Intel/AMD VS. IBM
First, s far as the expected updates are concerned, so long as they update it and get rid of the liquid cooled system, I'll be happy. (I just don't trust liquid cooling, take it as you will)
I've let some crap on these boards go on for long enough without saying anything, and I'm sure by the end of this someone's gonna try and tell me I'm wrong about something or other, but I do ask if you plan on arguing with me when it comes to the true performance of x86 VS. PPC and Power4, at least look into what I've said below, then cite your information so I can verify and comment.
I'll make this easy here, x86 is CRAP compared to Power4 and the PPC subset. You can have the most Ghz in the world, but if the rest of your system is crawling along an endless pipeline where every peice of information is fighting for bandwidth controlled by an OS that doesn't know how to properly handle it (thats why Windows users have to defrag), what the hell kinda speed does that give you. I will give credit where credit is due, AMD does a much better job with the way it handles x86 (Eliminating the Northbridge was a great idea)
Let me explain, the x86 Architecture forces every subsystem's throughput into a single fronside bus which then communicates to the CPU and Main Memory. This has worked fine, I was a Wintel user for 10 years myself before switching. The problem with that is every component of your system (USB, Ethernet, Graphics, PCI Cards, etc) is all fighting for bandwidth, so if your say trying to use multiple programs and access hardware functions (ie CD/DVD Burning) this causes system slowdown and because Windows is written the way it is (I'm not sure the tech side of the software) can cause programs to stall, crash, and freeze the system (remember when you couldn't do anything while your CD's were being burnt, that's because of x86's and Windows limitations) This is even worse due to the fact that the intel/amd CPU's only offer limited instrution sets (OK "1" on any non Hyper-Threading/Hyper-Transport System, those offer a whopping "2", Perhaps more on newer AMD's) There are more limitations I may mention later for now, lemme compare whats above to the current Power4 Architecture. Oh one more little thing before I move on, the maximum throughput of any Intel system is 6.4GBps
Power4 at it's core, allows EVERY subsystem to talk directly to the main memory and the CPU INDEPENDENTLY, NOTHING fights for bandwidth due to whats at the core of the Architecture, the System Controller. Forget Northbridge/Southbridge, The sytem controller regulates data moving through the system, it can direct data where it needs to go directly, making for efficient data throughput. On top of that, add to it, Branch Prediction Logic, with 95% accuracy (roughly), BPI can detect where data needs to go before it reaches the cpu, making dataflow even faster. Also add Enchanced Velocity engines in each CPU, DUAL INDEPENDENT SYSTEM BUSES Maxing out at 1.25GHz (For the moment) TWO CPU's with 2 Double Precision Floating Point Units per (Intel/AMD = 1) and you have a system that's total throughput max's out at 20GBps with much more efficiancy than that of any x86 system on the market today. A 3.8 GHz Pentium or for that matter a Dual Core 3.4GHz Pentium can't even compare when it comes to true RAW POWER, The Dual Core Pentiums still use a single System bus and the same architecute thats been around since 1989, ok so that's that. By the Way, Each G5 has 8 In-Flight Instruction Sets, VS 2 on any x86 system (AMD may have more as cited earlier, I'm not 100% on that)
OK clock speed, Max right now 3.8Ghz VS 2.5Ghz Dual, who wins.....what'd I say earlier, yeah, IBM/Apple wins, hell even on a single 2.5, though not by as considerable a margin. Clock speeds help the overall performance of the system in both respects, but a better system will stand to take better advantage of the CPU. To anyone who wants to refute this by comparing BENCHMARKS, don't waste your time or mine, there is no real way to compare these, one's running Windows, the other Mac OS, they treat filesystems and executables VERY differently. Not even the ones on Apples website really show the speed, efficiancy, and performance. Hell if there was a way to rate efficiancy with benchmarks, Apple would win hands down. Not to say benchmarks don't work on AMD VS. Intel, they're great for that, and yes AMD kicks Intels rears in most of those.
Moving on, this is really simple, Steve Jobs will not go x86 if he can help it, and for good reason, they cannon handle multitasking the way the PPC subset in Power4 can, Apple has always been about better multitasking and more stability and security, a little tidbit I picked up the other day actually about PowerPC, it actually helps secure the Mac Platform, unlike x86, PowerPC treats data as data and executables as exacutables, x86 treats data and executables, like executables, meaning someone could write a virus that when you open a word document, the file itself sends the virus to the system without you being any the wiser, because on-screen it could work exactly like a document.
OK, This was really long, I'm trying to get everything into this one message, but I know I might be forgeting some points, I apoligize ahead of time, though I know I'll get slammed by some of you making those points, then I have to come back and explain and.......well you all know the drill.