MacNN Forums (http://forums.macnn.com/)
-   Mac Desktops (http://forums.macnn.com/mac-desktops/)
-   -   The 970MP and low power 970FX have been officially announced. (pic) (http://forums.macnn.com/65/mac-desktops/262341/970mp-low-power-970fx-have-been/)

 
Eug Wanker Jul 7, 2005 11:25 AM
The 970MP and low power 970FX have been officially announced. (pic)
http://www.xlr8yourmac.com/

" 970MP and low power 970FX are emerged
IBM announced 970MP and low power 970FX at Power Everywhere Forum 2005 in Tokyo.
Japanese press release

970MP
dual core
1.4-2.5GHz
each core has 1MB L2 cache (2MB total)
one core can be swiched off for low power operation

Low Power 970FX
13W@1.4 GHz, 16W@1.6GHz (Typical?)

Die photo of 970MP

-M. Isobe "


http://ascii24.com/news/i/topi/artic...ages779599.jpg
 
gudrummer Jul 7, 2005 12:58 PM
Oh well...i have to save money again...hehehehe.
Just kiddin',more than happy with mine.
 
Eug Wanker Jul 7, 2005 02:01 PM
Quote, Originally Posted by gudrummer
Oh well...i have to save money again...hehehehe.
Just kiddin',more than happy with mine.
Yeah, but those quads are a comin'... :)
 
Simon Jul 7, 2005 02:27 PM
Too little, too late.
 
Eug Wanker Jul 7, 2005 03:02 PM
Quote, Originally Posted by Simon
Too little, too late.
"Too late" to keep Apple past 2007, but there are still 2 years of PowerPC Macs coming. And by no means is a dual-core 2.5 GHz G5 "too little".

Maybe you should have said... "Good stuff, but too late." :)
 
action Jul 7, 2005 03:07 PM
wasn't everyone anticipating these chips for a while? the issue isn't whether these chips exist or existed, the issue is how long will it take before they are fabbed in quanity to be put into a mac?

i think the release of this info is more pr damage control to show everyone they had chips for apple.

two years is a long way to wait for macintel powermac obviously, and i hope the dual cores are released soon but with ibm's prior history with apple i do not have much confidence we will see these machines within the next 6 months at the earliest and at the rate powermacs have been updated, that would put it on schedule for one bump prior to the macintel's release.

i can't wait for the bench marks.

chung lee
 
Eug Wanker Jul 7, 2005 03:14 PM
Quote, Originally Posted by action
wasn't everyone anticipating these chips for a while? the issue isn't whether these chips exist or existed, the issue is how long will it take before they are fabbed in quanity to be put into a mac?

i think the release of this info is more pr damage control to show everyone they had chips for apple.

two years is a long way to wait for macintel powermac obviously, and i hope the dual cores are released soon but with ibm's prior history with apple i do not have much confidence we will see these machines within the next 6 months at the earliest and at the rate powermacs have been updated, that would put it on schedule for one bump prior to the macintel's release.
Everyone was anticipating these chips for a while, but still, IBM never did officially announce the chips... until today.

I fully expect to see 970MP chips in retail machines within 6 months. I expect quad Macs too.

I'm not so sure about the low power G5 for laptops though. It's quite possible Apple will stick with the G4 7448, and then go to Yonah after that, skipping the G5 entirely for PowerBooks.
 
d.fine Jul 7, 2005 04:28 PM
I'm very curious about the next months...
 
Chinasaur Jul 7, 2005 07:16 PM
Hey Eug,

Estimates on increased heat generation? I'm no rocket scientist so I can't intuitively tell whether two cores are going to be hotter than one at a given clock speed given improvements that might be included in the new dies/materials/etc.

Any ideas?
 
Superchicken Jul 7, 2005 08:46 PM
I dono about for the PB. Are we sure Apple is going to want to put the Yohan in the PB ASAP? Seems a bit weird, move your Pro books to a new chip architecture ASAP before most apps are even developed for it? I don't know that we'll see that honestly. I wouldn't be surprised if we see a G5 PowerBook produced and then see that co-exist with the Pentium M based PowerBooks for a while.
 
