Query queueing/Resource reservation

I am using FPGA to accelerate some of graph algorithms. FPGA is a single resource that can’t be used by two queries at the same time. Does TG have any queueing mechanism so that all queries can time share the FPGA?