Nope. LOL NOTHING in Wintel is that smart. You set it all in BIOS. everything runs at the speed of the dumbest component usually. EG my FSB runs at 133 so the 333 memory has to run at 2x which is 266 MHz. If you installed everything in default configuration it would have the FSB running at 100 MHz and the 1800 MHz chip running at 1500 MHz and the RAM at 133 MHz and AGP 1x... it would be solid as a rock but slow as a snail. I often wonder how many people have their systems set up like this.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
IBM Preps for Apple Announcements
- Thread starter MacRumors
- Start date
- Sort by reaction score
Nope. LOL NOTHING in Wintel is that smart. You set it all in BIOS. everything runs at the speed of the dumbest component usually. EG my FSB runs at 133 so the 333 memory has to run at 2x which is 266 MHz. If you installed everything in default configuration it would have the FSB running at 100 MHz and the 1800 MHz chip running at 1500 MHz and the RAM at 133 MHz and AGP 1x... it would be solid as a rock but slow as a snail. I often wonder how many people have their systems set up like this.
So I wonder how this is going to work on the Mac. I get the feeling that the RAM DIMM doesn't have the info on its max speed in its eeprom (or whatever); if it did, some mobo manufacturer would have already implemented this "intelligent" functionality in bios/firmware/whatever.
But we can't fiddle with the bios on the Mac. So what does that mean? Will the Mac not recognize the new PC4000 DIMMs and speed up the ram timing when you install them? That would kinda suck.
Guys;
I know you are all Mac fans but come on at least look at what is happening itn the PC world from time to time.
Amd had implemented open transport. Intel has 800MHz bus systems ( which by the way over clock easily) Serial ATA has been around for a while now, mother boards are often shipping with both Serial ATA and Standard ATA ports. Many chipsets in the Intel world now support Firewire or is other wise impl;emented on the motherboard. To think that any of this is new and that Apple is leading the way is a little sad.
970's competition is not the Pentium in any of its current implementations. What we need to look out for is the Opterion and Athlon64 and the latest from intel. The 970 has to be compared with chips that are currently being introduced in the intel domain otherwise we are pulling the wool over our own eyes.
I stand by my assertion that these specs, if true, only brings Apple into parity with the PC world. The only advantage that is clear at this moment is the 970's true 64 bit address space and its raw performance. In a very short amount of time we will know more, but do understand that the darkside is also introducing hardware this month and next.
It is very possible that this massive peformance lead will last less than 30 days. I do hope that Apple and IBM do have plans to move ahead quickly with peformance gains. I'm talking a 150 to a couple of hundred MHz a month by the end of the year. Otherwise the weenies will be laughing the behinds off again. One thing you have to realize is that once you remove the MicroSoft OS from an Intel machine the performance is pretty darn good, it is in apples best interest to serve up economical machines that change the perspective on this completely.
If the posted spec are true than we can determine that there are some innovation such as 64 bit computing and firewire 800. But honestly these are not unexpected or breath taking. Lets face it a processor upgrade is long over due. Breath taking would have been a machine with an OpenTransport driven graphics controller, or the ability to address 128 GB of main memmory, or this much power in an LC size box. Nope we are just pulling up even maybe slightly ahead for a few days.
Dave
I know you are all Mac fans but come on at least look at what is happening itn the PC world from time to time.
Amd had implemented open transport. Intel has 800MHz bus systems ( which by the way over clock easily) Serial ATA has been around for a while now, mother boards are often shipping with both Serial ATA and Standard ATA ports. Many chipsets in the Intel world now support Firewire or is other wise impl;emented on the motherboard. To think that any of this is new and that Apple is leading the way is a little sad.
970's competition is not the Pentium in any of its current implementations. What we need to look out for is the Opterion and Athlon64 and the latest from intel. The 970 has to be compared with chips that are currently being introduced in the intel domain otherwise we are pulling the wool over our own eyes.
I stand by my assertion that these specs, if true, only brings Apple into parity with the PC world. The only advantage that is clear at this moment is the 970's true 64 bit address space and its raw performance. In a very short amount of time we will know more, but do understand that the darkside is also introducing hardware this month and next.
