Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

KD7IWP

macrumors 6502a
Original poster
Mar 8, 2004
689
40
American living in Canada
Sorry about the last post, it was accidental.
Last semester one of my prof's was telling our class that the PPC chip was designed well and one of the reasons was that it handled divide by zero errors on the chip. He also warned that although the x86 chips are faster in raw speed, they are not designed quite as well. If this interests anyone, I had a kernel panic last night and had to reboot, and when I checked the panic.log, it was from a divide by zero error. Sad how that error has to take a whole system down
 

Shamus

macrumors 6502a
Feb 26, 2006
651
0
yeah, theres nothing wrong with the PPC. I quite like the architecture. Little things like this make you realise just how well designed the PPC is. I wonder why intel hasnt fixed a problem like that?
 

contoursvt

macrumors 6502a
Jul 22, 2005
832
0
So just because you got a divide by zero error that means that the universe is falling apart? Sorry but any number of hardware instabilities can cause this to happen including bad ram and a clock thats not stable.....
 

Kingsly

macrumors 68040
Does this mean the Intel macs are more prone to Kernel panics?
I was under the impression that the switch to Intel was not going to change a thing but for extra speed and some app incompatibilities. Oh well. kernel panics are not so bad. Much more friendly than the BSOD, and I get to learn about six different languages in the process :)
 

jhu

macrumors 6502a
Apr 4, 2004
854
1
there are plenty of well designed chips. mips comes to mind. but the cheap and fast ones are still made by amd and intel.
 

eXan

macrumors 601
Jan 10, 2005
4,738
134
Russia
Shamus said:
yeah, theres nothing wrong with the PPC. I quite like the architecture. Little things like this make you realise just how well designed the PPC is.

Same here :)
 

Josias

macrumors 68000
Mar 10, 2006
1,908
1
Actually, you realize how little you need. My bigbrother has a 14" iBook G4 with 1.07 G4 and 256 MB DDR RAM. It never hacks doing anything. He even does pretty heavy video editing with the 32 MB Ati. The 2.16 GHz Core Duo is around 8 times faster i guess. I can't imagine anything would consume that much power. Is it just to say you got the worlds fastest computer??:confused: hehe
 

Morn

macrumors 6502
Oct 26, 2005
398
0
A 68k is better designed than x86 even. Lets all go back to that and enjoy lovely 40mhz processing.:rolleyes:
 

MisterMe

macrumors G4
Jul 17, 2002
10,709
69
USA
KD7IWP said:
Sorry about the last post, it was accidental.
Last semester one of my prof's was telling our class that the PPC chip was designed well and one of the reasons was that it handled divide by zero errors on the chip. He also warned that although the x86 chips are faster in raw speed, they are not designed quite as well. If this interests anyone, I had a kernel panic last night and had to reboot, and when I checked the panic.log, it was from a divide by zero error. Sad how that error has to take a whole system down
Presumeably, when you say design you mean ISA. The x86 has its roots in processor designs which predate the personal computer. After pouring $10 billions into materials and manufacturing, Intel now enjoys an overwhelming advantage in processing power/watt and processing power/$ with the x86. There are far superior CISC ISA's. Intel has rolled over them. Most RISC ISA's, including the PPC, are far superior. However, none can overcome Intel's advantages.

Imagine a world in which the superior processor architectures had access to Intel's materials and manufacturing prowess. Imagine ....
 

dpaanlka

macrumors 601
Nov 16, 2004
4,869
34
Illinois
I believe that, theoretically, PowerPC and RISC itself is still technically superior than x86.

It's just that Intel has spent so much time, effort, and money into making really fast x86 chips.

Oh well :(
 

jhu

macrumors 6502a
Apr 4, 2004
854
1
dpaanlka said:
I believe that, theoretically, PowerPC and RISC itself is still technically superior than x86.

It's just that Intel has spent so much time, effort, and money into making really fast x86 chips.

