Formal Verification Method for Fault Coverage Capability of March Algorithm based on SpinalHDL

March algorithm is a series of algorithms commonly used for memory testing, aiming to detect and diagnose faults in memory. The continuously evolving March algorithms are capable of detecting an increasing number of complex faults. However, the question of whether March algorithm can achieve the intended design goal, i.e., achieving full coverage of theoretically detectable faults, currently lacks practical validation methods. One approach to address this issue is to conduct formal verification of the fault detection capability of March algorithm on a software platform. Therefore, this paper proposes a method that utilizes SpinalHDL to model the fault behavior of memory cells and March algorithm. By simulating the progression of March algorithm within the memory cells, reaching sensitization conditions during the progression, injecting faults into the memory cells, and checking whether March algorithm can detect these faults, the fault detection capability of March algorithm can be validated to determine if it meets the design expectations.