why the dm6446 dsp cpu load is so high(90%) while doing simple calculation?

I just add  simple code in  function VIDENCCOPY_TI_process in file videnc_copy.c. then compile , link and generate video_copy.x64P. 

for(ii=0;ii<700;ii++)
for(jj=0;jj<400;jj++)
            {
             *(outBufs->bufs[curBuf]+jj*720+ii)= (*(outBufs->bufs[curBuf]+jj*720+ii)*2+10;
            }

I modify the encodedecode example with dvsdk_1_30_01_41 and codec_engine_2_00_01 on DM6446 to call video_copy server.

I run encodedecode, but the DSP cpu Load is 90%,  frame rate 18fps. why cpu is so slow. I know after code optimization the speed will up.

But DM6446 has a C64x+ DSP core which is 4000MIPS. How to calculate its process power?

thank you!

.