Eug Wanker Jul 7, 2005 08:50 PM
Quote, Originally Posted by Chinasaur
Estimates on increased heat generation? I'm no rocket scientist so I can't intuitively tell whether two cores are going to be hotter than one at a given clock speed given improvements that might be included in the new dies/materials/etc.
Dunno, but I'm guessing in the ballpark of 130ish Watts for the dual-core 970MP 2.5 GHz.

Six months ago, Apple said the G5 970FX 2.3 GHz was 55 Watts:

Efficiency Pays
And all that power won't heat up your server closet nearly as much as the competition. Even at 2.3 GHz, the advanced G5 processor consumes at most 55W per processor. Compare that to 89W for an Opteron or 110W for a Xeon. The Xserve G5 consumes less power from the wall, you get lower heat output into your data center, reducing its cooling needs. That results in lower HVAC bills, or lets you put more Xserve G5 servers in the same amount of space for the same cost. An added bonus, lower power draws make for longer run-times on UPS systems, if configured.
 
Eug Wanker Jul 7, 2005 10:48 PM
Press release in English

IBM today announced the newest member of the Power Architecture family of microprocessors -- the PowerPC 970MP. The new processor is a dual-core version of IBM's award winning PowerPC 970FX, targeted for clients who desire a low-cost, high performance, 64-bit, symmetric multiprocessing (SMP)-capable system in a small package with ranges from 1.4 to 2.5 GHz. The microprocessor also provides power-saving features that system architects can use to dynamically control the system power.

The IBM PowerPC 970MP microprocessor builds on the proven 64-bit IBM Power Architecture family and is designed for entry level servers as well as to provide new levels of performance and power management for the embedded marketplace. The increased computing density of the PowerPC 970MP brings a new level of performance to a variety of applications, from HPC clusters to demanding embedded system applications such as high performance storage, single board computer and high performance networking applications.

Each of the two 64-bit PowerPC 970MP cores has its own dedicated 1MB L2 cache, resulting in performance more than double that of the PowerPC 970FX. This design provides clients with a wide range of performance and power operating points that can be selected dynamically to match system processing needs. The frequency and voltage of both cores can be scaled downward to reduce the power during periods of reduced workload. For further power savings, each core can be independently placed in a power-saving state called doze, while the other core continues operation. Finally, one of the cores can be completely de-powered during periods of less stringent performance requirements.

IBM also announced today new low-power extensions to its award-winning PowerPC 970FX offering. This newest offering is targeted for clients who desire a low-cost 64-bit processor featuring high performance, a sub-20 Watt power envelope and SMP. The new offering is targeted to provide an operating power of 13W at 1.4 GHz and 16W at 1.6GHz under typical workloads. The microprocessor also provides power-saving features that system architects can use to dynamically control the system power.

The 64-bitPowerPC 970FX microprocessor builds on the proven 64-bit IBM Power Architecture family and is suited to embedded applications including imaging and networking, and provides new levels of performance and power management for the embedded marketplace. Designed to run at frequencies up to 2.7 GHz, the PowerPC 970FX includes a 512KB L2 cache, provides native 64-bit and 32-bit application compatibility and uses a high bandwidth processor bus capable of delivering up to 7.1 GB/s to keep the processor core and the SIMD/Vector engine fed with data. The processor core can dispatch five instructions per cycle, and issue one instruction per cycle to each of its ten execution units, including two fixed point, two floating point, two load store, two vector and two system units. The L1 instruction cache holds 64 KB, the L1 data cache holds 32 KB, and each processor has its own dedicated 1MB L2 cache.
 
Talleyman Jul 8, 2005 12:03 AM
I agree...too little to late. The announcement is not the availability or the clock speed that SJ or IBM promised. Dual core is nice, but what will the availability be? How will the low power chip really work.

IBM/Apple screwed the pooch on the G5. Apple should have seen some volume before they committed and IBM shouldn't have made those wonderful promises without something to back it up.
 
