TMS320F280039: How to improve the accuracy of float variables in CLA

Julia Li

Part Number: TMS320F280039

Tool/software:

Hi Champs,

The number of significant decimal places for float32 is about 6 to 7 digits. Now it is not enough for customer. Now customer need more significant decimal places.

Customer can use float64 in CPU, but in CLA it will cost more time like below thread, this will cause the customer's code to not run properly. Customer need to reduce code run time.

https://e2e.ti.com/support/microcontrollers/c2000-microcontrollers-group/c2000/f/c2000-microcontrollers-forum/1248583/tms320f28388d-double-precision-floating-point-type-make-cla-code-more-large/4722268?tisearch=e2e-sitesearch&keymatch=CLA%20AND%20float64#4722268

So is there a way to increase the significant decimal places of floating-point variables in CLA? Thank you!

Best Regards,

Julia

9 months ago

0 Ozino Odharo 8 months ago

TI__Mastermind 20885 points

Hi Julia,

Please take a look at the compiler's guide. Pay attention to the section 10.2.1 Variables and Data Types it describes the options. Let us know if you have more questions

https://downloads.ti.com/docs/esd/SPRU514/variables-and-data-types-t365212-3.html#t365212-3

It should also be noted that there are No 64-bit Types With CLA Compiler. The CLA compiler does not support 64-bit types.

Regards,

Ozino

C2000™︎ microcontrollers

C2000 microcontrollers forum

TMS320F280039: How to improve the accuracy of float variables in CLA