Systems, methods and computer program products for prediction of defect-related failures in integrated circuits

• US 5,822,218 A
• Filed: 08/27/1996
• Issued: 10/13/1998
• Est. Priority Date: 08/27/1996
• Status: Expired due to Term

• Current Assignee: Clemson University
• Original Assignee: Clemson University
• Inventors: Moosa, Mohamed S., Poole, Kelvin F.
• Primary Examiner(s): Voeltz, Emanuel Todd
• Assistant Examiner(s): GARBOWSKI, LEIGH M
• Application Number:
  US08/703,518
• Time in Patent Office: 777 Days
• Field of Search: 364/488, 364/489, 364/490, 364/491, 364/578, 364/468.28
• US Class Current: 716/52
• CPC Class Codes: G01R 31/31835 Analysis of test coverage or failure detectability

Abstract

Systems, methods and computer program products for predicting defect-related failures in integrated circuits produced by an integrated circuit fabrication process identify objects in a circuit layout for the integrated circuit design, each object having a location in the circuit layout and a reliability connectivity in the integrated circuit design. Sample object defects are generated for the identified objects, each sample object defect representing a defect produced in an object by the integrated circuit fabrication process and having a defect magnitude associated therewith. An accelerated life defect influence model is identified for each sample object defect, relating the lifetime of an object to the defect magnitude of a defect in the object. Sample object lifetimes are generated from the defect magnitudes associated with the sample object defects according to the corresponding identified accelerated life defect influence models. A prediction of the reliability of integrated circuits is generated from the sample object lifetimes according to the reliability connectivity of the associated objects in the integrated circuit design. Preferably, the accelerated life defect influence models include log-linear regression models, which may include deterministic object lifetime functions, each relating the defect magnitude of the at least one sample object defect to one object lifetime value, and log-linear object lifetime distributions, each relating the defect magnitude of a sample object defect to a plurality of object lifetime values.

Claims

1. A method of predicting defect-related failures in integrated circuits produced by an integrated circuit fabrication process according to a circuit layout for an integrated circuit design, the method comprising the following steps which are performed in a data processing system:

identifying objects in the circuit layout, each object having a location in the circuit layout and a reliability connectivity in the integrated circuit design;

generating sample object defects for the identified objects, each sample object defect representing a defect induced in an object by the integrated circuit fabrication process, each sample object defect having a defect magnitude associated therewith;

identifying an accelerated life defect influence model for each sample object defect, each accelerated life defect influence model relating the lifetime of an object to the defect magnitude of a defect in the object;

generating sample object lifetimes from the defect magnitudes associated with the sample object defects according to the corresponding identified accelerated life defect influence models; and

generating a prediction of the reliability of integrated circuits produced by the integrated circuit fabrication process according to the circuit layout, from the sample object lifetimes according to the reliability connectivity of the associated objects in the integrated circuit design.

2. A method according to claim 1 wherein said step of generating a prediction of the reliability of integrated circuits comprises the step of:

generating sample integrated circuit lifetimes from the sample object lifetimes according to the reliability connectivity of the associated objects.

3. A method according to claim 1 wherein said step of identifying an accelerated life defect influence model comprises the step of identifying a log-linear regression model for at least one of the sample object defects, the log-linear regression model having the form

4. A method according to claim 3:

wherein said step of identifying a log-linear regression model comprises the step of identifying a deterministic object lifetime function relating the defect magnitude of the at least one sample object defect to one object lifetime value; and

wherein said step of generating a sample object lifetime corresponding to the at least one sample object defect comprises the step of computing a sample object lifetime corresponding to the at least one sample object defect from the defect magnitude of the at least one sample object defect according to the deterministic object lifetime function to thereby generate a sample object lifetime for the at least one sample object defect.

5. A method according to claim 3:

wherein said step of identifying a log-linear regression model comprises the step of identifying a log-linear object lifetime distribution relating the defect magnitude of a sample object defect to a plurality of object lifetime values; and

wherein said step of generating a sample object lifetime corresponding to the at least one sample object defect comprises the step of sampling the log-linear object lifetime distribution to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

6. A method according to claim 3:

wherein said step of identifying a log-linear regression model comprises the step of identifying a predetermined generalized parametric object lifetime distribution function derived from intrinsic reliability data and having a distribution parameter related to the defect magnitude of a sample object defect by a distribution parameter function; and

wherein said step of generating a sample object lifetime corresponding to the at least one sample object defect comprises the steps of;

computing a value for the distribution parameter for the at least one sample object defect from the defect magnitude of the at least one sample object defect according to the distribution parameter function to thereby define a particularized object lifetime distribution function for the at least one sample object defect; and

sampling the particularized object lifetime distribution function to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