Wet Jimmy Jul 8, 2005 12:29 AM
Decisions, decisions.
Well, this put a spanner in the works - Next week I was intending on purchasing a Dual 2.7 PowerMac - now I don't know if I should wait until the Dual Core's hit the market...

Aside from bragging rights, what are the reasons to wait for PowerMacs containing the new chips? Will I get better performance from a dual-core 2.5GHz than from a dual processor 2.7GHz?

Any insights?

Cheers,

WJ.
 
Talleyman Jul 8, 2005 01:05 AM
I wouldn't wait. If you need it, get it now. If you can wait, possibly indefinately, wait. But there is NO info on when or if this will appear in a G5. Go for it, and in two or three years you can buy a Mactel machine.

My 2 bits.
 
goMac Jul 8, 2005 01:09 AM
Quote, Originally Posted by Superchicken
I dono about for the PB. Are we sure Apple is going to want to put the Yohan in the PB ASAP? Seems a bit weird, move your Pro books to a new chip architecture ASAP before most apps are even developed for it?
The feeling I've gotten in the development community is that most people have already made it to x86. The only sticky part I see in an x86 move is that while Safari has been ported, it won't work with any browser plugins built for PPC. This means to use.. say... Flash... you have to boot a PowerPC copy of Safari.

I seriously expect the only company not to have updates ready in time for x86 to be Microsoft. We'll have to wait a while again for Virtual PC and Windows Media Player (which is were the stupid x86 browser plugin comes into play). Office won't probably be native until the next update, but thats ok because it can run under Rosetta.

Remember, nearly any app that isn't ready runs under Rosetta anyway. So really, we're ready now to move.
 
Obi Wan's Ghost Jul 8, 2005 01:33 AM
In practical terms a dual dualcore 2.5 with 1 meg cache each is equal to a 5 or 6 ghz computer. That's a Mac built to last and handle anything you throw at the beast for 5 years or so.
 
G5man Jul 8, 2005 02:15 AM
Apple please keep PPC on top of the line models! sorry it maybe too late but you never know if Intel doesn't come to Apple expectations I am going to have a laugh
 
Don Pickett Jul 8, 2005 02:22 AM
Quote, Originally Posted by Obi Wan's Ghost
In practical terms a dual dualcore 2.5 with 1 meg cache each is equal to a 5 or 6 ghz computer. That's a Mac built to last and handle anything you throw at the beast for 5 years or so.
Nope. The article doesn't give details, but I would imagine that both cores share one bus, so a dual dual-core 2.5 GHz would cut the effective bus speed for each processor core down to 625 MHz. All four cores also share the same memory and I/O systems, so it is common to get higher latencies in dual-core systems as well.

Not saying dual-cores aren't good, but you don't get anything like linear performance scaling.
 
Big Mac Jul 8, 2005 02:22 AM
Unfortunately, Apple has committed itself to an all PC lineup, so I don't think there's much hope.
 
Talleyman Jul 8, 2005 02:25 AM
Quote, Originally Posted by G5man
Apple please keep PPC on top of the line models! sorry it maybe too late but you never know if Intel doesn't come to Apple expectations I am going to have a laugh
Huh? That's too simple of a calculation. Doesn't work like that....I don't think it does anyway. Everything I've read ( I am a total layman when it comes to this ) is that you will see faster speeds, but not necessarily double. And definately not more than double.

G5man: not gonna happen. done deal. IBM couldn't deliver.
 
Wet Jimmy Jul 8, 2005 02:26 AM
Quote, Originally Posted by Obi Wan's Ghost
In practical terms a dual dualcore 2.5 with 1 meg cache each is equal to a 5 or 6 ghz computer. That's a Mac built to last and handle anything you throw at the beast for 5 years or so.
But isn't this exactly what a dual processor is - two processors, two lots of cache and two independent FSB's...

Please correct me if I'm wrong, I really don't understand all this (hence the questions), but I really fail to see the benefit of dual-core aside from heat/power and manufacturing costs.
 
Don Pickett Jul 8, 2005 02:33 AM
Quote, Originally Posted by Wet Jimmy
But isn't this exactly what a dual processor is - two processors, two lots of cache and two independent FSB's...

