Timers and timing, was: MySQL Performance 6.0rc1

[Mark Kirkwood]

Mark Kirkwood wrote:
> To add to the confusion :-), here are results from a dual PIII 1G (Tyan 
> S2510) running 6.0-RC1 (minus all debugging and malloc checking):
> 
>        null function: 0.01143
>             getpid(): 0.51563
>               time(): 5.27039
>       gettimeofday(): 5.24489
> 
> Interesting - kinda like the 5.4-STABLE results (but 2x slower for 
> time....). Again a pretty stock kernel ('options SMP' added obviously).
> 

Sorry - I am mistaken, there some debugging support enabled:

makeoptions DEBUG=-g        # Build kernel with gdb(1) debug symbols

Not sure if that would explain the poor showing for the time tests tho.