Hi,
Can you help me with some calculation numbers on:
The performance numbers on the calculations parts,:
- C7x floating point vector DSP, 1.0 GHz, 80 GFLOPS, 256 GOPS
- 2x C66x floating point DSP, 1.35 GHz, 40 GFLOPS, 160 GOPS
- Deep learning matrix multiply accelerator up to 8 TOPS @ 1,0 GHz
- 3D GPU PowerVR Rogue 8XE GE8430, 750 MHz, 96 FLOPs, 160 GOPS
Which kind of data types have you been using, for example FP32, INT32, INT 8 etc?
- TDA4VM performance numbers for different data types and sizes of 2dfir-filters and FFT?
Best regards
Jorgen