Please correct me if I'm wrong, I really don't understand all this (hence the questions), but I really fail to see the benefit of dual-core aside from heat/power and manufacturing costs.
Dual cores have definite advantages with applications which are easily multi-threaded; another core means another processor on which to run threads. Like I said above, there are disadvantages with memory and I/O access, but you will generally see performance gains. How much depends on the application and the OS.
 
Wet Jimmy Jul 8, 2005 02:43 AM
Quote, Originally Posted by Don Pickett
Dual cores have definite advantages with applications which are easily multi-threaded; another core means another processor on which to run threads. Like I said above, there are disadvantages with memory and I/O access, but you will generally see performance gains. How much depends on the application and the OS.
... but don't Dual Processors take advantage of applications which are multi-threaded, too? And don't dual processor systems have two lots of cache, just like the dual core?

Again, what advantage, if any, does Dual CORE have over Dual PROCESSORS?
 
Don Pickett Jul 8, 2005 03:18 AM
Quote, Originally Posted by Wet Jimmy
... but don't Dual Processors take advantage of applications which are multi-threaded, too? And don't dual processor systems have two lots of cache, just like the dual core?

Again, what advantage, if any, does Dual CORE have over Dual PROCESSORS?
Because you'd typically have two dual-core processors in one box – four processor cores versus two.
 
Big Mac Jul 8, 2005 03:31 AM
If 2.7GHz G5s have to be water cooled (and still run at high temperatures), how in the world would dual dual cores be cooled?
 
Simon Jul 8, 2005 03:42 AM
Problem is, Apple won't sell PowerMacs at the current price point (or less as they should) with dual 970MPs running at 3.0GHz. They'll sell single MPs running at a max of 2.5GHz. Basically, the deal is that instead of dual CPUs we get dual cores but with more cache because IBM fscked up the clock increase. Meh, blech. This IBM update sucks. I'm glad Steve lit some Intel fire under their asses.
 
Wet Jimmy Jul 8, 2005 03:52 AM
Okay, so there is no real performance advantage to a single dual core over a dual processor. That's what I was trying to understand.
 
Don Pickett Jul 8, 2005 04:04 AM
Quote, Originally Posted by Wet Jimmy
Okay, so there is no real performance advantage to a single dual core over a dual processor. That's what I was trying to understand.
What? Dude, pay attention: a dual-core processor will be faster than a single core processor. It won't be twice as fast, but it will be faster, especially with operations which can be multi-threaded and/or highly parallelized. I imagine that a two processor dual-core G5 crunching the kind of data suited for AltiVec would be an incredibly fast machine. In general purpose computing you would see a speedup as well.
 
Obi Wan's Ghost Jul 8, 2005 04:12 AM
Quote, Originally Posted by Don Pickett
Nope. The article doesn't give details, but I would imagine that both cores share one bus, so a dual dual-core 2.5 GHz would cut the effective bus speed for each processor core down to 625 MHz. All four cores also share the same memory and I/O systems, so it is common to get higher latencies in dual-core systems as well.

Not saying dual-cores aren't good, but you don't get anything like linear performance scaling.
5 or 6 ghz is not linear. It's based on each additional processor adding about 33% extra performance. That comes to about 833Mhz out of each extra cpu. So 833Mhz x 3 = 2.5ghz + 2.5ghz CPU0 = 5Ghz. Some apps have very good altivec performance and so the computer can give even better performance than a 5Ghz single cpu.
 
Don Pickett Jul 8, 2005 04:23 AM
Quote, Originally Posted by Obi Wan's Ghost
5 or 6 ghz is not linear. It's based on each additional processor adding about 33% extra performance. That comes to about 833Mhz out of each extra cpu. So 833Mhz x 3 = 2.5ghz + 2.5ghz CPU0 = 5Ghz. Some apps have very good altivec performance and so the computer can give even better performance than a 5Ghz single cpu.
I still don't think you'd only see that performance in the best of circumstances – cache issues, memory, etc.
 
