topic Timing such short intervals in IntelĀ® oneAPI Math Kernel Library & IntelĀ® Math Kernel Library
https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/computation-time-for-1024-pt-complex-FFT-using-MKL-FFT/m-p/960117#M15835
Timing such short intervals is impractical in Fortran on Windows. It's possible but ugly to interface with QueryPerformanceCounter (less ugly, call a C function invoking __rdtsc()).
Under linux, I would suggest system_clock with integer(8) arguments.Fri, 30 Nov 2012 14:32:59 GMTTimP2012-11-30T14:32:59Zcomputation time for 1024 pt complex FFT using MKL_FFT
https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/computation-time-for-1024-pt-complex-FFT-using-MKL-FFT/m-p/960115#M15833
<P>HI,</P>
<P>i have been evaluating FFT of MKL.while computing , FFT takes approx. 1 milisecond which is too high.</P>
<P>i'm using following part of code to measure time of fft computation.</P>
<P>s_initial=dsecnd();</P>
<P>fft computation;</P>
<P>s_elapsed=dsecnd()-s_initial;</P>
<P>printf("time in milisec=%f",s_elapsed*1000);</P>
<P>can any one tell me how to reduce time compuation time gurther.</P>
<P>i'm using intel processor.</P>
<P>Thanks,</P>Thu, 29 Nov 2012 04:30:58 GMThttps://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/computation-time-for-1024-pt-complex-FFT-using-MKL-FFT/m-p/960115#M15833rakesh2012-11-29T04:30:58ZThis size FFT *computation*
https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/computation-time-for-1024-pt-complex-FFT-using-MKL-FFT/m-p/960116#M15834
This size FFT *computation* should take on the order of 10 microseconds.
If timing includes the planning/commit time, then it will take longer, of course.
Yet, 1 millisecond is very large.
DimaFri, 30 Nov 2012 12:48:30 GMThttps://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/computation-time-for-1024-pt-complex-FFT-using-MKL-FFT/m-p/960116#M15834Dmitry_B_Intel2012-11-30T12:48:30ZTiming such short intervals
https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/computation-time-for-1024-pt-complex-FFT-using-MKL-FFT/m-p/960117#M15835
Timing such short intervals is impractical in Fortran on Windows. It's possible but ugly to interface with QueryPerformanceCounter (less ugly, call a C function invoking __rdtsc()).
Under linux, I would suggest system_clock with integer(8) arguments.Fri, 30 Nov 2012 14:32:59 GMThttps://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/computation-time-for-1024-pt-complex-FFT-using-MKL-FFT/m-p/960117#M15835TimP2012-11-30T14:32:59Zactually dsecnd() is wrapper
https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/computation-time-for-1024-pt-complex-FFT-using-MKL-FFT/m-p/960118#M15836
actually dsecnd() is wrapper for __rdtsc, therefore customer can use this timing routine even for this short intervals too.Sat, 01 Dec 2012 02:31:49 GMThttps://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/computation-time-for-1024-pt-complex-FFT-using-MKL-FFT/m-p/960118#M15836Gennady_F_Intel2012-12-01T02:31:49ZThanks a lot .
https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/computation-time-for-1024-pt-complex-FFT-using-MKL-FFT/m-p/960119#M15837
Thanks a lot .Sat, 01 Dec 2012 05:55:27 GMThttps://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/computation-time-for-1024-pt-complex-FFT-using-MKL-FFT/m-p/960119#M15837rakesh2012-12-01T05:55:27ZHi,
https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/computation-time-for-1024-pt-complex-FFT-using-MKL-FFT/m-p/960120#M15838
Hi,
earlier i was calculating for only one iteration once i increased the no of iteration from one to 100 the calculation time reduced in the order of microsecond. now i get computation time around 0.02 ms(approx) for 1024 pt complex FFT.Mon, 03 Dec 2012 12:42:52 GMThttps://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/computation-time-for-1024-pt-complex-FFT-using-MKL-FFT/m-p/960120#M15838rakesh2012-12-03T12:42:52ZQuote:rakesh wrote:
https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/computation-time-for-1024-pt-complex-FFT-using-MKL-FFT/m-p/960121#M15839
<P></P><BLOCKQUOTE>rakesh wrote:<BR /><P></P>
<P>Hi,<BR /> earlier i was calculating for only one iteration once i increased the no of iteration from one to 100 the calculation time reduced in the order of microsecond. now i get computation time around 0.02 ms(approx) for 1024 pt complex FFT.</P>
<P></P></BLOCKQUOTE><P></P>
<P>While measuring time needed to complete 4096 sine values FFT routine I got ~121245 nanoseconds i.e 121 microseconds.Compiler intrinsic __rdtsc() was used and averaged.I hope that time interval needed to complete the FFT calculation was not influenced by the rdtsc latency which is thousand times slower. </P>Sat, 12 Jan 2013 18:41:09 GMThttps://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/computation-time-for-1024-pt-complex-FFT-using-MKL-FFT/m-p/960121#M15839Bernard2013-01-12T18:41:09Z