I would like to use the "Arm compute library" in a microprocessor (Cortex A9 based).
Is it possible to compile the arm compute library directly by importing source code in the IDE, in my case eclipse?
And compile it, I am using the "GCC ARM embedded" compiler.
I don't want to use a Linux OS.
Thanks in advance.
There is an example of how to build the Arm Compute Library for a Cortex-R52 using gcc and no operating system (bare metal).
No IDE but there is a build script that will show you the steps.
Let me know if you have any questions about it.
It's run on an Arm Fast Model to test. You can do the same with a Cortex-A9 Fast Model.
Thank you for your response, Jason.
I imported all source code into the eclipse IDE. I added the settings as described in the makefile for the Cortex-R52,
on the project properties.
I have an error during the compilation coming from a missing include file :
I don't know how I can resolve this issue.
I checked the bare metal example I referenced and it doesn't use OpenMP. This is where omp.h comes from. There is an option with the Arm Compute Library to use OpenMP but not for a bare metal target. The documentation about the Compute Library configuration is at https://arm-software.github.io/ComputeLibrary/v19.05/index.xhtml#S3_1_build_options
The configuration in the build.sh uses these flags (you can see openmp=0)
scons Werror=0 debug=1 neon=1 opencl=0 os=bare_metal arch=armv7a build=cross_compile cppthreads=0 openmp=0 standalone=1
Make sure you have the same configuration and the omp.h error should not show up.
View all questions in Machine Learning forum