DSPRelated.com
Forums

[OT]Apple to embrace Intel chips -- Vector Processing??

Started by Himanshu June 12, 2005
Hello!

I read somewhere on the net that Apple is moving towards embracing
Intel chips. What about PowerPC's Vector processing power they boasted
about earlier? Isn't PowerPC powerful than its contemporary chips?
Apple provices benchmarks for its Altivec extensions. The benchmark
uses FFT performance and convolution. Weren't they _Serious_ about
this?

Any comments?

Regards
--Himanshu

Himanshu wrote:
> > Hello! > > I read somewhere on the net that Apple is moving towards embracing > Intel chips. What about PowerPC's Vector processing power they boasted > about earlier?
Its called Altivec and its very similar to the SSE instruction set of the Petium 4 and current AMD CPUs.
> Isn't PowerPC powerful than its contemporary chips?
Maybe in specially chosen micro benchmarks, but when running real code a PowerPC CPU of X MHz runs code at roughtly the same speed as an X MHz Pentium CPU.
> Apple provices benchmarks for its Altivec extensions.
Altivec can be faster than non-Altivec code. I have never seen any benchmarks that compare Altivec with Pentium SSE. A lot of what Apple was doing was benchmarketing rather than benchmarking. Erik -- +-----------------------------------------------------------+ Erik de Castro Lopo nospam@mega-nerd.com (Yes it's valid) +-----------------------------------------------------------+ "I invented the term Object-Oriented, and I can tell you I did not have C++ in mind." -- Alan Kay
Himanshu wrote:
> Hello! > > I read somewhere on the net that Apple is moving towards embracing > Intel chips. What about PowerPC's Vector processing power they boasted > about earlier? Isn't PowerPC powerful than its contemporary chips? > Apple provices benchmarks for its Altivec extensions. The benchmark > uses FFT performance and convolution. Weren't they _Serious_ about > this? > > Any comments? > > Regards > --Himanshu >
Apple couldn't get a G5 suitable for laptops. Laptops are a "premium" product, i.e. people pay more for them.
But does that in any way (if we don't compare it with anything else)
mean that Altivec extension is not really faster? I think it is
__faster__ than the normal mode (mode?? they say there is no switch
between the modes as is need in pentiums). Altivec extention does take
less time in dreaded Multiply + accumulate processes in loop. Aint it ?

Thanks and regards
--Himanshu

in article 1118684027.973586.240760@g43g2000cwa.googlegroups.com, Himanshu
at hs.chauhan@gmail.com wrote on 06/13/2005 13:33:

> But does that in any way (if we don't compare it with anything else) > mean that Altivec extension is not really faster? I think it is > __faster__ than the normal mode (mode?? they say there is no switch > between the modes as is need in pentiums). Altivec extention does take > less time in dreaded Multiply + accumulate processes in loop. Aint it ?
it will. it sounds to me that entire applications will need to be at least rebuilt if not partly rewritten. i was sorta shocked by the announcement. looks like the victory of the Microsoft/Intel duopoly is nearly complete. next year, we will all have 0110 0110 0110 tattooed on our foreheads. -- r b-j rbj@audioimagination.com "Imagination is more important than knowledge."
robert bristow-johnson wrote:
> looks like the victory of the Microsoft/Intel duopoly is nearly complete.
Or not. Robert Cringely has an interesting take on the Apple/Intel announcement: http://www.pbs.org/cringely/pulpit/pulpit20050609.html -- Jim Thomas Principal Applications Engineer Bittware, Inc jthomas@bittware.com http://www.bittware.com (603) 226-0404 x536 A pessimist is an optimist with experience
Jim Thomas wrote:
> > robert bristow-johnson wrote: > > looks like the victory of the Microsoft/Intel duopoly is nearly complete. > > Or not. Robert Cringely has an interesting take on the Apple/Intel > announcement: > > http://www.pbs.org/cringely/pulpit/pulpit20050609.html
I realy don't agree with Cringely on this one. Intel cannot afford to annoy its other customers in the PC industry because if they do, their customers will just switch to AMD. I'm also pretty sure that a single vendor like Dell or HP is a significantly bigger customer than Apple will be at any time in the near future. If Inetl simply cannot afford to show favouritism to any one of its customers including Apple. Erik -- +-----------------------------------------------------------+ Erik de Castro Lopo nospam@mega-nerd.com (Yes it's valid) +-----------------------------------------------------------+ "C++ is an atrocity, the bletcherous scab of the computing world, responsible for more buffer overflows, more security breaches, more blue screens of death, more mysterious failures than any other computer language in the history of the planet Earth." -- Eric Lee Green
Erik de Castro Lopo wrote:
> Jim Thomas wrote: > >>robert bristow-johnson wrote: >> >>>looks like the victory of the Microsoft/Intel duopoly is nearly complete. >> >>Or not. Robert Cringely has an interesting take on the Apple/Intel >>announcement: >> >>http://www.pbs.org/cringely/pulpit/pulpit20050609.html > > > I realy don't agree with Cringely on this one. > > Intel cannot afford to annoy its other customers in the PC industry > because if they do, their customers will just switch to AMD. I'm also > pretty sure that a single vendor like Dell or HP is a significantly > bigger customer than Apple will be at any time in the near future. > > If Inetl simply cannot afford to show favouritism to any one of its > customers including Apple.
Strictly, Microsoft isn't an Intel customer. I doubt that Microsoft can write its OS to run on AMD but not Intel. Intel might be able to bundle an OS with its chips. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
It comes down to the fact that the operating system, word processing, 
TCP/IP stacks, and drivers do not benefit with Altivec.  Graphics, 
sound, and modem (dialup/DSL) are handled by external hardware so no 
benefit there.  Apple's mainstay of audio/video production, a small 
percentage of Apple's pie, will transition to Intel/AMD easily since 
Digidesign ProTools audio app and many of the popular video apps now run 
on both platforms (Apple/Wintel).  IBM does not seem too upset, nor will 
they be that hurt by the switch.

