I am looking to use parallel invoke on two functions, these two functions are themselves tbb::parallel_for functions.
My question is this even possible and if so what will be the effect of this on performance on an 8 CPU machine.
Thanks
Yes, this is possible. You will need to wrap parallel_for into functors or lambdas to pass to parallel_reduce.
The effect on performance depends on what the code does. But if your question is really about the number of thread and CPU utilization: there will be 8 threads running, one of those is the main application thread and 7 more will be created by TBB.