[X4U] Re: New Pentium Macs (Altivec)

[Philip J Robar]

On Jan 24, 2006, at 10:20 AM, Michael J. Prevost wrote:

>>>   . . .PowerPC could no longer
>>>   compete, regardless of broken promises from Motorola and IBM.
>>>   Things like the Velocity Engine (i.e. AltiVec) were stopgaps.

AltiVec was not a "stopgap", it was a clear step forward in SIMD  
(a.k.a. vector processing) implementations and remains a clearly  
superior solution to Intel's SSE.


>> We do serious number-crunching on the G5s (many jobs run for  
>> days). The
>> Altivec engine has made certain kinds of applications SCREAM such  
>> that a
>> single G5 can cometimes compete with a cluster of dozens of  
>> Winboxes, and
>> "dozens" is an understatement. Of course, that has requireAltivec- 
>> specific
>> programming. We are NOT looking forward to massive re-writes of  
>> our code.
>>
>> Oh, well, the price of wanting to be bleeding edge.
>>
>> Vard Nelson
>> Detect Geophysical

Then why rewrite? As long as the G5 is a superior solution keep using  
it. G5's will be available in the used market and will be supported  
by OS X for the foreseeable future.


> Vard,
>
> I'm glad you posted this message because I've been wondering about  
> applications optimized to use Altivec. Does the Altivec silicon do  
> anything more than accelerate the floating point performance?

AltiVec is not an FP accelerator. AltiVec is an implementation of  
SIMD (single instruction multiple data). In simple terms it  
accelerates digital signal processing and multimedia algorithms by  
applying the same instruction to multiple pieces of data at the same  
time. In other words it is a specialized form of parallel processing.  
Start here for more information: http://developer.apple.com/ 
hardwaredrivers/ve/


> The SPECmark performance for FP that Apple is reporting is much  
> better than the Altivec-equipped G5. So if this line of thinking is  
> correct, will it be necessary for your team need to write a lot of  
> low-level code for the Intel Core Duo? or just use a top-shelf  
> optimizing compiler to take advantage of the chip's basic  
> architecture?
>
> Mike
> Dot4, Inc.

You're confusing floating point math with SIMD, they're not the same  
thing. The FP and AltiVec are different execution units in the G5  
CPU. They do different jobs. The Power and PowerPC family's floating  
point performance was generally (Always?, I'm too lazy to check.)  
better then Intel chips of the same time frame and target market. The  
Core 2 chips are several years newer than the G5 so it's not  
surprising that it has faster floating point performance. The G5's  
SIMD performance is much better than Intel's and will remain so for  
the near future, unless Intel decides to significantly rev the SSE  
architecture.


Phil