Performance of a C66x DSP core calculating Double-Precision floating point values

Daisuke Maeda

Guru 20205 points

Single-Precision floating point calculation performance as follows:

- Floating point MACs/cycle: 8

- Arithmetic Floating point operations/cycle: 16

Please let me know Double-Precision floating point calculation performance.

Best regards,

Daisuke

over 12 years ago

0 Chad Courtney over 12 years ago

TI__Mastermind 30825 points

Floating Point Double-Precision is going to be 2 MACs/cycle and 6 Double-Precission Floating point calculations (2 Multiplies, and 4 Artimatic functions - Load/Stores are not included in this number and neither are DP->INT INT->DP conversions - these operate on the .L units.)

Best Regards,
Chad

0 Daisuke Maeda over 12 years ago in reply to Chad Courtney

Guru 20205 points

Hi Chad,

Thank you for your reply.
Sorry for my late reply.

I understand that C66x Double-Precision Floating point calculations performance as follows:

- Floating point MACs/cycle: 2

- Arithmetic Floating point operations/cycle: 6 (2 Multiplies and 4 Arithmetic functions, except Load/Stores and DP conversions)

In addition, here is performance for C674x.

http://e2e.ti.com/support/dsp/omap_applications_processors/f/42/t/281160.aspx

Therefore, I understand that a C66x core is equivalent to a C674x core in Double-Precision Floating point calculations performance.

Best regards,

Daisuke

Processors

Processors forum

Performance of a C66x DSP core calculating Double-Precision floating point values