Forum Home
Press F1
 
Thread ID: 116959 2011-03-27 19:09:00 Should I go to 64bit? Bryan (147) Press F1
Post ID Timestamp Content User
1189840 2011-03-28 18:19:00 That's not exactly "doubling" ;)

64 bit = 2 ^ 64
128 bit = 2 ^ 128 = (2 ^ 64)*(2 ^ 64)

In other words, 128 bit is the square of 64 bit.

I know this is not the exactly the right way to describe the difference between 64 bit and 128 bit architectures, but I am just showing how computing potential increases massively with 128 bit technology.

Cheers :)
Renmoo (66)
1189841 2011-03-28 19:05:00 Its the way the numbers work (not exactly sure what the proper terms are :p)

its like the way memory sizes go up in set units (I forgot what the pattern is) From memory its because the base unit it around 2 so the memory goes 2 4 6 8 16 32 64 128 256 512 1024 2048 etc I assume the communications between the hardware components use the same multiplier ratio such as 64 uses 64 bits of information

This multiplication factor is probably the reason 32x hardware can run on a 64x but

Then again i'm probably remembering this horribly incorrectly and am expecting a correction from the more wise among us
The Error Guy (14052)
1189842 2011-03-28 19:36:00 I often don't explain myself well.
I meant doubling as in double the amount of bits in an address.
I also meant 2^65 is double 2^64 numerically speaking.
At some point doubling the address length will just be ridiculous surely. 2048 bit computing?? hard to imagine. It's an exponential growth but there has to be a limit somewhere.

And they don't have to do it this way, any multiple of an 8 bit word should theoretically be workable, so we could go from 64 to 72 bit for example, or 96 which is 64 + 32 and could work out nicely. I'm sure I remember reading something about 48 bit devices somewhere, proprietry stuff but still.
dugimodo (138)
1189843 2011-03-28 19:55:00 I recently upgraded to a new computer and decided to go 64-bit as I had to reinstall all my programmes. By going 64-bit on new hardware I was future proofing my new computer. I do video editing so 64-bit was to my advantage. I now have 8 GB of ram whereas on my old 32-bit machine I was restricted to 4-GB of ram.

I had problems with the printer driver and some other older peripheral devices but everything is working fine at the moment. Most of my 32-bit programmes worked on 64-bit.
Bobh (5192)
1189844 2011-03-28 20:02:00 I had problems with the printer driver and some other older peripheral devices but everything is working fine at the moment. Most of my 32-bit programmes worked on 64-bit.

Ditto. I only have 4GB but I could upgrade if I need to ...
Software all work on mine but my printer won't so I use my laptop which runs Win7 x32 to print.

I understand HP might be better in this regard thou ...
Nomad (952)
1189845 2011-03-28 21:50:00 I had problems with the printer driver and some other older peripheral devices but everything is working fine at the moment. Most of my 32-bit programmes worked on 64-bit.

I had printer driver problems too but was able to solve it by browsing for a driver on my PC and selecting a 64bit one for the same brand and the same type of printer (inkjet) - didn't expect it to work but it did!
Richardd150 (13927)
1189846 2011-03-28 22:57:00 3 instances where 64bit (and memory to make use of it) might be useful:
- Adobe video editing - CS5 actually requires 64bit
- Audacity audio capture/editing - has an option to do everything in memory
- disc operations. A big memory cache might speed things up?? Certainly 8GB plays havoc with Windows' speedometer
BBCmicro (15761)
1189847 2011-03-28 23:55:00 I often don't explain myself well.
I meant doubling as in double the amount of bits in an address.
I also meant 2^65 is double 2^64 numerically speaking.
At some point doubling the address length will just be ridiculous surely. 2048 bit computing?? hard to imagine. It's an exponential growth but there has to be a limit somewhere.

And they don't have to do it this way, any multiple of an 8 bit word should theoretically be workable, so we could go from 64 to 72 bit for example, or 96 which is 64 + 32 and could work out nicely. I'm sure I remember reading something about 48 bit devices somewhere, proprietry stuff but still.The thing is, large memory address space isn't entirely the point - it's a useful byproduct of changing other things in the system.

If you have the ability to utilise 64bit whatevers inside the CPU anyway, you might as well take advantage of that for memory addressing.
Erayd (23)
1189848 2011-03-29 03:33:00 I often don't explain myself well .
I meant doubling as in double the amount of bits in an address .
I also meant 2^65 is double 2^64 numerically speaking .
At some point doubling the address length will just be ridiculous surely . 2048 bit computing?? hard to imagine . It's an exponential growth but there has to be a limit somewhere .




Well it doesn't HAVE to double, its just that a 48 bit address wouldn't be very useful because the architecture wouldn't offer any (or much advantage) over 32 bit . 96 is theoretical but I don't posess enough knowledge to point out pros/cons

I guess its like Mhz/Ghz surely an PC running 6 2Ghz cores (12000MHz) would have seemed impossible to people running 20MHz CPU's 20 years ago

Time will tell
The Error Guy (14052)
1 2