11 posts.
Fifteen design verification interview questions with answers: test plans, constrained random, coverage, assertions, UVM, scoreboards and debugging.
From specification to silicon: architecture, hardware and software partitioning, RTL and IP integration, verification, implementation, sign-off and bring-up.
A layered SystemVerilog testbench for an APB register slave: driver, monitor, scoreboard, coverage and protocol assertions, and the planted bug it caught.
Design engineers write the RTL; verification engineers prove it works. What each does all day, the skills each needs, and how to try both before you choose.
Concurrent assertions check timing and results on every clock. SVA properties for the ALU, and how one assertion caught an injected subtraction bug.
UVM builds a testbench from standard parts: items, sequences, driver, monitor, scoreboard, agent and test. The ALU test rebuilt in UVM and run in xsim.
The course's final project: a constrained-random testbench for the synchronous FIFO with a queue model, flag assertions and coverage, and a planted bug.
Random stimulus kept legal and steered to corner cases by constraints, then measured, with real xsim counts and a randcase fallback for weights.
Functional coverage records which cases a random test hit. A covergroup for the ALU, and how a cross exposed gaps that per-field coverage missed.
Generator, driver, monitor and scoreboard, joined by mailboxes inside an environment: a layered testbench for the ALU that checks 200 random transactions.
Classes for transactions, a reference model, queues, tasks and assertions: the SystemVerilog a testbench leans on, in a first self-checking test of an ALU.