Difference between the PPC and the x86?
I am trying to convert one of my friends from the Windows world over to the Mac platform. So far, I have been able to amaze him at all the awesome things Mac OS X can do.
The other day, though, he asked me what advantages does the PPC have over Pentiums or even AMD processors and such.
mostly know the answersto most hardware questions, but this one left me with nothing but a blank face.
Can anybody help?
The other day, though, he asked me what advantages does the PPC have over Pentiums or even AMD processors and such.
mostly know the answersto most hardware questions, but this one left me with nothing but a blank face.
Can anybody help?
Comments
Intel & AMD both go for the GHz thing in a big way, its their main marketing tool - But basically a 1GHz PPC G4 more or less gets the same done as a 2Ghz Pentium 4.
What might persuade your friend is that the x86 architecture is near the end of its life - its only hung on this long because intel manage to keep pumping the MHz and there are a SHITLOAD of apps out there that only run on x86
The 970 is comming, its new technology. Its almost deffinately gonna be better.
At the end of the day, unless your friend is a low level programmer, does it really matter whats driving the machine - If its fast enough for him to do what he wants, and he likes it better than windows or linux then buy one. If he's the type of guy that will get one as a toy for novilty factor then start moaning about not being able to play games in 3 months, tell him to stick with his intel setup.
He was just curious (and so am I) if the PPC processors have a much cleaner architecture and are more powerful. That's pretty much it all he wanted to know.
I heard before that the G4 handles more instructions per second because it has a shorter pipeline compared to Intel's pentiums.
Is this true?
http://a784.g.akamai.net/7/784/51/54..._TO_012003.pdf
Pages 7-9 of the Tech Overview link above may also help out.
Originally posted by allformac
He loves Unix and loves playing around with it.
Be sure you send him to the developer pages on the Apple site so he can read about X11. If he's a Sourceforge junkie, he'll be thrilled to see all those X11 apps he can run on a Mac. There are lots of crossplatform developments in the works. And after he reads the G5 specs tomorrow...who could resist?
Originally posted by allformac
He loves Unix and loves playing around with it.
He was just curious (and so am I) if the PPC processors have a much cleaner architecture and are more powerful. That's pretty much it all he wanted to know.
I heard before that the G4 handles more instructions per second because it has a shorter pipeline compared to Intel's pentiums.
Is this true?
It doesn't exactly handle more instructions per second. That is determined more by bandwidth to memory (how many instructions you can pull over in a given amount of time), although it's rather complicated when caching and prediction comes into play.
It's got far more registers (32 general purpose registers, and a host of others). It's instruction set is a lot cleaner, if not perfect (at least all instructions are 32-bits... talk about easier to follow, and to prefetch!). It's shallower pipeline means less instructions have to be tossed out on a bad branch prediction. Their SIMD implementation (Altivec) is supposed to be much better than the x86.
Both have (at their best) a 64-bit data bus and a 36-bit address bus, I believe, so that's pretty much a wash.
I can't comment a lot on the strengths and weaknesses of the x86 architecture, as I don't do much work on it. But I am very experienced with the PowerPC (I've done embedded development on them for the past 5 years or so). I currently have a CompactPCI board I've been working on that has three 800MHz 750FX (G3) processors on it (it's a NICE board
Put it this way... if you were to ask a Mac to sort a pile of information the result would be the same as a PC (x86) but the process of sorting is different.
In the end each side (PC/MAC) has it's highlights but the winner is decided by the end-user. Why? Because the end-user needs a platform that works for them. If you only need a word processor/web browser it is cheaper to buy a PC, much cheaper. If photo editing, or minor video editing is your game then a Mac would be you best investment. If Gaming is your aim, then a PC is the path to take. Please don't convert any PC users to Mac if they would be better off down the line with a PC.
Anyway...
Originally posted by \\/\\/ickes
Not that this makes a major difference but in the basics of it all a PPC system will handle its byte code (ones and zeros) backwards compared to the x86. So when a PC stores 110100 is really written and read 001011 when it is on a Mac. It is really more complicated than that but that was as simple as I can possibly convey.
Put it this way... if you were to ask a Mac to sort a pile of information the result would be the same as a PC (x86) but the process of sorting is different.
In the end each side (PC/MAC) has it's highlights but the winner is decided by the end-user. Why? Because the end-user needs a platform that works for them. If you only need a word processor/web browser it is cheaper to buy a PC, much cheaper. If photo editing, or minor video editing is your game then a Mac would be you best investment. If Gaming is your aim, then a PC is the path to take. Please don't convert any PC users to Mac if they would be better off down the line with a PC.
Anyway...
Actually, it's bytes are "backward", not the bits. 0x1011 on the PC would be a 0x1110 on the PowerPC. Note, however, that it is the x86 that is "backwards". If you have the value 0x12345678 in a variable on the x86, it's actually stored as "0x78563412" in memory. When it has to be put out on the internet (as part of an IP header, for example), it has to be converted to "big-endian" (the same as the PowerPC uses).
When a variable has a value of 0x12345678 on the PowerPC, it's stored in memory the same way. This makes things a LOT easier to deal with (trust me!).
However, there is a reason Intel made this standard: effeciency. I don't have to change addresses when looking at a value as a byte, word, or dword. You do on a PowerPC (although it's "transparent" for the most part).
As for the "don't convert the users", I don't agree. While it's certainly CHEAPER to buy a PC for your web-browsing and email, I think Apple has a lot to offer for it's price. Note that Consumer Reports gave Apple top-rank for both customer support and reliability of its systems. You get what you pay for. I'd happily steer my parents and friends towards Apple for whatever their computer needs are (although I'd probaby steer gamers towards the PC, or more likely a console these days).
Boy it felt good to say that (Although, we are comparing vapourware due in the next two months to a PC you can buy now)
EG: To get dhcp ethernet working on a Mac, you plug the ethernet cable into the Mac. Try that on OpenBSD or linux.
I did say Byte... and the idea was to keep it simple for the... uninformed. I know how it works.
I know apple does a great job with support and a usable OS but I think a cheap (<$125) PC used just to serf the net and read e-mail is a great value. Yes, it does use windows... or even Linux but to get the, easy to use, OSX you need to pay the overhead of apples consumer line. I am not knocking apple... not at all, I like OSX and I love the look apple gives to its computers. That is why I think the use of the computer will decides which to buy. You don't need a sledge hammer to hit a nail.
Originally posted by \\/\\/ickes
Dear John Whitney
I did say Byte... and the idea was to keep it simple for the... uninformed. I know how it works.
Yes, I saw you say "byte code". However, in your example you reversed the bits, not the bytes (otherwise, 110100 would have gone to 000111). You also used all 0's and 1's, which also look like bits, so I wanted to clarify.