Reducing the communication bottleneck via on-chip cosimulation of gate-level HDL and C-models on a hardware accelerator | IEEE Conference Publication | IEEE Xplore