G3 and G4 Compatibility

[Galen_of_Edgewood]

Hey all,

Does anyone know if the G4 and the G3 is compatible? I really need to know for sure, as in 99-100% positive. If you know of where this is documented, could you post me a link?

Thanks in advance.

Oh yeah, I don't give a flip that it's a Mac chip, that is irrelavent for what I'm considering. Let's not start a flame war, please....

------------------
"I don't think so", said Rene Descart. Just then, he disappeared.

[zombor]

the g4 is just a souped up version of the g3 chip. There are prolly extra instructions i nthe g4(just like from p2 to p3), but architecturlly, the are the same chip, just alot faster. what are you doing that it needs to be compatible

------------------
Compaq Armada E500
650MHz PIII 128MB RAM
8MB ATI rage mobility
11GB hard drive/DVD
dual booting win98/2k

who said you cant game on a notebook????

[Galen_of_Edgewood]

Originally posted by zombor:
the g4 is just a souped up version of the g3 chip. There are prolly extra instructions i nthe g4(just like from p2 to p3), but architecturlly, the are the same chip, just alot faster. what are you doing that it needs to be compatible

Try real-time embedded stuff. (I'm not kiddin' either!)

I'm not doing it personally, but I have been asked to research finding a simulation computer and I know that the G3 is compatible with what PowerPC model that we're using. I couldn't find anything easily on the Motorola website, so I thought that this would be dramatically faster. Guess what, it is.

That and I was curious myself (I'm an x86 junkie, though I do know that the PowerPC chips perform increadble, considering their clock speed). Good opportunity to learn, I say.

------------------
"I don't think so", said Rene Descart. Just then, he disappeared.

[PotatoChipMan]

G4 has an alti-vec processing unit. For alti-vec optimized applications this can dramitcally improve performance. In any non-altivec app the performance gain is neglable. Fortunatly most apps are alti-vec optimized. Same thing for the MP systems. MP's won't shine until OSX final is here.

PCM

[Arcadian]

Originally posted by PotatoChipMan:
G4 has an alti-vec processing unit. For alti-vec optimized applications this can dramitcally improve performance. In any non-altivec app the performance gain is neglable. Fortunatly most apps are alti-vec optimized. Same thing for the MP systems. MP's won't shine until OSX final is here.
PCM

Can you explain alti-vec? I haven't heard of it until now.

[zombor]

Originally posted by Galen_of_Edgewood:
Try real-time embedded stuff. (I'm not kiddin' either!)

I'm not doing it personally, but I have been asked to research finding a simulation computer and I know that the G3 is compatible with what PowerPC model that we're using. I couldn't find anything easily on the Motorola website, so I thought that this would be dramatically faster. Guess what, it is.

That and I was curious myself (I'm an x86 junkie, though I do know that the PowerPC chips perform increadble, considering their clock speed). Good opportunity to learn, I say.

the only reason PowerPC is faster at a lower clock pseed is because of it's lower pipeline. A higher pipe enables the clock to go up, but take more time for the instruction to go thru the pipe, therefore, negqating the higher clock speed.

As for alti-vec, it is kinda like apples SEE2 and MMX. But the OS can utilize these instructions to its full potential. I might be a little off because i havent read about it since it came out.


------------------
Compaq Armada E500
650MHz PIII 128MB RAM
8MB ATI rage mobility
11GB hard drive/DVD
dual booting win98/2k

who said you cant game on a notebook????

[johnsonfromwisconsin]

Basically, MMx, SSE, 3dnow!, and Altivec (also known as the Velocity Engine) operate under the same principal, although Altivec is a superior implementation.

All four of them are structures known as SIMD, which stands for Single Input Multiple Data.

Traditionally, computers have been SISD (Single Input Single Data) type machines.

How do they differ?

In a traditional machine, you added (for example) one number to another to get one result. With SIMD, you can add one number to 2 or three or more numbers at the same time. For instance, consider the following additions a computer is about to do:

i + j

i + k

i + l

noticce that they all consist of additions of "i". A traditional computer would have to compute these 3 additions sequentially, one by one, but with SIMD, they can be computed all at once:

i + <j, k, l>

I hope that helped someone

------------------
Remember, there can be only One!

[Arcadian]

Thanks for the info, johnsonfromwisconsin.

But, can you explain the differences between these extensions as well? I am unfamiliar with Altivec, so I was wondering how it was different from Intel's or AMD's implementations.

[johnsonfromwisconsin]

THe best article I've seen on the different SIMD sets is here ( http://arstechnica.com/cpu/1q00/simd/simd-1.html )

Note, I made a common mistake on the SIMD acronym, it actually stands for Single Instruction Multiple Data. I'm always messing that up

------------------
Remember, there can be only One!

[Duo]

100% compatibility.

Duo

------------------
I am King Duo, King of Cat Fish