Catfish_Man Jul 8, 2005 04:35 AM
Quote, Originally Posted by Don Pickett
What? Dude, pay attention: a dual-core processor will be faster than a single core processor. It won't be twice as fast, but it will be faster, especially with operations which can be multi-threaded and/or highly parallelized. I imagine that a two processor dual-core G5 crunching the kind of data suited for AltiVec would be an incredibly fast machine. In general purpose computing you would see a speedup as well.
He didn't say a single single core processor. He's comparing a single dual core, to dual single cores, and is correct that there's not much (if any) gain. Bandwidth would go down slightly, and latency between cores would improve some. Not sure how much effect that would have in total.
 
Wet Jimmy Jul 8, 2005 05:43 AM
Quote, Originally Posted by Catfish_Man
He didn't say a single single core processor. He's comparing a single dual core, to dual single cores
Finally! Thankyou, CM!

That's what I thought. So in light of this, while I understand a quad processor (Dual Dual-Core) would be nice, a plain old Dual Processer in the form of a PM DP2.7GHz is good enough for me.

I'm kidding myself if I'd use all the extra grunt, anyway!
 
Gamicoulas Jul 8, 2005 05:45 AM
Please guys stop this too late stuff. The whole switch has nothing to do with the performance!!! It's a fact that the PPC is a better architecture than Intel. End of argument...

The switch is all about moving on the other side of the fence and compete with the other OSes. This is armageddon in the making and an attempt for world domination in the consumer desktop OS. And that can only happen if OSX is running on Intel... If Windows where running on ARM processors that's what Apple would have gone with. This is win or lose all scenario, that if successful will require licensing OSX in the near future when some momentum is there...

If you wanna compete in the States on Nascar or Indy car racing then you don't bring your F1 car along, you build a Nascar or an Indy car and go and compete with it. It does not matter if your F1 car is running circles around the other cars, you are not allowed to compete with that one...

It might be a crap example but do you dig it now???
 
Link Jul 8, 2005 06:01 AM
Woohoo! We're back in business!

Some of you guys sound like such morons. When IBM was acting like Motorola (quite literally), you trashed motorola and cheered IBM on, and now that IBM isn't the main player, you're cheering on Intel!

Hypocrites. The whole thing is that IBM is actually ahead of the game. Have you actually seen any dual core machines shipping yet from ANYONE? Any Opterons over 2.7ghz? (feel free to correct me if I'm wrong)... Last I checked AMD likes IBM's production so much they use them and Intel is about as reliable as motorola when it comes to things like this....

Motorola's not far from the fact either, there's that 8641D that's been burning a hole in the production table for a while now, wonder what's better a dualcore 8641D or a single core 970fx... hrm faster bus vs higher CPU bandwidth... tricky one.
 
DrBoar Jul 8, 2005 06:38 AM
PPC might be "better" than the cludgy x86 but the CPU also work in a context. In personal computers the G4 is limited by tha slow bus and the G5 might have a very fast bus for streaming data but latencies is way worse than for Pentium 4 or especially the the modern AMDs with IMC.
http://www.anandtech.com/mac/showdoc.aspx?i=2436&p=9
The doubling of cache might help some but I think that the AMD will be king of the hill anyhow.

The low power G5 looks nice but not better than a currently aviable G4 in most apps...
 
otheronenorehto Jul 8, 2005 07:01 AM
My question is will the dual core processors come with DDR2 and PCI-e?
 
Lateralus Jul 8, 2005 07:29 AM
Quote, Originally Posted by otheronenorehto
My question is will the dual core processors come with DDR2 and PCI-e?
Most likely.
 
otheronenorehto Jul 8, 2005 07:54 AM
will they be out before cristmas though? probably not...
 
Big Mac Jul 8, 2005 08:04 AM
Quote, Originally Posted by Lateralus
Most likely.
It took Apple a damn long time to finally get to DDR. So why would Apple expend the resources to get PCIe and DDR2 on board a Mac platform that has no future? I'm a bit incredulous concerning the notion that dual core G5s will ever appear. Mactel is, for better or worse, Apple's future. We have to resign ourselves to the fact that the Mac is at EOL.
 