It is very possible that this massive peformance lead will last less than 30 days. I do hope that Apple and IBM do have plans to move ahead quickly with peformance gains. I'm talking a 150 to a couple of hundred MHz a month by the end of the year. Otherwise the weenies will be laughing the behinds off again. One thing you have to realize is that once you remove the MicroSoft OS from an Intel machine the performance is pretty darn good, it is in apples best interest to serve up economical machines that change the perspective on this completely.
If the posted spec are true than we can determine that there are some innovation such as 64 bit computing and firewire 800. But honestly these are not unexpected or breath taking. Lets face it a processor upgrade is long over due. Breath taking would have been a machine with an OpenTransport driven graphics controller, or the ability to address 128 GB of main memmory, or this much power in an LC size box. Nope we are just pulling up even maybe slightly ahead for a few days.
Dave
Right, and that is why Apple would do well by trying to employ more of the same sort of options that Dell offers. I don't want a choice between, 4 different Office suites, 3 OSes, 3 NICs, etc, etc; all I'm looking for is the ability to select the number and speed of processors in my new Mac.
Re: A simple question
Apple cannot recognize revenue until a product ships, so if Apple does not ship any existing products for 2-3 months, then Apple's quarterly statements will be killed.
However, as you point out Apple can essentially "lock up" the consumer's money to prevent him from spending it on another product, or at least the consumer may hesitate buying that "other" product.
A balance must be struck.
Apple cannot recognize revenue until a product ships, so if Apple does not ship any existing products for 2-3 months, then Apple's quarterly statements will be killed.
However, as you point out Apple can essentially "lock up" the consumer's money to prevent him from spending it on another product, or at least the consumer may hesitate buying that "other" product.
A balance must be struck.
Yep. Actually, we don't even know that yet, since we don't know the prices. If they stay at current G4 price points, then I think we are pretty close to parity.
The only sorta big deal in this (relative to wintel) is PCI-X. But that's nothing earth-shattering, especially since there are no cards out. And PCs will have that soon enough. There's no real advantage here since the cards will come out for Mac/PC at the same time.
I'm still excited about the new Macs though. God, it'll be nice to finally have a decently powerful machine!
Re: Re: PCI-X
PCI on standard PCs is 32-bit / 33MHz (133MB/s (around 128MiB/s maximum).
On Macs it is 64-bit / 33Mhz (266MB/s)
A PCI-X controller has 3 configurations:
- 1 slot, 64-bits, 133MHz (1GB/s)
- 2 slots, 64-bits, 100MHz (800MB/s shared)
- 4 slots, 64-bits, 66MHz (533MB/s shared)
I expect that the new Apple motherboard will be using an AMD 8131 dual PCI-X controller and possibly an AMD 8151 AGP8x controller, probably in this layout:
[ 970 ] --- [ Apple System Controller ] --- [ AMD 8131 ] --- [ AMD 8151** ] --- [ IO controller*** ]
The Apple System Controller would have a dual-DDR memory interface, one (or two) E-Bus interfaces for the processor(s), and a HyperTransport downlink to the next chip.
** AGP8x might be incorporated in the System Controller by Apple, thus negating the need for this chip.
***There is also a fair chance that Apple designed the IO controller as a PCI-X device, hence why there are only 3 PCI-X slots on the motherboard. OTOH it could simply be a HyperTransport IO device like a SATA variant of the AMD 8111 or nVidia nForce MCP-T.
I hope that we get nice pictures of the new Apple motherboard when reviews hit the streets next week.
PCI on standard PCs is 32-bit / 33MHz (133MB/s (around 128MiB/s maximum).
On Macs it is 64-bit / 33Mhz (266MB/s)
A PCI-X controller has 3 configurations:
- 1 slot, 64-bits, 133MHz (1GB/s)
- 2 slots, 64-bits, 100MHz (800MB/s shared)
- 4 slots, 64-bits, 66MHz (533MB/s shared)
I expect that the new Apple motherboard will be using an AMD 8131 dual PCI-X controller and possibly an AMD 8151 AGP8x controller, probably in this layout:
[ 970 ] --- [ Apple System Controller ] --- [ AMD 8131 ] --- [ AMD 8151** ] --- [ IO controller*** ]
The Apple System Controller would have a dual-DDR memory interface, one (or two) E-Bus interfaces for the processor(s), and a HyperTransport downlink to the next chip.
