

So, even if an Itanium or Sparc were great server CPU's they wouldn't _necessarily_ be good as desktop user applications.

I think the HyperTransport 2.0 standard and the 970mp will come hand-in-hand to the PowerMac. Hopefully, this will be soon. :)
Click to expand...


I don't disagree with your premise.
The only thing I would add is the lines have blurred between what is a server CPU and what is a desktop one. ONE REASON is that chips that used to be considered as "Server" chips had larger cache sizes. This has changed somewhat recently. Another reason is that "Server" chips almost always had the ability to be used in Multi-processor configurations whereas desktop cpus were always single processor only.

For example.
Many people compare Apple's PowerMac line to lets say a Dell or Compaq.
Typically the comparisons are made between dual processor PMs vs single processor X86 machines. These are not direct comparisons. A more appropriate comparison would be the iMac to the Dell/Compaq. Since these are both single processor machines. Those comparisons are usually not made though because the top Single processor x86 boxes handily out perform the iMac.
PowerMacs are Apple's equivalent of a WORKSTATION quality offering. For several years Workstation type machines have been dual processor and in x86 land you are talking about Xeons and Opterons. Even the price of these machines reflect this. For example the top PowerMac starts at around $3000 and so do dual Processor Opteron/Xeon workstations from HP/Compaq.
Yes Xeons and Opterons can be sold as single processor machines but almost never are. When they are sold as such they come on motherboards that allow you to add the second processor later. Also I agree with some that have posted that these cpus are traditionally thought of as entry level server processors but I think as I have shown they are also the ones that are sold in the 2 processor Workstation class market( the same market segment as the PowerMac)
 
~loserman~ said:
Also I agree with some that have posted that these cpus are traditionally thought of as entry level server processors but I think as I have shown they are also the ones that are sold in the 2 processor Workstation class market( the same market segment as the PowerMac)
Click to expand...

To emphasize your point, note that Xeon (DP) chips are dual capable, and are sold into the workstation and small server market.

Also note that Xeon MP chips are capable of 3-way and higher operation, and end up in larger servers.

Similarly, Opterons are available in the 100, 200 and 800 series. These are suitable for 1-way, 2-way, and up to 8-way operation respectively.

Some of these changes reflect packaging more than actual design (a Pentium Extreme Edition is a Xeon chip in a 1-way socket configuration).
 
Bear said:
I suspect (and have stated so elsewhere) that the "One More Thing"™ at this years WWDC will be something that may have been talked about in rumors briefly but is not really expected to be announced. And no I have no clue which it might be, this is just the feeling I have.
Click to expand...
Spoken like a true psychic. Hahaha! :p
 
AidenShaw said:
To emphasize your point, note that Xeon (DP) chips are dual capable, and are sold into the workstation and small server market.

Also note that Xeon MP chips are capable of 3-way and higher operation, and end up in larger servers.

Similarly, Opterons are available in the 100, 200 and 800 series. These are suitable for 1-way, 2-way, and up to 8-way operation respectively.

Some of these changes reflect packaging more than actual design (a Pentium Extreme Edition is a Xeon chip in a 1-way socket configuration).
Click to expand...

Why don;t apple have something for serves like AMD has oct Opterons for sale :eek: :confused:
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back