What's the deal with 64-bit processors?

[Playfrsbee]

Hi All,

I'm looking to improve my geek vocabulary. As it sounds like the 970 processor is a done deal, it appears that the future of Apple Computing will be in 64 bits. But what does that mean exactly. Does it mean that the code is written in "phrases" of 64 combinations of 0/1 (i.e. 2^64th power). If so, does this mean that the os will be more powerful? It would seem to me that this would also slow down the computer considerably. Am I totally off with my understanding of 64-bit computing?

Thanks for your time, I look forward to the lesson.

Ethan

 

[cc bcc]

This topic has been discussed over and over again. Have you tried the search button? I don't know the answers to your questions, perhaps a search would help.

 

[rainman::|:|]

Short answer is a 64-bit processor will handle twice as much information as the 32-bit ones we're using now. I don't think that necessarily means twice as fast, I think it just means that it works more efficiently at high-data volume.

I don't really know long answer, processor architecture isn't one of my speciality. As suggested try a search, someone's bound to have spelled it out...


pnw

 

[Catfish_Man]

Re: What's the deal with 64-bit processors?

Originally posted by Playfrsbee
Hi All,

I'm looking to improve my geek vocabulary. As it sounds like the 970 processor is a done deal, it appears that the future of Apple Computing will be in 64 bits. But what does that mean exactly. Does it mean that the code is written in "phrases" of 64 combinations of 0/1 (i.e. 2^64th power). If so, does this mean that the os will be more powerful? It would seem to me that this would also slow down the computer considerably. Am I totally off with my understanding of 64-bit computing?

Thanks for your time, I look forward to the lesson.

Ethan

Your impression is correct. The extra 32 bits in each variable add to bus traffic and use cache space, which can slow things down. The advantages of 64 bit chips are:

1) 64 bit memory addressing (so you can have 2^64 bits of memory instead of 2^32. 2^32 is 4GB)
2) Programs that use really big integers (previously they've had to fake them using multiple 32 bit integers)

For most current programs, 64 bit computing is about the same as 32 bit.

The reason why the PowerPC 970 is so cool is not really because it's 64 bit (although that will be useful when 4GB of ram isn't enough), it's because it's an very powerful desktop oriented chip. The G4+ on the other hand, is really meant for high end embedded (routers and such), which requires very low power and low latency, sometimes at the expense of overall performance.