** AGP8x might be incorporated in the System Controller by Apple, thus negating the need for this chip.
***There is also a fair chance that Apple designed the IO controller as a PCI-X device, hence why there are only 3 PCI-X slots on the motherboard. OTOH it could simply be a HyperTransport IO device like a SATA variant of the AMD 8111 or nVidia nForce MCP-T.
I hope that we get nice pictures of the new Apple motherboard when reviews hit the streets next week.
Yes the wait is getting to be painful. Even more painfull is the realization that I will have to work a bunch of overtime to even consider one of these machines.
Hopefully they can come in slightly lower than G4 prices. I really love the MAC but performance and pricing have always been an issue. There is a differrence between paying a little more for value and getting taken to the cleaners. Lately the price performance ratio on the PowerMacs has been leaving people at the cleaners. If they can squeeze the low end into the 800 to 1000 dollar range that would be a start.
PCI-X is interesting, but my understanding is that there are differrent levels of implementation. We could get a significant implementation or we could something lesser that a maximum implementation. What would be really neat is if they stuck an Opentransport connection on the machine. My only concenr is backwards compatability with existing cards.
You aren't the only one that is excited. People need options besides the MS/Intel world, these machines may put apple into the running again. How long they remain in the running is another matter. That is why I believe that the IBM seminars may have the most important information of the conference. Things in the intel world have not slowed down, IBM is going to have a major challenge on their hands to get and keep leadership via the 970 & 980.
Dave
Hopefully they can come in slightly lower than G4 prices. I really love the MAC but performance and pricing have always been an issue. There is a differrence between paying a little more for value and getting taken to the cleaners. Lately the price performance ratio on the PowerMacs has been leaving people at the cleaners. If they can squeeze the low end into the 800 to 1000 dollar range that would be a start.
PCI-X is interesting, but my understanding is that there are differrent levels of implementation. We could get a significant implementation or we could something lesser that a maximum implementation. What would be really neat is if they stuck an Opentransport connection on the machine. My only concenr is backwards compatability with existing cards.
You aren't the only one that is excited. People need options besides the MS/Intel world, these machines may put apple into the running again. How long they remain in the running is another matter. That is why I believe that the IBM seminars may have the most important information of the conference. Things in the intel world have not slowed down, IBM is going to have a major challenge on their hands to get and keep leadership via the 970 & 980.
Dave
I guess you mean HyperTransport. Open Transport was a networking technology that was developed by Apple way back when.
IMO, the problem with this part of your analysis is that while these technologies are available on PCs there are few if any systems from major manufacturers that implement all of them in one product, particularly at the "personal computing" level. So, if you add Firewire 800, the rumored 1GHz system bus, the built-in wireless support, HyperTransport, PCI-X, etc. you are talking about a very high-end configuration.
I disagree with this also, but in this case I think you are overstating the "advantage" that the PPC970 has over PCs. The only way the PPC970 will outperform high-end Pentium/AMD systems is to run in dual configurations. At best, a single 2GHz PPC970 will be comparable to a 3GHz Pentium and the 64-bitness of the PPC970 won't really benefit most users.
That will be true only if PC manufacturers go mainstream with dual processors, which I don't think will happen anytime soon. But, it's true that in single processor configurations PCs will most like remain in the lead.
I will admit, however, that Apple has a bad habit of trying to compare the performance/features of the __next__ generation Mac to what is currently shipping in the PC market. So, yes it's a moving target, Apple will have to work hard to keep up. This is another reason why I'm skeptical about the "leaked" specs on the G5, these sound like a next generation product that may not ship for several more months.
Again, I guess you mean HyperTransport. However, I think more likely (and "Breath taking") would be PCI-Express. I suspect that HyperTransport will never be used on a high-end PC to serve as a graphics port (8X AGP will do that until PCI-Express arrives).
