SECDEVTOOL-OMAPL138C6748: TMS320C674x complex vector multiplication

Yuehai Zhou1

Part Number: SECDEVTOOL-OMAPL138C6748
Other Parts Discussed in Thread: OMAPL138

Dear TI,

I have omapl138 evaluation board, I want to perform complex vector multiplication in DSP core (C6748), but in the dsp lib, I find the real vector multiplication. I program a function that performs complex vector multiplication, it seems very slow. Could you tell me how to perfom complex vector multiplication fastly and efficiently? Thank you very much.

Best regards,

Yuehai Zhou

over 4 years ago

0 Shankari G over 4 years ago

TI__Mastermind 25535 points

Hi Yuehai,

In the link below, we may find the performance of few DSPLIB and MATH LIB functions for OMAPL138/C6748. https://www.ti.com/microcontrollers-mcus-processors/processors/digital-signal-processors/core-benchmarks/core-benchmarks.html

Please do check for your required function. Let me also check.... mean while.

The below doc has the list of DSPLIB functions for C674x.

https://www.ti.com/lit/ug/spru657c/spru657c.pdf?ts=1623842149564&ref_url=https%253A%252F%252Fwww.google.com%252F

Regards

Shankari G

0 Aravind Batni over 4 years ago in reply to Shankari G

TI__Expert 8885 points

Hi Yuehai,

I see dsp lib supporting "DSPF_sp_dotp_cplx". Is this something you can use for your need?

Also, please note that you can speed up the implementation which you had done in "C" with the help of "intrinsic" which can help the compilers to schedule the code in best possible ways to get an optimal implementation.

Here is the link for the Optimizing Compiler guide

https://www.ti.com/lit/ug/sprui04b/sprui04b.pdf

Also, it may be better to consider to keep the buffers in internal memory and have Cache enabled to speed up the execution time.

Thanks.

Processors

Processors forum

SECDEVTOOL-OMAPL138C6748: TMS320C674x complex vector multiplication