Oh well :(

so that means x86 is still superior
 

mcmillan

macrumors 6502
Jul 4, 2004
318
0
Kingsly said:
(...) and I get to learn about six different languages in the process :)
See... Apple thinks on us all the time.:p

On topic: can perfectly remember Jobs talking about how the G5 had many less pipelines that the Intel processor. I don't know about the new processors, but I have always thought that PPC is better in a technological way.

IBM's poor administration might be the fault of the switch... who knows.
 

asencif

macrumors 6502
Dec 21, 2005
323
0
What makes Intel more attractive it's their roadmap of rolling out different types of mulit-core chips with lower power consumption especially for the laptops. However, I think their overall design is very ancient and just being tweaked with different variations. While they can put out more Ghz of speed their stability still remains in question as AMD and PPC both had lower clock speed chips that outperformed Intel. Now we are to believe that Intel's roadmap is superior, that they've gotten their act together and the technology is going leaps and bounds over PPC. Well to an extent is it true when it comes to the laptops, IBM's very slow production cycles, and Intel's ability to pump out tons of chips at a fast pace.

If you look at a company being superior because of their production speed then Intel wins out, however if you look at architecture then PPC still can hold it's own. In many people's opinions, PPC is still superior and can outperform anything Intel will make, if it keeps evolving. This part here is just speculation on my part, but I believe many of the Apple native apps that run fast or open fast on the new CD are based on their design and optimization for 32-Bit. If every Mac OS X apps including the OS had been converted and built especially for 64-Bit then the speed would be incredible. However, Apple could not do this either because the G4 chip is 32-Bit and they had the whole laptop line on it. It would cost more to have two different set of apps and OS's.
 

jhu

macrumors 6502a
Apr 4, 2004
854
1
dpaanlka said:
Speed wise, yes... but not like technology wise.

i beg to differ. just take away the x86 decoding front-end (which is pretty impressive by itself) and you have an architecture that rivals any current chip.
 

solvs

macrumors 603
Jun 25, 2002
5,684
1
LaLaLand, CA
The PPC is great and all, but woefully underfunded and underdeveloped. The G4 itself is actually a pretty good chip in and of itself, but since development stalled a bit, and what was developed was aimed at the embedded market, we were stuck with what we have. The G5 had potential, but not enough to overtake what Intel had in the pipeline. Money was a big issue. The P4 was designed to clock high, but was really poorly designed. When it hit the wall, it lost it's advantages. The newer systems are designed more like the Pentium M, which was more like the P3. I like the PPC and all, but I think Steve saw it just wasn't going to go where we needed it to.

Both AMD's chips and Intel's new ones actually share some of the good stuff from RISC, so not all is lost. x86 has come a long way from it's early CISC roots. But yeah, there are some advantages of RISC that we'll be loosing. Best case would have been a faster PPC, but that just wasn't a cost effective option. If at all.
 

jhu

macrumors 6502a
Apr 4, 2004
854
1
solvs said:
Both AMD's chips and Intel's new ones actually share some of the good stuff from RISC, so not all is lost. x86 has come a long way from it's early CISC roots. But yeah, there are some advantages of RISC that we'll be loosing. Best case would have been a faster PPC, but that just wasn't a cost effective option. If at all.

the only "risc" advantage missing from modern x86 processors is fixed-length instructions, and that's not really an advantage anymore since cisc instructions are smaller. thus more instructions can be fit into the processor's l1 and l2 cache. besides, if you look at the amount of instructions in the ppc instruction set, there's no way that it could possibly be called "reduced instruction set computer." a better term would be load-store, which all modern processors, x86 and ppc, are.
 

Morn

macrumors 6502
Oct 26, 2005
398
0
PPC might be a nicer instruction set for programmers and OS developers to work with. In that in has fewer exotic weirdnesses. But even if you gave the PPC intel's production and developement resources that wouldn't necessarily mean you'd get products better than x86, the main difference between x86 and PPC chips todays is just the front end, x86 chips need some more instruction decoding to feed the execution units at a decent rate. So you just need a few more transistors not a big deal with today's hundreds of millions of transistors chips.
x86 today outperforms chips that see plenty of developement like IBM's POWER5+ and the Itanium.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.