AM3359 ALU Performance Floating/Fixpoint operation

Jaikey

Prodigy 230 points

I was building a PI regulator for our generators. I'm not sure in which way the ALU is working?

I want to get the best performance for my regulation algorithm, so I must know more details

about the ALU of the ARM Cortex A8.

Does operations with fixpoint format works faster than floatpoint?

Does the ALU works with 32Bit operands and can it build a muliplication result of 64Bit?

Does the ALU detects overflows by themselves?

Thanks for helpful Links and replies

over 9 years ago

0 Biser Gatchev-XID over 9 years ago

TI__Guru**** 393215 points

Complete information about the ARM Cortex-A8 MPU can be found on the ARM web site: infocenter.arm.com/.../index.jsp The AM335X MPU revision is r3p2.

0 Jaikey over 9 years ago in reply to Biser Gatchev-XID

Prodigy 230 points

Thanks for the information but I'm still unsure if a fixpoint format like Q3.28 makes sense for faster operation,

because they seems to have a floating point unit.

0 JJD over 9 years ago in reply to Jaikey

TI__Guru* 86550 points

If you make use of Neon or the VFP, you will be able to get better performance because of its parallelism. You can get more information using Neon and the VFP in the AM335x here: processors.wiki.ti.com/.../Cortex-A8

Regards,

James

Processors

Processors forum

AM3359 ALU Performance Floating/Fixpoint operation