Method and system for using deep learning to improve design verification by optimizing code coverage, functional coverage, and bug detection
Methods, systems, and devices for tuning a set of simulation parameters associated with a design verification environment are described that include: simulating a circuit design according to a set of simulation runs; providing, to a machine learning network, an indication of functional coverage results associated with simulating the circuit design according to the set of simulation runs; receiving an output in response to the machine learning network processing the functional coverage results; and simulating the circuit design based on a recommended set of simulation parameters. The output includes: an indication of a near miss event associated with the functional coverage results; and the recommended set of simulation parameters. Simulating the circuit design based on the recommended set of simulation parameters includes generating a set of component signals associated with triggering the near miss event.
1 . A system comprising:
a processor; and
a memory storing instructions thereon that, when executed by the processor, cause the processor to:
simulate a circuit design according to a set of simulation runs;
provide, to a machine learning network, an indication of functional coverage results associated with simulating the circuit design according to the set of simulation runs;
receive an output in response to the machine learning network processing the functional coverage results, wherein the output comprises:
an indication of a near miss event associated with the functional coverage results; and
a recommended set of simulation parameters; and
simulate the circuit design based on the recommended set of simulation parameters, wherein simulating the circuit design based on the recommended set of simulation parameters comprises generating a set of component signals associated with triggering the near miss event.
2 . The system of claim 1 , wherein the instructions are further executable by the processor to:
parse a design abstraction of the circuit design; and
identify, based on parsing the design abstraction, the set of component signals associated with triggering the near miss event,
wherein the machine learning network determines the recommended set of simulation parameters based on identifying the set of component signals.
3 . The system of claim 1 , wherein the near miss event is associated with a failure to generate all component signals of the set of component signals.
4 . The system of claim 1 , wherein processing the coverage results by the machine learning network comprises tracking the functional coverage results as a function of a set of simulation parameters associated with simulating the circuit design.
5 . The system of claim 1 , wherein a result associated with simulating the circuit design based on the recommended set of simulation parameters comprises achieving a target frequency associated with triggering the near miss event, generating the set of component signals, or both.
6 . The system of claim 1 , wherein the machine learning network:
identifies, in response to processing the functional coverage results, that a frequency associated with satisfying one or more functional coverage statements corresponding to the circuit design is less than a target frequency; and
generates a second recommended set of simulation parameters associated with satisfying the one or more functional coverage statements according to the target frequency.
7 . The system of claim 6 , wherein the machine learning network:
identifies a second set of component signals associated with triggering the one or more functional coverage statements by expanding the one or more functional coverage statements; and
generates the second recommended set of simulation parameters in association with triggering all component signals of the second set of component signals.
8 . The system of claim 1 , wherein the instructions are further executable by the processor to:
receive, from the machine learning network, a set of simulation parameters corresponding to the set of simulation runs; and
provide training data to the machine learning network based on the functional coverage results, wherein providing the training data comprises providing positive reinforcement, negative reinforcement, or both with respect to the set of simulation parameters and the functional coverage results.
9 . The system of claim 1 , wherein the machine learning network comprises a reinforcement learning network.
10 . A system for tuning a set of simulation parameters associated with a design verification by:
simulating a circuit design according to a set of simulation runs;
providing, to a machine learning network, an indication of functional coverage results associated with simulating the circuit design according to the set of simulation runs;
receiving an output in response to the machine learning network processing the functional coverage results, wherein the output comprises:
an indication of a near miss event associated with the functional coverage results; and
a recommended set of simulation parameters; and
simulating the circuit design based on the recommended set of simulation parameters, wherein simulating the circuit design based on the recommended set of simulation parameters comprises generating a set of component signals associated with triggering the near miss event.
11 . The system of claim 10 , wherein tuning the set of simulation parameters associated with the design verification includes:
parse a design abstraction of the circuit design; and
identify, based on parsing the design abstraction, the set of component signals associated with triggering the near miss event,
wherein the machine learning network determines the recommended set of simulation parameters based on identifying the set of component signals.
12 . The system of claim 10 , wherein the near miss event is associated with a failure to generate all component signals of the set of component signals.
13 . The system of claim 10 , wherein processing the coverage results by the machine learning network comprises tracking the functional coverage results as a function of a set of simulation parameters associated with simulating the circuit design.
14 . The system of claim 10 , wherein a result associated with simulating the circuit design based on the recommended set of simulation parameters comprises achieving a target frequency associated with triggering the near miss event, generating the set of component signals, or both.
15 . The system of claim 10 , wherein the machine learning network:
identifies, in response to processing the functional coverage results, that a frequency associated with satisfying one or more functional coverage statements corresponding to the circuit design is less than a target frequency; and
generates a second recommended set of simulation parameters associated with satisfying the one or more functional coverage statements according to the target frequency.
16 . The system of claim 15 , wherein the machine learning network:
identifies a second set of component signals associated with triggering the one or more functional coverage statements by expanding the one or more functional coverage statements; and
generates the second recommended set of simulation parameters in association with triggering all component signals of the second set of component signals.
17 . The system of claim 10 , wherein tuning the set of simulation parameters associated with the design verification includes:
receiving, from the machine learning network, a set of simulation parameters corresponding to the set of simulation runs; and
providing training data to the machine learning network based on the functional coverage results, wherein providing the training data comprises providing positive reinforcement, negative reinforcement, or both with respect to the set of simulation parameters and the functional coverage results.
18 . The system of claim 10 , wherein the machine learning network comprises a reinforcement learning network.
19 . A method comprising:
simulating a circuit design according to a set of simulation runs;
providing, to a machine learning network, an indication of functional coverage results associated with simulating the circuit design according to the set of simulation runs;
receiving an output in response to the machine learning network processing the functional coverage results, wherein the output comprises:
an indication of a near miss event associated with the functional coverage results; and
a recommended set of simulation parameters; and
simulating the circuit design based on the recommended set of simulation parameters, wherein simulating the circuit design based on the recommended set of simulation parameters comprises generating a set of component signals associated with triggering the near miss event.
20 . The method of claim 19 , further comprising:
parsing a design abstraction of the circuit design; and
identifying, based on parsing the design abstraction, the set of component signals associated with triggering the near miss event,
wherein the machine learning network determines the recommended set of simulation parameters based on identifying the set of component signals.