This must be embarassing. After Nemetschek posted benchmarks showing that the Vectorworks on the G5 outperforms Vectoworks on the Intel Xeon, the company is being required by Apple to switch over to Intel CPUs.
The benchmark tested:
2D Redraws -- general 2D CAD drawing performance.
Renderings -- calculation times.
Solids Modeling -- modeling operations.
Architosh asks the question, "What's going on to make this [G5] machine so quick?" The Nemetschek CTO admits he doesn't know why. Now it no longer matters; the answer will be assigned to the dustbin of history.
You can read the details on the benchmark at Architosh, and then read reaction to the Apple-Intel love-in next week by subscribing to upFront.eZine.
Well, aside from misspelling my publication, I'm delighted that a PC-centric CAD publication would pick up this issue. Though, admittedly, I'm dissappointed in the negative spin. Sure, the PowerPC G5 is not only a true 64-bit chip with backward 32-bit compatibility (32 registers are simply left empty) but is one terribly fast processor, especially as applied to floating-point and vector math. Notwithstanding this performance, Apple is now claiming that the PowerPC road map is devoid of the progress they are looking for in future hardware designs. Thus Intel.
That claim is partially false. The truth is more likely that, while Apple pays as little as $65 per chip for PowerPC parts, developing new designs takes hundreds of millions. And Apple simply doesn't want to pay any part of that.
Steve has had many choices. Sony wanted Apple to join them in the use and expansion of the IBM Cell chip. But reportedly, Apple engineers don't like the Cell chip, and Steve told Sony no.
So what does it take to impress Steve Jobs these days? And why did Bill Gates not get impressed with Intel's future road map (that same map Steve says is so much better)?
The answer must be more complex than we are allowed to discover at this point in time. The importance of keeping it a secret is worth (+$$) more than the shock (-$$) and disturbance that secret will entail and the Mac platform.
As for near future buying decisions? Buy speedy PowerPC Macs. A dual-core version of the current G5 would be a killer! And for support? Apple has a very good history of supporting legacy stuff, just look at "Classic" inside OS X today. It works very well.
Posted by: Anthony Frausto-Robledo | Jun 13, 2005 at 02:49 PM
Yeah, I somehow think that these Intel chips will be a step backward, too. From a layman's point of view, there's something about Intel which smells CHEAP. Crashes, General Protection faults. Apple's jes cuttin' corners ta make cheaper computers. Their products are gonna get worse, not better. Jobs has sold out--but this rush to cut costs will lower the quality of Apple computers, not make them better. I'm gonna be one of the last people ta buy the G5 cause it'll probably be the last great computer that Apple'll ever make. I mean, the G5's a good chip. It was designed as a 64 bit chip from the ground up, along with the Motherboard which accompanies them. It seems that Intel ADDS on the 64 bit capability after the fact. They have too many legacy problems. If the PC's I'm usin' are any indication, I'll go with the G5 just before they get phased out.
Posted by: Joe | Oct 31, 2005 at 09:51 AM