Eriamjh Jul 8, 2005 08:20 AM
Quote, Originally Posted by Big Mac
Unfortunately, Apple has committed itself to an all PC lineup, so I don't think there's much hope.
Actually, an all Intel lineup, an that is not until June-July (or even later) 2007. I just checked and it's still 2005. :err:

Apple BETTER use these processors and quickly. I can't imagine them dropping the 2.7GHz model in favor of, say, a Twin Dual-core 2.5GHz machine, but it is possible that such a beast be sold side-by-side with a dual 2.7.

I'm ready to buy anytime now. I don't want to get the last PPC model, nor do I want the first Mac-In-tel (I hate that name). I want to avoid both machines and enjoy what I'd be buying now.

Quote, Originally Posted by BIG Mac
It took Apple a damn long time to finally get to DDR. So why would Apple expend the resources to get PCIe and DDR2 on board a Mac platform that has no future? I'm a bit incredulous concerning the notion that dual core G5s will ever appear. Mactel is, for better or worse, Apple's future. We have to resign ourselves to the fact that the Mac is at EOL.
Apple didn't move to DDR because the G4 has such a sh*tty bus design that it wasn't much of an improvement over PC133. They jumped on PCI-X pretty fast, but it looks like PCI-e should have been the choice. It's not the first time they goofed (look at the DVD-RAM machines of 2000).

Apple still has to design Macs that sell, and much like people who hang onto OS9, there will be those who hang onto PPC well into the Intel-era (like those who still need classic). Whose to say that IBM won't have any breakthroughs that help the PPC linger and allow Apple to sell both machines? (Actually, even I don't believe that will happen, but Jobs is a fickle man. You never know.)

Hang on. It's gonna be a wild ride.
 
blackwind Jul 8, 2005 08:47 AM
Quote, Originally Posted by Link
Woohoo! We're back in business!

Some of you guys sound like such morons. When IBM was acting like Motorola (quite literally), you trashed motorola and cheered IBM on, and now that IBM isn't the main player, you're cheering on Intel!

Hypocrites. The whole thing is that IBM is actually ahead of the game. Have you actually seen any dual core machines shipping yet from ANYONE? Any Opterons over 2.7ghz? (feel free to correct me if I'm wrong)... Last I checked AMD likes IBM's production so much they use them and Intel is about as reliable as motorola when it comes to things like this....

Motorola's not far from the fact either, there's that 8641D that's been burning a hole in the production table for a while now, wonder what's better a dualcore 8641D or a single core 970fx... hrm faster bus vs higher CPU bandwidth... tricky one.
Dual-core machines are available.

The closest competitor to the G5 would be the Opteron. Currently, dual-core 2.2-GHz Opterons (model numbers 275 or 875 depending on 2-way or 8-way capability) are available, but each processor alone costs a ludicrous amount of money. By mid-2006, I would expect dual-core 2.6-GHz Opterons being available. If a quad-core Power Mac G5 at 2.5 GHz appear at that time, it won't actually be in a particularly bad position.

The dual-core Xeons (which can be used in multi-processor machines) will appear in early 2006. It is unlikely for them to clock higher than 3.4 GHz (initially).

The dual-core Athlon 64 X2 is also available at 2.4 GHz. Dual-core Pentium D's are also available, but they are crippled with no Hyperthreading... The 3.2-GHz Pentium XE 840 has Hyperthreading, and it is available. Of course, none of these chips are multi-processor-capable, unlike the Opteron.

The fastest single-core processor from AMD is the 2.8-GHz Athlon 64 FX-57. The fastest single-core and multi-processor-capable processor from AMD is the 2.6-GHz Opteron x52 (x being either "2" or "8").

The fastest single-core processor from Intel is the 3.8-GHz Pentium 4 670 (well, actually it is probably the 3.73-GHz Pentium 4 EE). The fastest single-core and multi-processor-capable processor from Intel is the 3.6-GHz Xeon.

The current single-core 2.7-GHz G5 is capable of competing with any of the above single-core processors, so a dual-core 2.5-GHz G5 (with its larger L2 per core) should be just fine.
 
