Machine learning model training to assist in system debug
A process to train a machine learning model to predict fault location in a system includes generating a data set for training the machine learning model. The generating includes injecting, at a selected location of the system, a test fault into a simulation of the system using a workload, and recording a respective error syndrome generated by the simulation. Further, the generating includes repeating the injecting, at other selected location(s) of the system, of other test fault(s) into the simulation of the system, and the recording of respective, generated error syndromes. In addition, the process includes training, using the data set, the machine learning model, and providing the trained machine learning model for use in debugging the system, where the debugging includes predicting, using the trained machine learning model, a fault location within the system based on an error syndrome generated by the system due to the fault.
1. A computer-implemented method of facilitating processing within a computing environment, the computer-implemented method comprising:
generating a data set for training a machine learning model to predict fault location in a system, the generating including:
injecting, at a selected location of the system, a test fault into a simulation of the system using a workload;
recording a respective error syndrome generated by the simulation resulting from injecting the test fault into the simulation at the selected location;
repeating injecting, at one or more other selected locations of the system, of one or more other test faults into the simulation of the system using the workload, and the recording of respective error syndromes generated by the simulation resulting from injecting the one or more other test faults into the simulation of the system to generate the data set; and
training, using the data set, the machine learning model to predict fault location within the system; and
providing the trained machine learning model for use in debugging the system, where the debugging includes predicting, using the trained machine learning model, the fault location within the system based on an error syndrome generated by the system due to the fault.
2. The computer-implemented method of claim 1 , wherein generating the data set further comprises generating, for the workload, a table of test faults injected into the simulation of the system, along with respective potential paths of error propagation within the system and respective recorded error syndromes.
3. The computer-implemented method of claim 2 , wherein test faults injected into the system include one or more test faults indicative of a hardware fault.
4. The computer-implemented method of claim 1 , wherein generating the training data set further comprises deterministically determining the selected location in the system for injecting the test fault into the simulation of the system, and deterministically determining the one or more other selected locations in the system for injecting the one or more other test faults into the simulation of the system.
5. The computer-implemented method of claim 1 , wherein generating the training data set further comprises semi-deterministically determining the selected location in the system for injecting the test fault into the simulation of the system, and semi-deterministically determining the one or more other selected locations in the system for injecting the one or more other test faults into the simulation of the system.
6. The computer-implemented method of claim 1 , wherein generating the training data set further comprises randomly selecting the selected location in the system for injecting the test fault into the simulation of the system, and randomly selecting the one or more other selected locations in the system for injecting the one or more other test faults into the simulation of the system.
7. The computer-implemented method of claim 1 , wherein generating the training data set further comprises exhausting training data set generation by obtaining a netlist of hierarchal nodes across the system, and for a given sub-block of the system, sequentially injecting test faults into all nodes of the given sub-block of the system, and recording respective error syndromes generated by the simulation resulting from injecting test faults into all nodes of the given sub-block of the system.
8. The computer-implemented method of claim 1 , further comprising obtaining a further data set from deploying the machine learning model to localize faults during system test, with a different workload, and refining the machine learning model using the further data set.
9. The computer-implemented method of claim 1 , wherein the machine learning model comprises a multi-class classifier to predict fault location within the system using the error syndrome generated by the system due to the fault.
10. A computer program product for facilitating processing within a computing environment, the computer program product comprising:
a set of one or more computer readable storage media; and
program instructions, collectively stored in the set of one or more computer readable storage media, for causing at least one processor set to perform computer operations comprising:
generating a data set for training a machine learning model to predict fault location in a system, the generating including:
injecting, at a selected location of the system, a test fault into a simulation of the system using a workload;
recording a respective error syndrome generated by the simulation resulting from injecting the test fault into the simulation at the selected location;
repeating injecting, at one or more other selected locations of the system, of one or more other test faults into the simulation of the system using the workload, and the recording of respective error syndromes generated by the simulation resulting from injecting the one or more other test faults into the simulation of the system to generate the data set; and
training, using the data set, the machine learning model to predict fault location within the system; and
providing the trained machine learning model for use in debugging the system, where the debugging includes predicting, using the trained machine learning model, the fault location within the system based on an error syndrome generated by the system due to the fault.
11. The computer program product of claim 10 , wherein generating the data set further comprises generating, for the workload, a table of test faults injected into the simulation of the system, along with respective potential paths of error propagation within the system and respective recorded error syndromes.
12. The computer program product of claim 11 , wherein test faults injected into the system include one or more test faults indicative of a hardware fault.
13. The computer program product of claim 10 , wherein generating the training data set further comprises deterministically determining the selected location in the system for injecting the test fault into the simulation of the system, and deterministically determining the one or more other selected locations in the system for injecting the one or more other test faults into the simulation of the system.
14. The computer program product of claim 10 , wherein generating the training data set further comprises semi-deterministically determining the selected location in the system for injecting the test fault into the simulation of the system, and semi-deterministically determining the one or more other selected locations in the system for injecting the one or more other test faults into the simulation of the system.
15. The computer program product of claim 10 , wherein generating the training data set further comprises randomly selecting the selected location in the system for injecting the test fault into the simulation of the system, and randomly selecting the one or more other selected locations in the system for injecting the one or more other test faults into the simulation of the system.
16. The computer program product of claim 10 , wherein generating the training data set further comprises exhausting training data set generation by obtaining a netlist of hierarchal nodes across the system, and for a given sub-block of the system, sequentially injecting test faults into all nodes of the given sub-block of the system, and recording respective error syndromes generated by the simulation resulting from injecting test faults into all nodes of the given sub-block of the system.
17. A computer system for facilitating processing within a computing environment, the computer system comprising:
at least one processor set;
a set of one or more computer readable storage media; and
program instructions, collectively stored in the set of one or more computer readable storage media, for causing the at least one processor set to perform computer operations comprising:
generating a data set for training a machine learning model to predict fault location in a system, the generating including:
injecting, at a selected location of the system, a test fault into a simulation of the system using a workload;
recording a respective error syndrome generated by the simulation resulting from injecting the test fault into the simulation at the selected location;
repeating injecting, at one or more other selected locations of the system, of one or more other test faults into the simulation of the system using the workload, and the recording of respective error syndromes generated by the simulation resulting from injecting the one or more other test faults into the simulation of the system to generate the data set; and
training, using the data set, the machine learning model to predict fault location within the system; and
providing the trained machine learning model for use in debugging the system, where the debugging includes predicting, using the trained machine learning model, the fault location within the system based on an error syndrome generated by the system due to the fault.
18. The computer system of claim 17 , wherein generating the data set further comprises generating, for the workload, a table of test faults injected into the simulation of the system, along with respective potential paths of error propagation within the system and respective recorded error syndromes.
19. The computer system of claim 18 , wherein test faults injected into the system include one or more test faults indicative of a hardware fault.
20. The computer system of claim 17 , further comprising obtaining a further data set from deploying the machine learning model to localize faults during system test, with a different workload, and refining the machine learning model using the further data set.