Your dreaming
Double the chips frequency in little over a year?! Your dreaming, IBM's 9XX processors don't have enough pipeline stages to match the Pentium 4s frequency. If IBM's 9XX chips get to 3.4GHz at the 90-nanometer level by the end of 2004, then that will be tremendous progress.
The 9XX chips not moving to 4GHz in 2004 doesn't necessarily mean lower performance than the Pentium 4. IBM has already stated that a upcoming Power5 computer is testing out at up to 4X faster than a Power4 computer, so a 980 processor based on the Power5 should have significantly higher performance than a 970 processor. But much of that speed increase will likely be in something other than frequency boosts. IBM has already indicated that the transistor count will not increase significantly moving to the Power5. Most of the performance increases will likely not be geared for a single processor desktop computer.
Double the chips frequency in little over a year?! Your dreaming, IBM's 9XX processors don't have enough pipeline stages to match the Pentium 4s frequency. If IBM's 9XX chips get to 3.4GHz at the 90-nanometer level by the end of 2004, then that will be tremendous progress.
The 9XX chips not moving to 4GHz in 2004 doesn't necessarily mean lower performance than the Pentium 4. IBM has already stated that a upcoming Power5 computer is testing out at up to 4X faster than a Power4 computer, so a 980 processor based on the Power5 should have significantly higher performance than a 970 processor. But much of that speed increase will likely be in something other than frequency boosts. IBM has already indicated that the transistor count will not increase significantly moving to the Power5. Most of the performance increases will likely not be geared for a single processor desktop computer.
I largely agree with the wizard's statements, and that is why I think Apple needs to bring out the IBM engineers to show that there is a solid path to the future.
The 970 is based on IBM's Power4, which was introduced in 2001. It is already an old design. It just happens to be a lot better than what Apple has been using.
IBM is already producing Power5 chips (probably test quantities), and they had Linux booting on it by this past March. I should hope that Apple is at least compiling Darwin on the Power5 chips.
See CNet article and post from an IBM engineer.
Re: PCI-X = 16x PCI
You're thinking of PCI Express, which will only start to appear in systems towards the end of this year, and is not PCI-X which is the same speed as AGP4x at best. PCI-X 2 doubles the speed of PCI-X. PCI Express 16x yet again doubles the speed, but with a completely different connector and low-level protocol.
You're thinking of PCI Express, which will only start to appear in systems towards the end of this year, and is not PCI-X which is the same speed as AGP4x at best. PCI-X 2 doubles the speed of PCI-X. PCI Express 16x yet again doubles the speed, but with a completely different connector and low-level protocol.
wiz,
pci-x and serial ATA aren't revolutionary, only evolutionary.
Opterion and Athlon64 are mainly used as server chips, with very few being used as home, soho and portable markets. the 970 will be used in all markets. IBM will use the 970 in desktops, workstations and servers. Apple will use the970 in portables, desktops, workstations, and servers. To compare the 970 to the Opterion and Athlon64 is the same as compering the Opterion and Athlon64 to the Power4 and Power 5.
To say the collection of the upcoming parts only brings the Macs up to parity to Intel/AMD levels is mostly garbage. Most people I know still buy PIV and Athalon mobos and proccessers. Most institions I know still buy PIV.The up coming Apple across the product line. Therefor, the Apple product line, taken as a whole, will be more advanced the Intel/AMD product offerings taken as a whole. For now.
Moving beyond the my platform is better than yours pissing match, I can see a more natural computing environment coming soon. Truly accurate voice recognition, smart operating systems that learn what you like and need - and setting them up as defaults. With nonvolital carbon nanotube RAM ( up to 10 GB) projected in the next 2 - 5 years and the nanotube processor components it won't matter any more. Your house, tranportation and workplace(s) will run something probabilly none of us can predict.
As for proving me wrong, I'll take you word on the mobos that have the included technologies. I didn't run across them while shopping to upgrade my 2 Pentiums and 1 PII. I'll look again.
pci-x and serial ATA aren't revolutionary, only evolutionary.