Eug Wanker Jul 8, 2005 09:05 AM
Quote, Originally Posted by Big Mac
Unfortunately, Apple has committed itself to an all PC lineup, so I don't think there's much hope.
I doubt the Power Macs will go x86 until 2007. Apple's contract with IBM doesn't end until 2007 anyway.

Quote, Originally Posted by Wet Jimmy
But isn't this exactly what a dual processor is - two processors, two lots of cache and two independent FSB's...

Please correct me if I'm wrong, I really don't understand all this (hence the questions), but I really fail to see the benefit of dual-core aside from heat/power and manufacturing costs.
Heat/power and manufacturing costs are very important. Furthermore, quads become a lot more likely when dual-core SMP-able chips are available.

Quote, Originally Posted by Big Mac
If 2.7GHz G5s have to be water cooled (and still run at high temperatures), how in the world would dual dual cores be cooled?
They'd likely be watercooled too at 2.5 GHz. Remember, the supposed main reason for watercooling in the top end Power Macs was not overall heat, but heat density. Heat density in a dual-core design doesn't increase, even though overall heat does.

Quote, Originally Posted by Simon
Problem is, Apple won't sell PowerMacs at the current price point (or less as they should) with dual 970MPs running at 3.0GHz. They'll sell single MPs running at a max of 2.5GHz. Basically, the deal is that instead of dual CPUs we get dual cores but with more cache because IBM fscked up the clock increase. Meh, blech. This IBM update sucks. I'm glad Steve lit some Intel fire under their asses.
:rolleyes: Stop with the FUD. Of course Apple won't sell Power Macs at the current price point with dual 970MP running at 3.0 GHz, because there is no such thing as a 970MP 3.0. And given the speeds of the 970FX, we shouldn't have expected them either.

The cache increase is no surprise. It's common sense, and the Think Secret rumour predicted this amount of cache 1 year ago. This IBM update is great, even if it is a bit late.

And yes, I do expect to see quad Macs, although only at the high end:

http://www.pbase.com/eugwanker/image/42930828.jpg

A quad G5 2.5 would be a killer machine, by any reasonable measure.

Quote, Originally Posted by Catfish_Man
He didn't say a single single core processor. He's comparing a single dual core, to dual single cores, and is correct that there's not much (if any) gain. Bandwidth would go down slightly, and latency between cores would improve some. Not sure how much effect that would have in total.
According to IBM:

"Each of the two 64-bit PowerPC 970MP cores has its own dedicated 1MB L2 cache, resulting in performance more than double that of the PowerPC 970FX."

This makes a lot of sense. I would expect that despite the bandwidth issues, a single 970MP 2.5 would likely be faster in most circumstances than a dual 970FX 2.5. However, there are specific circumstances where the dual 970FX 2.5 could be faster.

Furthermore, as mentioned earlier, the 970MP brings the strong possibility of quads.

Quote, Originally Posted by Gamicoulas
It's a fact that the PPC is a different architecture than Intel x86.
Fixed.

Quote, Originally Posted by Link
Some of you guys sound like such morons. When IBM was acting like Motorola (quite literally), you trashed motorola and cheered IBM on, and now that IBM isn't the main player, you're cheering on Intel!

Hypocrites. The whole thing is that IBM is actually ahead of the game. Have you actually seen any dual core machines shipping yet from ANYONE? Any Opterons over 2.7ghz? (feel free to correct me if I'm wrong)... Last I checked AMD likes IBM's production so much they use them and Intel is about as reliable as motorola when it comes to things like this....

Motorola's not far from the fact either, there's that 8641D that's been burning a hole in the production table for a while now, wonder what's better a dualcore 8641D or a single core 970fx... hrm faster bus vs higher CPU bandwidth... tricky one.
I don't have a lot of confidence in Freescale's ability to be a desktop chip pioneer, and indeed that is not their aim.

