Compiler/TDA2PXEVM: Floating point operations in VCOP

Santhiya R1

Part Number: TDA2PXEVM

Tool/software: TI C/C++ Compiler

Hi,

I need to implement algorithm in EVE VCOP kernel which has floating-point operations.

1. How to handle floating-point operations in VCOP kernel?

2. If Q number system can be used to convert float type to Q number type, will this be supported in VCOP? Found that Q number operations are different from normal arithmetic operations. So, how to handle multiplication/division of Q numbers in VCOP kernel?

over 4 years ago

0 Anshu Jain over 4 years ago

TI__Guru 56820 points

Hi,
VCOP is a fixed point processor and hence cannot directly work with floating point number. You will have to write a fixed point implementation of your kernel and then map it to VCOP.

Regards,

Anshu

0 Santhiya R1 over 4 years ago in reply to Anshu Jain

Prodigy 150 points

Hello Anshu,

Thanks for the response.

Any idea on 2 nd point which is Q number system on VCOP?

0 Anshu Jain over 4 years ago in reply to Santhiya R1

TI__Guru 56820 points

Hi Santhiya,

Q Point is just a way to represent a floating point number to fixed point number and it doesn't need any support from the processor. You will have to convert your algorithm to use this to convert it into fixed point algorithm.

Regards,

Anshu

Processors

Processors forum

Compiler/TDA2PXEVM: Floating point operations in VCOP