Opterion and Athlon64 are mainly used as server chips, with very few being used as home, soho and portable markets. the 970 will be used in all markets. IBM will use the 970 in desktops, workstations and servers. Apple will use the970 in portables, desktops, workstations, and servers. To compare the 970 to the Opterion and Athlon64 is the same as compering the Opterion and Athlon64 to the Power4 and Power 5.
To say the collection of the upcoming parts only brings the Macs up to parity to Intel/AMD levels is mostly garbage. Most people I know still buy PIV and Athalon mobos and proccessers. Most institions I know still buy PIV.The up coming Apple across the product line. Therefor, the Apple product line, taken as a whole, will be more advanced the Intel/AMD product offerings taken as a whole. For now.
Moving beyond the my platform is better than yours pissing match, I can see a more natural computing environment coming soon. Truly accurate voice recognition, smart operating systems that learn what you like and need - and setting them up as defaults. With nonvolital carbon nanotube RAM ( up to 10 GB) projected in the next 2 - 5 years and the nanotube processor components it won't matter any more. Your house, tranportation and workplace(s) will run something probabilly none of us can predict.
As for proving me wrong, I'll take you word on the mobos that have the included technologies. I didn't run across them while shopping to upgrade my 2 Pentiums and 1 PII. I'll look again.
Thanks for the PCI/X info, knew there where differrent implementations but not the specifics.
IBM it seems has not come real clean on the I/O bus implementation for the 970. Or to put it another way I haven't read in depth on it yet. What is the possibility that the 970 has opentransport ports built into the chip itself? It seems like all that IBM has done is introduced the chip but have keep alot of details out of the public eye. My thinking is that this chip is more specialized than we have been shown so far.
533 MB/sec is pretty impressive! On the ohter hand PCI in all of its recent implementations can be saturated. So one has to wonder if this is enough of an improvement. The image of Motorola being repalaced by a ocmbination of Apple, AMD and IBM is very strange but on the other hand very welcomed.
Dave
IBM it seems has not come real clean on the I/O bus implementation for the 970. Or to put it another way I haven't read in depth on it yet. What is the possibility that the 970 has opentransport ports built into the chip itself? It seems like all that IBM has done is introduced the chip but have keep alot of details out of the public eye. My thinking is that this chip is more specialized than we have been shown so far.
533 MB/sec is pretty impressive! On the ohter hand PCI in all of its recent implementations can be saturated. So one has to wonder if this is enough of an improvement. The image of Motorola being repalaced by a ocmbination of Apple, AMD and IBM is very strange but on the other hand very welcomed.
Dave
I suspect it won't be quite that miraculous. Ars Technica indicates the 970 is no better and maybe worse than the P4 at integer math. I would bet that a 2 Ghz 970 will wind up being about as fast as a 3 Ghz P4, since it is probably about 1.25 times as fast as a G4 at the same clock speed, and the G4 is generally considered to be about 1.5 times as fast as the P4 at similar clock speeds (The G4 would be a great processor if it could only clock up as high as the P4).
But even if a 2 Ghz 970 is "only" as fast as a 3 Ghz P4, it's very likely to clock up much more than 2 Ghz as time goes on, whereas the P4 may have trouble getting much faster than 3 Ghz and so hopefully as time goes on the Mac will again be a real competitor in all areas. It'll need to be because sooner or later Longhorn will show up and it will probably be a pretty good operating system, and again there will be fewer reasons to use a Mac.
Processor Speed
While many people keep pointing to the MHz gap beteen the Mac and Wintel machines, people forget that there is a similar MHz gap between Intel and AMD processors. AMD's new Athlon XP processors (3000+ and 3200+) run in the very low 2GHz range and compete very well with Pentium IVs running in the low 3 GHz range. There is no reason that an IBM chip running a more efficient instruction set shouldn't be able to hold its own at similar speeds to the AMD chips.
Also, although many of the technologies that were posted for the G5 are available for the Wintel market, they are not found in consumer products. The picture from the Apple site lists the G5 as a consumer product. Most of the technologies (such as S-ATA) are only found in professional (high-end) workstations or servers in the Wintel market. While Apple may not be the first to integrate the technology into a computer, they are definitely a leader in introducing the technology into consumer (and prosumer) computers. Anyway, price a Wintel computer with all that technology built into it, and you will be paying just as much if not more than the cost of a PowerMac. Don't compare the price of a PowerMac to an Emachines or HP box. And Dell prices scale up quickly into the $3000+ range.
