TDA2: Deep learning algorithms run very slowly on DSP(TDA2xx)

Hi,

Deep learning algorithms run very slowly on DSP(TDA2xx)

image size is 512x256.

### CPU [  DSP1], LinkID [ 49],
[IPU1-0]    831.681815 s: 
[IPU1-0]    831.681845 s:  [ ALG_TIDL ] Link Statistics,
[IPU1-0]    831.681937 s:  ******************************
[IPU1-0]    831.681998 s: 
[IPU1-0]    831.682028 s:  Elapsed time       = 633812 msec
[IPU1-0]    831.682120 s: 
[IPU1-0]    831.682272 s:  New data Recv      =   0.2 fps
[IPU1-0]    831.682364 s: 
[IPU1-0]    831.682394 s:  Input Statistics,
[IPU1-0]    831.682455 s: 
[IPU1-0]    831.682516 s:  CH | In Recv | In Drop | In User Drop | In Process
[IPU1-0]    831.682577 s:     | FPS     | FPS     | FPS          | FPS       
[IPU1-0]    831.682669 s:  --------------------------------------------------
[IPU1-0]    831.682760 s:   0 |   0. 1      0. 0      0. 0           0. 1

Computational complexity:

thanks!