7. A method according to claim 3:

wherein said step of identifying a log-linear regression model comprises the step of identifying a predetermined generalized parametric object lifetime distribution function derived from intrinsic reliability data and having a distribution parameter related to the defect magnitude of a sample object defect and a conditional probability factor by a distribution parameter function, the conditional probability factor being independent of the predetermined generalized parametric object lifetime distribution function; and

wherein said step of generating a sample object lifetime corresponding to the at least one sample object defect comprises the steps of;

computing a value for the distribution parameter for the at least one sample object defect from the defect magnitude of the at least one sample object defect and the conditional probability factor according to the distribution parameter function to thereby define a particularized object lifetime distribution function for the at least one sample object defect; and

sampling the particularized object lifetime distribution function to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

8. A method according to claim 1:

wherein said step of identifying an accelerated life defect influence model for each sample object defect comprises the steps of;

identifying a deterministic object lifetime function relating the defect magnitude of a first sample object defect to one object lifetime value, for the first sample object defect; and

identifying an object lifetime distribution relating the defect magnitude of a second sample object defect to a plurality of object lifetime values, for the second sample object defect; and

wherein said step of generating sample object lifetimes comprises the steps of;

computing an object lifetime value from the defect magnitude of the at least one sample object defect according to the deterministic object lifetime function to thereby generate a sample object lifetime corresponding to the first sample object defect; and

sampling the object lifetime relating to the defect magnitude of the second sample object defect to thereby generate a sample object lifetime corresponding to the second sample object defect.

9. A method according to claim 1:

wherein said step of identifying objects from the circuit layout comprises the steps of;

identifying cells in the circuit layout, each of the cells occupying an area of the circuit layout and having a reliability connectivity in the integrated circuit design; and

identifying objects within each cell, each of the objects within a cell having a location in the cell and a reliability connectivity in the cell;

wherein said step of generating sample object defects comprises the step of generating sample object defects for objects within the cells;

wherein said step of generating sample object lifetimes comprises the step of generating sample object lifetimes for objects within the cells; and

wherein said step of generating a prediction of the reliability of integrated circuits comprises the steps of;

generating sample cell lifetimes for cells from the sample object lifetimes for objects within the cells according to the reliability connectivity of the associated objects in the cells; and

generating a prediction of the reliability of integrated circuits produced by the integrated circuit fabrication process from the sample cell lifetimes according to the reliability connectivity of the associated cells in the integrated circuit design.

10. A method according to claim 1 further comprising the steps of:

modifying the integrated circuit design based on the predicted reliability of integrated circuits produced by the integrated circuit process according to the circuit layout; and

fabricating integrated circuits according to the modified integrated circuit design.

11. A method according to claim 1 further comprising the steps of:

modifying the integrated circuit fabrication process based on the predicted reliability of integrated circuits produced by the integrated circuit process according to the circuit layout; and

fabricating integrated circuits using the modified integrate circuit fabrication process.

12. A method of predicting defect-related failures in integrated circuits produced by an integrated circuit fabrication process according to a circuit layout for an integrated circuit design, the method comprising the following steps which are performed in a data processing system:

identifying objects in the circuit layout, each object having an object type, a location in the circuit layout and a reliability connectivity in the integrated circuit design;

providing a process flaw distribution associated with the integrated circuit fabrication process, the process flaw distribution representing a statistical distribution of flaws produced by the integrated circuit fabrication process;

generating sample flaws according to the process flaw distribution, each of the sample flaws having a location with respect to the circuit layout;

correlating the locations of the sample flaws with the locations of the identified objects to thereby generate sample object defects, each sample object defect representing a defect induced in an object by a sample flaw, each sample object defect having a defect magnitude associated therewith;

identifying an accelerated life defect influence model for each sample object defect, each accelerated life defect influence model representing an influence of the defect magnitude of a defect on the lifetime of an object;

generating sample object lifetimes for the identified objects from the defect magnitudes associated with the sample object defects according to the corresponding identified accelerated life defect influence models; and

generating a prediction of the reliability of integrated circuits produced by the integrated circuit fabrication process according to the circuit layout from the sample object lifetimes according to the reliability connectivity of the associated identified objects in the integrated circuit design.

13. A method according to claim 12 wherein said step of generating sample flaws comprises the step of sampling the process flaw distribution.

14. A method according to claim 12 wherein the process flaw distribution comprises at least one of a process flaw magnitude distribution, a process flaw spatial distribution, and a process flaw quantity distribution.

