Instruction Speed 
Author Message
 Instruction Speed

> As has been said, it varies from CPU to CPU.  On an 8088 it is
> probably faster to do the 3 MOVSW instructions.  This is because
> it takes less time to fetch the instruction bytes (3) than in the
> MOV CX,3; REP MOVSW (5).  The same may also be true on systems
> that have 2 or more wait states.
> Ironically, on a 486 it takes less time to use MOV AX,[DI]; MOV [SI],AX
> three times (even less if you use EAX!) if you are executing out of cache.

This example shows that on quite similar architectures, the particular
code sequence which gives fastest performance can still be architecture
dependent.  It is also true for high level languages.  Quite different
fortran or C code can be better on different architectures.  Thus it is
necessary for HLLs to have the option of the user, or the compiler writer,
or the library producer, providing alternative codings for a given purpose.

Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399
Phone: (317)494-6054

Mon, 18 Apr 1994 20:58:57 GMT  
 [ 1 post ] 

 Relevant Pages 

1. Instruction Speed and instruction availability

2. --- Instruction speed ---

3. --- Instruction speed ---

4. Instruction speeds

5. P4/Athlon instruction speed

6. pentium instruction clock speeds

7. instruction -> speed

8. Speed of WAM instructions

9. Reference: HW Instructions to Speed Up Lisp?

10. Speed..Speed..Speed

11. Perl speed vs. Python speed

12. integer*8 speed vs integer*4 speed


Powered by phpBB® Forum Software