Memory Guide for canmv

Viewed 97

1: Understanding Memory Distribution on CanMV-K230 V1.1

I am trying to understand the complete memory distribution for the CanMV-K230 V1.1 board. Specifically, I would like to know:

How much memory is required by RT-Smart to run?
How much memory is allocated for loading and running AI tasks?
How much temporary free memory is available for storing variables and runtime data?
And any other if missing?

Any insights or official references regarding memory allocation and management on this board would be highly appreciated.

2: Maximum AI Model Size for K230

Hello everyone,

I am working with the K230 and would like to understand the maximum AI model size that can be handled by the board. Specifically, I need help with:

What factors determine the largest model that can run on the K230?
Is there a method to predict whether a model will work on the K230 before deployment?
Are there any tools or techniques to optimize a model for running efficiently on this hardware?

1 Answers
  1. you can use make menuconfig, navigate to > Board Configuration > Memory Layout > Memory Layout for 1024M DDR ,here is the memory set for rtsmart and mmz(camera, display
  2. the model size is depend on the memory size.

@Wy001

But I am using canmv IDE and my firmware image is CanMV-K230_V1P0_P1_micropython_v1.2.2-0-g4b8cae1_nncase_v2.9.0.img

All board support this configure.