Anyways, I find it amusing that before the Intel announcement I used to have to defend x86 quite a bit in these types of discussions. However, these days, I find I have to defend PPC, cuz suddenly Intel is gonna save the world. :lol:

Quote, Originally Posted by otheronenorehto
My question is will the dual core processors come with DDR2 and PCI-e?
I damn well hope so. :p I think PCIe is a lock, and DDR2 is likely but not guaranteed.

Quote, Originally Posted by Big Mac
It took Apple a damn long time to finally get to DDR.
G4s can't use DDR, so it's irrelevant how long it took to get DDR (from a technical standpoint). DDR on a G4 is essentially equivalent to SDR. IMO the main reason for G4 Macs to support DDR was marketing. Fortunately, DDR is now cheaper than SDR too.
 
Simon Jul 8, 2005 10:36 AM
When was that that IBM promised 3.0GHz 970s? :rolleyes:

Now, more than a year later we get MPs at 2.5GHz. That's not FUD, it's plain failure to deliver on a promise. Competition is good. Intel's competition is good. Ideology isn't. It's nice Steve agrees. :thumbsup: :)
 
Eug Wanker Jul 8, 2005 10:43 AM
Quote, Originally Posted by Simon
When was that that IBM promised 3.0GHz 970s? :rolleyes:

Now, more than a year later we get MPs at 2.5GHz. That's not FUD, it's plain failure to deliver on a promise. Competition is good. Intel's competition is good. Ideology isn't. It's nice Steve agrees. :thumbsup: :)
IBM has never promised 3.0 GHz G5 chips of any sort. Stevie did. Get your facts straight. In fact, if you watch the actual keynote, when Steve announced 3.0 GHz, the IBM guy seemed quite taken aback.

And if you're gonna talk about failed roadmaps, where are those 5 GHz Tejas chips?
 
otheronenorehto Jul 8, 2005 10:47 AM
but will they be out by christmas...

Is it worth it to wait for the possibility of Pci-e or buy now...
 
blackwind Jul 8, 2005 11:01 AM
Quote, Originally Posted by Eug Wanker
And if you're gonna talk about failed roadmaps, where are those 5 GHz Tejas chips?
Beating the crap out of that 3-GHz PowerPC 970, which could only match a measly 4.5-GHz Tejas? :D

I'm with Eug with not understanding the huge deal. Intel had to scale back from 3.8 GHz to 3.2 GHz for adapting the Pentium 4 Extreme to the dual-core Pentium XE 840 and Pentium D. Even if that 3 GHz PowerPC 970 existed, it would have been scaled back for a dual-core G5 release.

AMD has a 2.8-GHz Athlon 64 FX-57, and it has a dual-core 2.4-GHz Athlon 64 X2 counterpart, so a 3-GHz G5 being toned down to 2.5-GHz for a dual-core G5 would have been expected anyway.

EDIT: Well, it's not quite fair, as the Athlon 64 FX-57 was released at about the same time as the Athlon 64 X2. Regardless, the old 2.6-GHz Opteron x52 series was toned down to 2.2 GHz for the dual-core Opteron x75 series.
 
Agent69 Jul 8, 2005 12:27 PM
Actually, the Pentium 4Ds start at 2.8ghz.
 
blackwind Jul 8, 2005 01:34 PM
Quote, Originally Posted by Agent69
Actually, the Pentium 4Ds start at 2.8ghz.
... but their maximum speed is 3.2 GHz.

I probably should have been clearer by mentioning that I was considering only the top-end models, so I apologize for any confusion. :)
 
Person Man Jul 8, 2005 05:18 PM
Quote, Originally Posted by Big Mac
Unfortunately, Apple has committed itself to an all PC lineup, so I don't think there's much hope.
Would you QUIT saying "PC?"

APPLE IS NOT GOING TO MAKE PCS! They're going to make MACINTOSHES, although they will have Intel chips in them.

A processor alone does not a generic PC make, and the rest of the hardware ain't going to be generic.
 
All times are GMT -4. The time now is 10:30 PM.

Copyright © 2005-2007 MacNN. All rights reserved.
Powered by vBulletin® Version 3.8.8
Copyright ©2000 - 2014, vBulletin Solutions, Inc.


Content Relevant URLs by vBSEO 3.3.2