topic hi in Intel® Integrated Performance Primitives

ipp lead to decrease performance

mmmmm__hamed — Sat, 21 Mar 2015 23:10:20 GMT

i have a system with 15 -750 cpu and windows 8.1 and I use IPP8.2 .

my program have a greate number Array Mul ,Div or sqrt.

i write a small code with ipp and use both ippm and ipps for mul array and compare that code performance with my simple array mul.

my code is this:

/-------------------------------------------------------------

Ipp64f* src=new Ipp64f[200]();
   for(int i=0;i<200;i++)
   {
       src=i; // initialize source array with values
   }
   Ipp64f* res=new Ipp64f[200]();

   double t1=GetTickCount();
   for(int k=0;k<5000000;k++)
   {
ippsMulC_64f(src,2,res,200);
   }
   double t2=GetTickCount();

double Ipp_Mul_Time=t2-t1;

   for(int k=0;k<5000000;k++)
   {
       for(int i=0;i<200;i++)
       {
res=src*2;
       }
   }
   double t3=GetTickCount();

   double MyMulTime=t3-t2;

//------------------------------------------------------------------------------------

i run this test program and in every test my simple Mul is faster than IPP Mul

i test this program for both small and big arrays and get same result.

i think that i can improve my progarm performance with intel ipp but it happend.

is exist a problem in my test code that lead to this abnormal behavor? or my test result is correct?

my cpu is i5 -750.

mmmmm__hamed — Sat, 21 Mar 2015 23:14:40 GMT

my cpu is i5 -750.

hi

mmmmm__hamed — Sun, 22 Mar 2015 07:44:11 GMT

my problem is solved

i find that if i mul to floating point number instead 2 , Ipp is faster.

i think in above case ipp doing 64 bit floating point mul and standard mul use another(i dont unknow) simpler operation.

with same operation ipp is winner.

Hi!

Sergey_K_Intel — Mon, 23 Mar 2015 05:21:59 GMT

Hi!

Please, be careful with measurements like these[:smile:])).

I have seen cases when optimizing compiler (like Intel C/C++) simply ignored some source code (even loops), if found that the results were not used later on.