Originally posted by JeffHendr
Parallel processing, once again, is related to optimizing programs for multiple threads by forking off different calculations to be done at the same time. It is a characteristic of programming, not architectures. OS X does not handle threads any different than Windows. OS X does not magically make a program multi threaded. Windows does not magically make a program multi threaded.
So is all this stuff on Apple's website about the velocity engine a bunch of garbage?
Do Macs make as much of a mess of the hard drive as PC's do concerning multiple copies of files and file fragments? Does OS X grow infinitely bigger the way that Windows does? Does OS X significantly slow down the way Windows does after a couple of months of usage?
In any case, thanks for the info. I'm still learning alot about Macs. I just got my first Mac-Powerbook 1 GHz with Superdrive. I thought all the info on the Apple website about the G4 processor was pretty cool, but maybe it's all smoke and lights to keep Apple users happy while their company is in trouble.
I hate to break it to you, but there's this think called marketing. They tell lies by twisting the truth in a complicated way to make people believe things that aren't necessarily true.
When apple claims to have 266MBs of bandwidth for PCI devices, they dont' tell you that is only for 64 bit pci devices. Most PCI devices, save high end scsi cards, are 32 bits, meaning they go at the same 133MBs speed that they claim pcs can only do. Secondly, the integration of a north bridge and south bridge chips into the system controller does not give an advantage- it still does the same work of the two, it still has the same elements. Not only that, but the two bridges in PCs are connected by a very fast bus, 266 or 533 MBs, and it isn't a bottleneck like Apple claims. In the old days (97-98) this wasn't so, but it has been since around 2000. If you want proof, Intel calls it part of their hub archietcutre, Via calls it the V-Link, and AMD calls is hypertransport, and it's also called a backside bus.
The velocity engine uses the same concept as the entire CISC instruction set does- work on mutiple data sets in the same instruction. THat is why people hated the CSIC instructino set- they were too complicated! But, what Apple has said was a disadvantage in the past is now a benefit. That's marketing.
"Jaguar includes enhanced preemptive multitasking, symmetric multiprocessing and multithreading capabilities..."
Hello? These have been in the NT kernel since it's inception with NT 3.1. These are basic computing concepts that Apple has marketed because they are fancy names. There are not unique to Apple; Apple is the only one that advertises them because they sound good. That's marketing. Look at Microsoft's marketing- see those same buzz words?
http://www.microsoft.com/ntworkstat...ctOverview/FeatureList/Windows95/techdiff.asp Of course you do, they're there to sell products, NOT be a technical comparison of two technologies.
I know this may be hard to believe, but a standard means that it's the same (supposedly) across different things. PCI bus is standard. Memory controllers operate on standards. The processing of 1's and 0's are standard. Just because they use a different instruction format does not mean that PowerPC and X86 are completely two different beasts. At the heart they still have the same ALU's, the same shift units, the same floating point units(even if different sizes). What is different is how they are implemented (using a carry & look ahead adder, putting in multiple decoders to have 3 different pipelines) and the physical electronics underneath. Apple can go and put on their website how their floating point conforms to IEEE standards, that does not make it unique.