CharLS embedded performance

Mar 18, 2015 at 2:15 AM
I am considering using CharLS for JPEG-LS encoding in a real time embedded application. Does anyone know how many clock cycles are required per pixel when running on an embedded 32 bit processor (ARM Cortex M0 or similar). No RTOS, it will be a "bare metal" application. I am trying to calculate the maximum pixels per second throughput for a given processor clock rate to see if real time encoding is feasible.

Thank you.
May 12, 2015 at 6:09 PM
CharLS comes with a simple test application that can also do some performance measurements. You could use that to do some performance measurements on your platform. CharLS is C++ code, so performance also depends on the compiler you are using.

You could do also some tests on a Raspberry PI (uses also an ARM CPU) and then convert the performance numbers to the CPU your embedded system has.