CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CFX (https://www.cfd-online.com/Forums/cfx/)
-   -   CFX 64bit single-double precision (https://www.cfd-online.com/Forums/cfx/23448-cfx-64bit-single-double-precision.html)

Tom December 19, 2006 15:40

CFX 64bit single-double precision
 
Hi. I am running CFX in a 64bit computer. But am having difficulty understanding what the difference between the single and double precision option is using CFX.

Is single and double still 8 and 16 decimal places respectively on a 64 bit computer?

or

Does it increase to 16 and 32 decimal places for single and double precision respectively?

Thanks for your help.

KB December 19, 2006 16:01

Re: CFX 64bit single-double precision
 
Single and double precision still means 8 resp. 16 digit precision for floating point numbers.

The 32 resp. 64 bit computers differ in the size of their virtual memory.

Charles December 19, 2006 16:39

Re: CFX 64bit single-double precision
 
To add to that, AFAIK x86 CPU's (both IA-32 and AMD64) actually do the maths in 80 bits, it's just a question of how much precision gets stored. So double precision is not primarily slower because of processing at higher precision, but rather because of the extra amount of information that has to be stored and retrieved from memory.


All times are GMT -4. The time now is 04:18.