Dan Ash

Himanshu wrote:
> But does that in any way (if we don't compare it with anything else) > mean that Altivec extension is not really faster? I think it is > __faster__ than the normal mode (mode?? they say there is no switch > between the modes as is need in pentiums). Altivec extention does take > less time in dreaded Multiply + accumulate processes in loop. Aint it ? > > Thanks and regards > --Himanshu >
On Tue, 14 Jun 2005 07:10:31 +1000, Erik de Castro Lopo
<nospam@mega-nerd.com> wrote:

>Jim Thomas wrote: >> >> robert bristow-johnson wrote: >> > looks like the victory of the Microsoft/Intel duopoly is nearly complete. >> >> Or not. Robert Cringely has an interesting take on the Apple/Intel >> announcement: >> >> http://www.pbs.org/cringely/pulpit/pulpit20050609.html > >I realy don't agree with Cringely on this one.
Ever since Cringely freaked out and took a chainsaw to his rear-drive airplane project I've wondered about the guy. He seems to live right on the border of lucidity with occassional excursions into the other side. Bright guy, just a little wobbly at times. I think he really missed the mark on this one as well.
>Intel cannot afford to annoy its other customers in the PC industry >because if they do, their customers will just switch to AMD. I'm also >pretty sure that a single vendor like Dell or HP is a significantly >bigger customer than Apple will be at any time in the near future. > >If Inetl simply cannot afford to show favouritism to any one of its >customers including Apple.
There's a lot of truth to that. The World's Most Paranoid Company (Andy Grove said it first, not me) goes way out of its way to avoid anything even resembling anti-trust or a lack of integrity in business practices. I know it doesn't seem that way on the outside sometimes, but from the inside I'm always a bit surprised at how big a deal it is to maintain integrity in relationships. I have to say that I'm not even close to being on the inside of this thing with Apple, but my understanding is that much of the motivation is that PowerPC has no substantial path for power reduction for mobile platforms (i.e., laptops), which is expected to become an increasingly more important segment of Apple's market (just like it is in the rest of the PC market). By most measures Intel has hit a homerun with Centrino, so AMD may not even appear as the best alternative from this perspective. As others have mentioned, I'm hoping that in the long term this provides a natural evolution that generates more competition in the OS market. While it seems unlikely at this point that Apple would ever consider porting their OS to other platforms, having Apple platforms as an additional alternative with closer compatibility to MS-based platforms can only be a good thing for the consumer. At least that's what I'm hoping. Who knows how this will shake out, but let's hope that it enables more capabilities than it hurts. Eric Jacobsen Minister of Algorithms, Intel Corp. My opinions may not be Intel's opinions. http://www.ericjacobsen.org