While many people keep pointing to the MHz gap beteen the Mac and Wintel machines, people forget that there is a similar MHz gap between Intel and AMD processors. AMD's new Athlon XP processors (3000+ and 3200+) run in the very low 2GHz range and compete very well with Pentium IVs running in the low 3 GHz range. There is no reason that an IBM chip running a more efficient instruction set shouldn't be able to hold its own at similar speeds to the AMD chips.
Also, although many of the technologies that were posted for the G5 are available for the Wintel market, they are not found in consumer products. The picture from the Apple site lists the G5 as a consumer product. Most of the technologies (such as S-ATA) are only found in professional (high-end) workstations or servers in the Wintel market. While Apple may not be the first to integrate the technology into a computer, they are definitely a leader in introducing the technology into consumer (and prosumer) computers. Anyway, price a Wintel computer with all that technology built into it, and you will be paying just as much if not more than the cost of a PowerMac. Don't compare the price of a PowerMac to an Emachines or HP box. And Dell prices scale up quickly into the $3000+ range.
Apple isn't going to put all their chips in one basket.
Motorola is expected to start manufacturing the G4 on a smaller process size in the third quarter of this year. Putting these into all the consumer Mac models would bump up the speeds by as much as 80% (7557 G4 tops out at 1.8GHz, according to an internal Motorola document).
It's unlikely that Apple would have IBM processors in more than 50% of the Mac computers in the near future. The 970 chips will almost certainly be restricted to the pro models to differentiate them from the consumer Macs and releasing the topend G4 for the consumer Macs will boost the low to midrange priced Macs performance considerably.
Motorola is expected to start manufacturing the G4 on a smaller process size in the third quarter of this year. Putting these into all the consumer Mac models would bump up the speeds by as much as 80% (7557 G4 tops out at 1.8GHz, according to an internal Motorola document).
It's unlikely that Apple would have IBM processors in more than 50% of the Mac computers in the near future. The 970 chips will almost certainly be restricted to the pro models to differentiate them from the consumer Macs and releasing the topend G4 for the consumer Macs will boost the low to midrange priced Macs performance considerably.
Re: Processor Speed
Don't be shocked when you find out that Apple has a PowerMac that is upwards of $5,000 with these high performance features. The listing of the availability of pro model graphics cards should give that away. Afterall, pro level graphics cards can cost $1,000+. Apple is not gearing the PowerMac for consumers. There is bound to be consumer Mac models announced very soon with faster G4 chips to compete against the low to mid priced Wintel based computers.
Don't be shocked when you find out that Apple has a PowerMac that is upwards of $5,000 with these high performance features. The listing of the availability of pro model graphics cards should give that away. Afterall, pro level graphics cards can cost $1,000+. Apple is not gearing the PowerMac for consumers. There is bound to be consumer Mac models announced very soon with faster G4 chips to compete against the low to mid priced Wintel based computers.
The Opterion may be marketed at workstations but that doesn't mean that motherobard manufacture aren't coming out with workstation boards. The reality is that the opterion does make a good work station chip nad the Athlon64 is apparently going to be marketed directly at tthe workstation market.
I respectfully disagree; the Opterion and Athlon64 are contemporaies to the 970 and will be competeing agianst the 970. It is silly to compare a 970 machine against 6 month old PC technology. To be fair we have to compare agianst what is coming on the market with it.
Not garbage at all, the 970 is apparently Apples future processor direction. To compare it against what has been produced in the PC world in the past is silly. Most institutions I know buy based on price. When they can justify an Apple they can justify the price of an higher priced PC. It doesn't matter overall, all you need to have is an equivalent solution is the PC world, they exist or are coming on line. AMD's new processors can not be dismissed, especially with all of the developement around them. I still maintian that Apple and IBM will have to work very hard to keep any performance advantage that they may have at introduction. Further if AMD feels threatened by these systems they can bring the Athlon64 on line pretty darn quick.
