Profiling AlexNet on Raspberry Pi and HiKey 960 with the Compute LibraryOverview Set up your Raspberry Pi NFS on Pi Build the Arm Compute Library on Pi Run the graph_alexnet application on Pi Start Streamline gatord on Pi Add Streamline annotations and rebuild on Pi Build the Arm Compute Library on HiKey 960 Profile with Streamline on HiKey 960 Next steps
Start Streamline gatord on Pi
The Streamline gator daemon is a userspace application that can be built using the Android NDK and the gator driver is a Linux kernel driver.
To profile the example, you need to run
sudo. It can be run without
sudo, but the sample based profiling information will not be available and this is one of the interesting things to see for this application. The
gator kernel module is not needed for this tutorial as
sudo will provide profiling via the Linux perf API.
Use NFS to copy the 32-bit
gatord from DS-5 directly to the Pi using:
$ cp $DS5_HOME/sw/streamline/bin/arm/gatord /mnt
ssh to the Pi and start the
$ sudo ./gatord
Now that everything is setup, the next step is to look at the performance details of the graph_alexnet application.