Re: QtHandles and performance

[John W. Eaton]

On 01/09/2014 04:37 PM, John W. Eaton wrote:
> On 01/09/2014 04:30 PM, Rik wrote:
> > On 01/09/2014 12:06 PM, address@hidden wrote:
>
> > Do we need to re-measure the performance penalty before we can make an
> > informed decision, or are we confident in the 10% number?
>
> Measuring again is probably a good idea.  I compiled with optimization
> enabled (default -O2 for GCC) and timed running the test suite to compare.

OK, I did that test again and this is what I see (default -g -O2
compiler options for both cases):

  without atomic refcount (current default):

    make[1]: Leaving directory `/scratch/jwe/build/octave-opt/test'
    401.45user 49.45system 7:50.31elapsed 95%CPU (0avgtext+0avgdata 
415228maxresident)k
    153352inputs+20904outputs (172major+16480393minor)pagefaults 0swaps

  with atomic refcount:

    make[1]: Leaving directory `/scratch/jwe/build/octave-opt-atomic/test'
    405.32user 49.07system 7:51.65elapsed 96%CPU (0avgtext+0avgdata 
415348maxresident)k
    85768inputs+20920outputs (54major+16128734minor)pagefaults 0swaps

That's about, like, no difference.  Hmm.  And yes, octave-opt/config.h
has

  /* #undef USE_ATOMIC_REFCOUNT */

and octave-opt-atomic/config.h has

  #define USE_ATOMIC_REFCOUNT 1

Well, that's not the result I was expecting.  So I guess there's no
reason to NOT use atomic refcount.  I ran the test with GCC 4.8.2 on an
x86_64 system.  Maybe it was worse with earlier versions of GCC?  I
guess I could try that just to see if I can confirm what I remember as
the previous result.  Or we could also look up the previous discussion
in the mailing list archives.

jwe