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

No Itanium = No G5 ???

How abou this: What if Apple also gets AMD into the PowerPC consortium? That would give apple another supplier that actually KNOWS how to compete. I mean, it's not like the athlon is the only chip that AMD produces, they also produce MIPS chips. If they can get over the hurdle of licensing the PowerPC to AMD, we'll really see some progress!
 
AMD...

Originally posted by topicolo
How abou this: What if Apple also gets AMD into the PowerPC consortium? That would give apple another supplier that actually KNOWS how to compete. I mean, it's not like the athlon is the only chip that AMD produces, they also produce MIPS chips. If they can get over the hurdle of licensing the PowerPC to AMD, we'll really see some progress!
Click to expand...

...does not have the resources to compete against Intel with two ISAs in the same market (desktop computers). Heck, they can't even make two separate chips to compete against Intel (the Hammer architecture is going to be a Server/Desktop/Mobile chip). Also, Motorola knows how to compete just fine, which is why it doesn't make good chips for Apple (that's debatable, the G4 is a pretty nice chip). Confused? Motorola is not a desktop chip maker. They happen to make a high end embedded chip that can be used in desktops, but it isn't designed for that. Also, all of AMD's engineers are used to squeezing performance out of x86, if they switched to PPC the result would be LESS performance, not more.

So, in summary, making AMD make PPC desktop chips would:

1) hurt AMD by wasting R&D on a small market
2) hurt Apple by giving them "first try" chips, instead of refined ones
3) hurt the consumers by giving them "first try" chips.

Now if we could get Transmeta to make chips for Apple, that would be a different story...
 
Re: AMD...

Originally posted by Catfish_Man


...does not have the resources to compete against Intel with two ISAs in the same market (desktop computers). Heck, they can't even make two separate chips to compete against Intel (the Hammer architecture is going to be a Server/Desktop/Mobile chip). Also, Motorola knows how to compete just fine, which is why it doesn't make good chips for Apple (that's debatable, the G4 is a pretty nice chip). Confused? Motorola is not a desktop chip maker. They happen to make a high end embedded chip that can be used in desktops, but it isn't designed for that. Also, all of AMD's engineers are used to squeezing performance out of x86, if they switched to PPC the result would be LESS performance, not more.

So, in summary, making AMD make PPC desktop chips would:

1) hurt AMD by wasting R&D on a small market
2) hurt Apple by giving them "first try" chips, instead of refined ones
3) hurt the consumers by giving them "first try" chips.

Now if we could get Transmeta to make chips for Apple, that would be a different story...
Click to expand...

First, AMD doesn't have to do the R&D in competition with Moto, they could set up a joint system with them like IBM did back in the day.
Second, AMD can use the excess capacity they have now to make PPCs and give apple a second supplier.
Third, how would Transmeta be a better candidate than AMD? they don't have their own foundries, they are starting to get killed by Intel on the low-power mobile chips front relegating them only to the low-powered server and tablet-pc niches.

just my 2cents
 
Re: Re: AMD...

Originally posted by topicolo


First, AMD doesn't have to do the R&D in competition with Moto, they could set up a joint system with them like IBM did back in the day.
Second, AMD can use the excess capacity they have now to make PPCs and give apple a second supplier.
Third, how would Transmeta be a better candidate than AMD? they don't have their own foundries, they are starting to get killed by Intel on the low-power mobile chips front relegating them only to the low-powered server and tablet-pc niches.

just my 2cents
Click to expand...

Transmeta's code morphing software could be modified to allow x86 AND PPC compatibility. This would allow for a Windows emulation layer similar to classic. No more incompatibility problems. Also, Transmeta's chips are much more sophisticated than either x86 or PPC, given the funding and fabs (IBM?) they could make a fabulous high performance chip. Remember, currently they're on .18 micron without SOI, and they're still doing pretty well, with a chip designed for performance instead of power efficiency, .13 micron, and SOI, they'd have a really great chip.
 
Re: Re: Re: AMD...

Originally posted by Catfish_Man


Transmeta's code morphing software could be modified to allow x86 AND PPC compatibility. This would allow for a Windows emulation layer similar to classic. No more incompatibility problems. Also, Transmeta's chips are much more sophisticated than either x86 or PPC, given the funding and fabs (IBM?) they could make a fabulous high performance chip. Remember, currently they're on .18 micron without SOI, and they're still doing pretty well, with a chip designed for performance instead of power efficiency, .13 micron, and SOI, they'd have a really great chip.
Click to expand...

The problem is, the code morphing technology is EXACTLY what is hindering performance. Also, transmeta chips have really sucky floating point performance, which is pretty huge drawback if you want to run 3d games/apps and transmeta's at .18 micron because they can't control the foundries (which they don't own) nearly as well. Look at what happened to cyrix and you'll see one very likely possibility for transmeta in the future
 
I sure as hell hope not. Even Intel would've moved on to a 64bit chip by then. Granted, a 64bit chip doesn't necessarily mean that it'll be faster, but it does mean that it'll be able to address more memory and such. I think Apple will have to move to a 64bit chip before the 7 years is up just so that they can support more memory on their macs.

Speaking of 64bit chips, preliminary benchmarks of the new AMD clawhammer chip has just been released. This 800Mhz chip can beat a P4 1.6Ghz in Quake III!!! Could this be a sign of how well the G4 will do?
AMD Clawhammer Benchmarks
 
Re: Re: Re: Re: AMD...

Originally posted by topicolo


The problem is, the code morphing technology is EXACTLY what is hindering performance. Also, transmeta chips have really sucky floating point performance, which is pretty huge drawback if you want to run 3d games/apps and transmeta's at .18 micron because they can't control the foundries (which they don't own) nearly as well. Look at what happened to cyrix and you'll see one very likely possibility for transmeta in the future
Click to expand...

The reason why Transmeta's performance sucks is because they don't care about it. The Crusoe has as many features as possible stripped off to save on space/power. With the chip foundry problem, that's why I suggested teaming up with IBM. Read arstechnica's article on the Crusoe, it explains why I think it has potential quite well.
 
Now that I think of it, is there REALLY anything different between a pc mobo and a mac mobo? Aside from the obvious differences in processor sockets and northbridge/southbridge chipsets, I can't really think of anything.

If the only things that really need to be redesigned are the chipsets, it doesn't make sense that apple is so far behind. I still think that Apple should be making more alliances with industry giants like VIA and AMD, if not to produce faster CPUS, then to update their mobo technologies.
 
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