Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
10 years ago today, Apple flipped the bird to PowerPC
- Thread starter Gamer9430
- Start date
- Sort by reaction score
I hear you, I guess I was just a bit confused by that particular upgrade path. Thanks for clarifying.
No problem. My own particular upgrade path during the intel transition was an iSight G5 iMac bought in December 2005....to, well, nothing - I didn't buy another Mac for 3 years and then it was an eMac from eBay.
I personally went from my 2003 era PowerMac G5 to a 2009 27" iMac. When I first laid eyes on that screen...I needed a towel. Now that was innovation!
Ding, ding, ding, ding! We have a winner. Now you know why I wrote:
"...had a slight, probably undetectable (outside of benchmarks) edge in performance due to twice..."
Beat it at what? What are the configurations of the systems? Making a generic statement is useless.
From a benchmark standpoint, based only on my quick research, a L05 dual core 2.0 GHz PowerMac G5 scores about a 1800 (per core?, 64bit) on GeekBench and a 2.0GHz intel core duo iMac gets around 2500 on a 32bit test. But these really aren't apples to apples. One is 64bit, the other is incapable of that. One has larger ram capacity and better graphics card options, bus speeds, etc. Geekbench numbers don't really compare to real world application use though...for that you need specialized testing.
PL595, when MacCubed refers to beat, I believe he's speaking on how fast the computers compare on tests. I found a nice write up here for those more curious on this topic: http://www.everymac.com/systems/apple/imac/faq_cd/imac-intel-speed-compared-to-imac-g5.html
Do I need to go over "expansion" again with a Power Mac G5 vs iMac? And still, the Core Duo iMacs saw up to Snow Leopard, which doesn't offer much these days versus Leopard (Well maybe Google Chrome or Firefox, but still you don't gain much by running a 1 year newer Mac).
I know he's referring to "tests". The question is: Which tests is he referring to? If he's not going to be specific I could merely point to Geekbench to prove him wrong.
[doublepost=1452887243][/doublepost]
Yes, you will need to go over expansion. How does one expand a processor?
I understand your point.
If you don't mind, point to the Geekbench scores you have in mind. I'm generally curious. I took a quick look this afternoon and found the Late 2005 PowerMac G5 (the two dual core machines 2.0 and 2.3) to have comparable geekbench scores to the Core Duo iMac released the following year. Granted, I only looked at the 2.0 Core Duo iMac and never bothered to check G5 iMacs. Everymac wasn't entirely clear on the scores for PMG5 so I assume those scores are 64bit only and CD is 32 bit. Speaking of, I'm confused why this discussion is comparing a PMG5 to a CDiMac? They are completely different animals. Seems more logical to compare G5iMac to CDiMac.
I didn't have anything specific in mind...I was merely running with the numbers you provided.
Damned if I know. My argument is the Core Duo and Core 2 Duo processors are almost identical in real world performance. I made this statement because someone mentioned Core Duo was a poor performer compared to the Core 2 Duo. Others have pulled out the disparate system specific comparisons. As you've noted such comparisons make little sense.
If the CD has weak ass GMA950 graphics, but that was on the consumer laptops only. At least compare like to like - the Pro kit had ATI X1600/nVidia 7300 graphics.
As I've said about 2 pages ago - the Core 2 Duo is a Core Duo plus added 64-bit and a smidge more cache. The main guts of the processor - the x86 units are damn near identical.
In 2016 is a Core Duo a bad investment against a marginally more expensive Core 2 Duo Mac from a couple of years later with the Nvidia chipset? - Yes. That doesn't make it a CPU that "really sucked" to quote the OP.
According to EveryMac, the first Core 2 Duo MacBook is able to run 10.7.5.
Yes, sure, from a theoretical point of view, Apple could have released Lion with a 32-bit kernel and made Mountain Lion and up compatible with 32-bit CPUs but it didn't and nothing can change that.
The full 64-bit transition went the following way:
- Panther: support for 64-bit CPUs but no 64-bit processes
- Tiger: support for 64-bit interface-less processes
- Leopard: support for 64-bit applications
- Snow Leopard: comes with a 64-bit kernel (but not enabled by default) and almost all pre-loaded applications are 64-bit
- Lion: boots on the 64-bit kernel by default on supported Macs
- Mountain Lion and up: no more support for Macs unable to boot on the 64-bit kernel
Another difference is that x64 comes with 16 general purpose registers (GPR) instead of only 8 on x86 (and 32 on PowerPC, be it 32 or 64 bits).
This usually translates to a real world performance improvement of 10-15% using x64 software whereas using PPC64 software usually comes with a slight performance penalty (since you don't get any more GPR and you have to deal with pointers using 64 bits instead of 32).
I'm aware of that, I just went with the important highlights rather than architectural advances for x64. I've got a degree in Computer Science, most people in here do not. Added 64-bit should cover it nicely.
My mistake. With so many different combinations I got confused as to what was what.
The only thing preventing Lion and later OS X releases from running on the Core Duo processors is Apples decision not to support it. Their decision to do so is understandable and I do not fault them for that decision. However the performance of the two processors, outside of benchmarks, was essentially identical. I owned both systems (and own the 2,1 MacBook Core 2 Duo version today) and there was no tangible, real world difference. Certainly not enough to say that "it really sucked" unless you're going to say the Core 2 Duo process really sucked too.
If the software wasn't optimized for the Intel chipsets and/or used Rosetta, and at the time of the switch NOTHING was optimized, the G5 would remove all doubt as the better processor. If it were a Universal Binary and the developer actually gave a damn, the G5's would certainly keep up with the Intel chips. If it were a crapsack UB that didn't care about PPC, the Intel processor would work better.
It all depends on the coding of the application as to whether or not the G5 would work just as well. Even today, I am still struck at how an old Universal Binary done right can make a G5 feel so snappy.
One of the "last gasps" of UBs was Adobe CS4, and I'd call it a universal binary done right. Even on a low end single 1.8 or on a G4, I find CS4 to still be quite usable and very peppy.
I mentioned this in another thread, but it's hard to say how many G4 and G5 era machines are still churning out day-to-day work in print shops, newspapers, and graphic design houses. My first G5 came from a local design agency and they'd only retired them in 2012 or so. They get the job done, and do so quite well in most cases at these places. Ask Erik how the G5 and the G4s are doing at the newspaper where he works.
While they can certainly get the job done the question is would a newer system get the job done faster. Most likely the answer is yes a newer system would. Whether there's benefit depends on the job and company.
I like PPC based systems but if I were running a business and time equals money I wouldn't continue to use one (all else being equal).