Manticore: Hardware-Accelerated RTL Simulation with Static Bulk-Synchronous Parallelism

The demise of Moore's Law and Dennard Scaling has revived interest in specialized computer architectures and accelerators. Verification and testing of this hardware heavily uses cycle-accurate simulation of register-transfer-level (RTL) designs. The best software RTL simulators can simulate designs at 1--1000~kHz, i.e., more than three orders of magnitude slower than hardware. Faster simulation can increase productivity by speeding design iterations and permitting more exhaustive exploration. One possibility is to use parallelism as RTL exposes considerable fine-grain concurrency. However, state-of-the-art RTL simulators generally perform best when single-threaded since modern processors cannot effectively exploit fine-grain parallelism. This work presents Manticore: a parallel computer designed to accelerate RTL simulation. Manticore uses a static bulk-synchronous parallel (BSP) execution model to eliminate runtime synchronization barriers among many simple processors. Manticore relies entirely on its compiler to schedule resources and communication. Because RTL code is practically free of long divergent execution paths, static scheduling is feasible. Communication and synchronization no longer incur runtime overhead, enabling efficient fine-grain parallelism. Moreover, static scheduling dramatically simplifies the physical implementation, significantly increasing the potential parallelism on a chip. Our 225-core FPGA prototype running at 475 MHz outperforms a state-of-the-art RTL simulator on an Intel Xeon processor running at $\approx$ 3.3 GHz by up to 27.9$\times$ (geomean 5.3$\times$) in nine Verilog benchmarks.