I wasn't attempting to get into a pissing match as you call it, I'm just trying to temper some wildly outrageous claims. Frankly my system is so old that it is pathetic, the informaiton so far is that the new systems from Apple will be a good replacement. The key is pricing, lets hope that Apple doesn't use a marginal performance advantage as justification for excessive pricing.
Some of the things you describe above are fantastic. Unfortunately we will have to wait beyond rev 2 of the 970 to see them.
You have not been able to find mother boards with serial ata, or Hypertransport???
Dave
LOL, how old is your PC?
My nForce2 based motherboard (with SATA and Hypertransport on board for those of you who think it is some amazingly new technology) allows you to adjust the settings in the BIOS, yes. But it also will run the memory and the processor and everything at the speeds that they want to run at.
DIMMs include a small ROM called the SPD ROM, that the BIOS reads to configure the memory speeds. The processor will do something similar. AGP is the same, it runs at the best speed for the card/controller out of the box.
Sheesh!
A few disagreements here. First if you are comparing this possible 970 introduction to a $400 dollar consumer PC machine of course not those are yesterdays technolgies. Shop around with the same amount of money that you will spend for that Apple and then you will see a differrent situation.
Besides I'm willing to bet that Apple and most people consider the PowerMac to be a professional system. We have seen nothing yet to indicate that Apple will introduce the 970 in their lower end machines (not htat I'm not hoping this). Maybe in the end we are trying to say the same thing, that comparable systems will exist. I just have trouble with the PowerMac being classified as a consumer model. I don't think anybody true believes that that is its target market.
Once one looks at pro level pc's and the supposed 970, the performance advantage doesn't look that fantastic.. It looks like Apple will have parity or a slight advantage at launch time that is all. If Apple can double performance by the end of the year then we will have something to crow about.
DAVE
Besides I'm willing to bet that Apple and most people consider the PowerMac to be a professional system. We have seen nothing yet to indicate that Apple will introduce the 970 in their lower end machines (not htat I'm not hoping this). Maybe in the end we are trying to say the same thing, that comparable systems will exist. I just have trouble with the PowerMac being classified as a consumer model. I don't think anybody true believes that that is its target market.
Once one looks at pro level pc's and the supposed 970, the performance advantage doesn't look that fantastic.. It looks like Apple will have parity or a slight advantage at launch time that is all. If Apple can double performance by the end of the year then we will have something to crow about.
DAVE
Ahh, now that's more like it. This makes much more sense. Hopefully the Mac mobos will do the same thing. Since you can't get PC4000 RAM yet, I am sure then that Apple will be putting PC3700 into the 1 gig bus machines. PC3200...nah, why cripple them even further? Especially with the exorbitant prices Apple charges for RAM! I just pricewatched both 3200 and 3700, and there's like a $30 difference between them at the 512mb level. Not worth the crippling.
Not all bus accesses involve memory. When doing I/O the majority of bus activity may involve RAM <-> device communication, but there is also direct CPU <-> device communication, like "Ok, the data's ready in RAM for you" when reading and writing. Also, depending on how the SMP is done, the multiple CPUs might have to talk to each other to keep their caches coherent, which involes CPU <-> CPU communication, again, not involving RAM.
Bus speeds change less frequently than RAM speeds do, so it makes sense to leave a little head-room, so that more engineering effort is not needed 6 months from now.
Plus, I think (don't know) that the bus speed will be 800MHz for the 1600MHz chip, 900MHz for the 1800MHz chip, and 1000MHz for the 2000MHz chip.
Re: Apple isn't going to put all their chips in one basket.
If the 970 is cheeper to make and faster why would apple not use it in all there machines?
If the 970 is cheeper to make and faster why would apple not use it in all there machines?
Re: Re: Apple isn't going to put all their chips in one basket.
Product differentiation. I would not be surprised to hear about a major upgrade to the G3 line Monday.
See The Register Article for some additional rumors.
Product differentiation. I would not be surprised to hear about a major upgrade to the G3 line Monday.
See The Register Article for some additional rumors.