15. A method according to claim 12 wherein the process flaw distribution comprises a process flaw magnitude distribution, a process flaw spatial distribution, and a process flaw quantity distribution.

16. A method according to claim 12:

wherein said step of identifying objects comprises the step of identifying objects having a critical area associated therewith;

wherein said step of generating sample flaws comprises the step of generating sample flaws having a location and an area; and

wherein said step of correlating the locations of sample flaws with the locations of the identified objects comprises the steps of generating a sample object defect if a critical area of an object intersects the area of a sample flaw.

17. A method according to claim 12 wherein said step of identifying an accelerated life defect influence model comprises the step of identifying a log-linear regression model for at least one of the sample object defects, the log-linear regression model having the form

18. A method according to claim 17:

wherein said step of identifying a log-linear regression model comprises the step of identifying a deterministic object lifetime function relating the defect magnitude of a sample object defect to one object lifetime value; and

wherein said step of generating a sample object lifetime corresponding to the at least one sample object defect comprises the step of computing a sample object lifetime corresponding to the at least one sample object defect from the defect magnitude of the at least one sample object defect according to the deterministic object lifetime function to thereby generate a sample object lifetime for the at least one sample object defect.

19. A method according to claim 17:

wherein said step of identifying a log-linear regression model comprises the step of identifying a log-linear object lifetime distribution relating the defect magnitude of a sample object defect to a plurality of object lifetime values; and

wherein said step of generating a sample object lifetime corresponding to the at least one sample object defect comprises the step of sampling the log-linear object lifetime distribution to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

20. A method according to claim 17:

wherein said step of identifying a log-linear regression model comprises the step of identifying a predetermined generalized parametric object lifetime distribution function derived from intrinsic reliability data and having a distribution parameter related to the defect magnitude of a sample object defect by a distribution parameter function; and

wherein said step of generating a sample object lifetime corresponding to the at least one sample object defect comprises the steps of;

computing a value for the distribution parameter for the at least one sample object defect from the defect magnitude of the at least one sample object defect according to the distribution parameter function to thereby define a particularized object lifetime distribution function for the at least one sample object defect; and

sampling the particularized object lifetime distribution function to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

21. A method according to claim 17:

wherein said step of identifying a log-linear regression model comprises the step of identifying a predetermined generalized parametric object lifetime distribution function derived from intrinsic reliability data and having a distribution parameter related to the defect magnitude of a sample object defect and a conditional probability factor by a distribution parameter function, the conditional probability factor being independent of the predetermined generalized parametric object lifetime distribution function; and

wherein said step of generating a sample object lifetime corresponding to the at least one sample object defect comprises the steps of;

computing a value for the distribution parameter for the at least one sample object defect from the defect magnitude of the at least one sample object defect and the conditional probability factor according to the distribution parameter function to thereby define a particularized object lifetime distribution function for the at least one sample object defect; and

sampling the particularized object lifetime distribution function to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

22. A method according to claim 12:

wherein said step of identifying an accelerated life defect influence model for each sample object defect comprises the steps of;

identifying a deterministic object lifetime function relating the defect magnitude of a first sample object defect to one object lifetime value, for the first sample object defect; and

identifying an object lifetime distribution relating the defect magnitude of a second sample object defect to a plurality of object lifetime values, for the second sample object defect; and

wherein said step of generating sample object lifetimes comprises the steps of;

computing an object lifetime value from the defect magnitude of the at least one sample object defect according to the deterministic object lifetime function to thereby generate a sample object lifetime corresponding to the first sample object defect; and

sampling the object lifetime relating to the defect magnitude of the second sample object defect to thereby generate a sample object lifetime corresponding to the second sample object defect.

23. A method according to claim 12:

wherein said step of identifying objects from the circuit layout comprises the steps of;

identifying cells in the circuit layout, each of the cells occupying an area of the circuit layout and having a reliability connectivity in the integrated circuit design; and

identifying objects within each cell, each of the objects within a cell having a location in the cell and a reliability connectivity in the cell;

wherein said step of generating sample object defects comprises the step of generating sample object defects for objects within the cells;

wherein said step of generating sample object lifetimes comprises the step of generating sample object lifetimes for objects within the cells; and

wherein said step of generating a prediction of the reliability of integrated circuits comprises the steps of;

generating sample cell lifetimes for cells from the sample object lifetimes for objects within the cells according to the reliability connectivity of the associated objects in the cells; and

generating a prediction of the reliability of integrated circuits produced by the integrated circuit fabrication process from the sample cell lifetimes according to the reliability connectivity of the associated cells in the integrated circuit design.

