Another new toy!

[Mark Smith]

Not C= related but still fun.

Just replaced my PowerPC based Mac Mini with a Core 2 Duo based Mac Mini ... and it is sooooooooo much faster!

"Backing up" a DVD with H.264 codec used to take I think 20 hours (overnight and through the rest of the work day), ... now takes 1 hour .. hmmmmmm, I didn;t realise there was so much space between the old 1.42Ghz G4 and a 2Ghz Core 2 Duo :-)

Oh and Parallels is fantastic ... probably not a good thing for an EMC contractor to say really :-)

Mark

[Guest]

The rest of the world wanted so badly to believe that PowerPC was better than Intel, but whenever the reality of truth is shown on a subject, popular anti-establishment is rarely as beautiful as expected.

[Golan Klinger]

Define "better".

[Guest]

Name it.  I've heard RISC zealots claim everything from faster, cheaper, cooler.  None of it has proven to be true.  HP, one of the biggest supports of RISC with their PA-RISC architecture realized in the early 90's that RISC is a fool's errand given Moore's law since more and more of the CISC CPUs are implemented in logic circuits instead of microcode in each successive generation of CPU.  Now, HP is the driving force behind the Itanium processor, which is not only a CISC design, but so efficient that it only needs half the clock speed to match the performance of Power and Sparc processors of the same generation.

[Mark Smith]

While the Pentium 4 was Intels main chip to desktop users I'd have always said the PowerPC was better, but since the big switch to the Core line of processors things turned around ... very impressed Intel ditched the old model in favour of the new, seems they realised you can;t just keep upping the hertz and have to be more elegant in their designs :-)

[Guest]

While the Pentium 4 was Intels main chip to desktop users I'd have always said the PowerPC was better, but since the big switch to the Core line of processors things turned around ... very impressed Intel ditched the old model in favour of the new, seems they realised you can;t just keep upping the hertz and have to be more elegant in their designs :-)

But what were you basing that comparison on between P4 and PowerPC?  Companies like Adobe went out of their way to make the Windows and Mac OS products so different that comparable tests were impossible.  Even now, with the Intel processors, the Mac version of Adobe Photoshop is more than twice as fast as the Windows version.  We've all known for a long time that Adobe is anti-Microsoft and was a huge backer of Apple, but now their secret is out and the reality has been all along that Intel has always been better than PowerPC.

[Mark Smith]

My comparison was on "user experience" .. alsways found MP3 ripping to be much faster on PowerPC machines, most things number crunching always were faster on the PowerPC.
Example being my old amiga with CyberstormPPC, it had a 233Mhz 604e, it would steam through RC5 key crunching ... rate was about twice of my work PC .. P-III 800 something or other.

I think these days with Universal binaries in OSX I think developers are tuning more for Intel now rather than PPC so maybe that is why it seems leagues faster ... still don;t care, the "user experience" is what counts for me :-)

I think a good test would be to get a task result from something being done on a G5 PowerMac, and then running the same benchmark on an Intel Mac but force the intel to use Rosetta and emulate the PowerPC ... wonder how close they'd be ?

[Guest]

Again, you're making my point about code not being optimized for a platform.  You can write code to do the exact same thing in Assembler and Ruby, but your performance in the two will be radically different on the same machine.  This holds true even using the same language.  You can write code to do an encryption algorithm in C, but compile it without MMX extensions (which was common in the days of Pentium-3 because AMD had just been able to legally implement MMX) and compare that with code that uses MMX extensions and see a huge performance difference.  What you're describing has nothing to do with the CPU's involved, but with the programming involved.  The big difference today, especially for Apple, is that MMX and SSE instructions have been around long enough in the Intel processors that ALL new compilers default to using these things and thus all code built with those compilers have these features enabled by default.  Since Apple never touched an Intel processor that didn't have at least SSE3 built in, it makes sense that their Intel stuff kicks the dog-snot out of G5.