vForce Performance

  • I recently posted a query on the Xcode-Users list but with no reply so far.

    I hopes of accelerating some standard C functions, I've run a few timing experiments to assess vForce.h function performance.  The results are disappointing.

    - Comparing the time to run vvsqrt( ) on a 10000 point data file to a simple for( ) loop,
      i.e., for( ii=0; ii<10000; ii++ ) yy[ii] = sqrt[ii] , shows a very slight improvement
        ( ~23 nsec per input data pt. for vvsqrt vice ~24 nsec for the loop),
    - Comparing the time to run vvsin( ) in similar fashion actually shows the loop to be better--
        ( ~39 nsec per data pt. for vvsin vice ~21 nsec for the loop).

    (FYI,  tests were run under Xcode 3.2.3 on a MacPro w Dual-Core, 3 GHz Xeon processor.)

    My questions are:
      - Are these timing results consistent with expected vForce performance?
    - If not, what does one need to do to get the expected performance?
    - And, regardless, are there other considerations that would favor use of the vForce functions?

    Don Thompson
  • On Mon, 23 Jan 2012 12:24:16 -0800, dct said:

    > I recently posted a query on the Xcode-Users list but with no reply so far.
    >
    > I hopes of accelerating some standard C functions, I've run a few timing
    > experiments to assess vForce.h function performance.  The results are
    > disappointing.

    I think the perfoptimization-dev list would be better:
    <perfoptimization-dev-request...>

    --
previous month january 2012 next month
MTWTFSS
            1
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31          
Go to today