Skip to content

Latest commit

 

History

History
107 lines (85 loc) · 2.77 KB

README.md

File metadata and controls

107 lines (85 loc) · 2.77 KB
Raw

Spresense SDK repository

Clone this repository and update submodules.

$ git clone --recursive https://github.com/sonydevworld/spresense.git

After repositories have been cloned, each submodule is in 'Detached HEAD' state. Master branch have to be checked out before you can start developing.

$ git submodule foreach git checkout master

Submodules

.
├── examples  - Spresense SDK examples
├── nuttx     - NuttX original kernel source + CXD5602 port.
└── sdk       - Spresense SDK source and PC tools.

Using docker

A pre-compiled docker container is available with all the pre-requisite that is needed in order to build the Spresense SDK.

In order to start using it simply type:

$ source spresense_env.sh

This script will create an alias spresense which should preceed the regular SDK build scripts and Make commands. examples

SpresenseSDK: $ spresense make

Spresense SDK build instructions

Prerequisites

Install the necessary packages and GCC ARM toolchain for cross-compilation.

$ sudo apt-get install git gperf libncurses5-dev flex bison gcc-arm-none-eabi

Install the kconfig-frontends package from nuttx-tools.

$ git clone -n https://[email protected]/nuttx/tools.git tools
$ cd tools
$ git config core.autocrlf false
$ git checkout 891510d2534287afef6b5c90a375f94442a7e94d
$ cd kconfig-frontends/
$ ./configure
$ make
$ sudo make install
$ sudo ldconfig

Build

Go to the folder where you cloned the {SDK_FULL}, and enter the sdk folder name:

$ cd spresense/sdk

Set up the nuttx kernel configuration

$ tools/config.py --kernel release

Set up the SDK configuration

$ tools/config.py examples/hello

Build the example image:

$ make buildkernel
$ make

A nuttx.spk file appears in the sdk folder when this step has successfully finished. This file is the final result and can be flashed into the your board.

Connect via serial port

  • Install minicom
  • Configure it: Bps/Par/Bits : 115200 8N1, Hardware Flow Control : No, Software Flow Control : No
  • Connect to serial port (usb) to get the NuttX terminal

Execute the Assymetric Multi-processing example

Install genromfs. Set up the compilation configuration to enable ASMP and choose ASMP ELF Loader Example under Examples

$ tools/config.py -m

Build the example image:

$ make buildkernel
$ make

This example implements a parallel execution of an SVM inference. The main program (asmp) allocates the workload and the workers execute the computations. The result is reduced and interprented in the main program. The main program is executed using "asmp #numberOfWorkers". If no value for number of workers is provided then it defaults to 1.