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
_______________________________________________
Do not post admin requests to the list. They will be ignored.
PerfOptimization-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden
