I have a written a backen compiler that generates x64 assembler opcodes. I can execute these code snippets properly with prologue and epilogue on both Windows and Linux.
The assembler contains an optimizer thats takes a higher level language and generates a very paralell set of programs and i would like to run these on the Phi coprocessor.
So lets say i have created 120 byte coded assembler snippets that i want to start in parallell. They are all different. Is this possible using this HW?
For more complete information about compiler optimizations, see our Optimization Notice.