24. A computer program product, comprising:

a computer-readable storage medium having computer-readable program code means embodied in said medium, said computer-readable program code means comprising;

first computer-readable program code means for storing a circuit layout for an integrated circuit design;

second computer-readable program code means, responsive to said first computer-readable program code means, for identifying objects in the circuit layout, each object having a location in the circuit layout and a reliability connectivity in the integrated circuit design;

third computer-readable program code means, responsive to said second computer-readable program code means, for generating sample object defects for the identified objects, each sample object defect representing a defect induced in an object by an integrated circuit fabrication process producing integrated circuits according to the circuit layout, each sample object defect having a defect magnitude associated therewith;

fourth computer-readable program code means, responsive to said third computer-readable computer program code means, for identifying an accelerated life defect influence model for each sample object defect, each accelerated life defect influence model relating the lifetime of an object to the defect magnitude of a defect in the object;

fifth computer-readable program code means, responsive to said third and fourth computer-readable program code means, for generating sample object lifetimes from the defect magnitudes associated with the sample object defects according to the corresponding identified accelerated life defect influence models; and

sixth computer-readable program code means, responsive to fifth computer-readable program code means, for generating a prediction of the reliability of integrated circuits produced by the integrated circuit fabrication process according to the circuit layout from the sample object lifetimes according to the reliability connectivity of the associated objects in the integrated circuit design.

25. A computer program product according to claim 24 wherein said fourth computer-readable program code means comprises seventh computer-readable program code means for identifying a log-linear regression model for at least one of the sample object defects, the log-linear regression model having the form

26. A computer program product according to claim 25:

wherein said seventh computer-readable program code means comprises computer-readable program code means for identifying a deterministic object lifetime function relating the defect magnitude of a sample object defect to one object lifetime value; and

wherein said eighth computer-readable program code means comprises computer-readable program code means for computing a sample object lifetime corresponding to the at least one sample object defect from the defect magnitude of the at least one sample object defect according to the deterministic object lifetime function to thereby generate a sample object lifetime for the at least one sample object defect.

27. A computer program product according to claim 25:

wherein said seventh computer-readable program code means comprises computer-readable program code means for identifying a log-linear object lifetime distribution relating the defect magnitude of a sample object defect to a plurality of object lifetime values; and

wherein said eighth computer-readable program code means comprises computer-readable program code means for sampling the log-linear object lifetime distribution to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

28. A computer program product according to claim 25:

wherein said seventh computer-readable program code means comprises computer-readable program code means for identifying a predetermined generalized parametric object lifetime distribution function derived from intrinsic reliability data and having a distribution parameter related to the defect magnitude of a sample object defect by a distribution parameter function; and

wherein said eighth computer-readable program code means comprises;

computer-readable program code means for computing a value for the distribution parameter for the at least one sample object defect from the defect magnitude of the at least one sample object defect according to the distribution parameter function to thereby define a particularized object lifetime distribution function for the at least one sample object defect; and

computer-readable program code means for sampling the particularized object lifetime distribution function to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

29. A computer program product according to claim 25:

wherein said seventh computer-readable program code means comprises computer-readable program code means for identifying a predetermined generalized parametric object lifetime distribution function derived from intrinsic reliability data and having a distribution parameter related to the defect magnitude of a sample object defect and a conditional probability factor by a distribution parameter function, the conditional probability factor being independent of the predetermined generalized parametric object lifetime distribution function; and

wherein said eighth computer-readable program code means comprises;

computer-readable program code means for computing a value for the distribution parameter for the at least one sample object defect from the defect magnitude of the at least one sample object defect and the conditional probability factor according to the distribution parameter function to thereby define a particularized object lifetime distribution function for the at least one sample object defect; and

computer-readable program code means for sampling the particularized object lifetime distribution function to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

30. A computer program product according to claim 24:

wherein said fourth computer-readable program code means comprises;

computer-readable program code means for identifying a deterministic object lifetime function relating the defect magnitude of a first sample object defect to one object lifetime value, for the first sample object defect; and

computer-readable program code means for identifying an object lifetime distribution relating the defect magnitude of a second sample object defect to a plurality of object lifetime values, for the second sample object defect; and

wherein said fifth computer-readable program code means comprises;

computer-readable program code means for computing an object lifetime value from the defect magnitude of the at least one sample object defect according to the deterministic object lifetime function to thereby generate a sample object lifetime corresponding to the first sample object defect; and

computer-readable program code means for sampling the object lifetime relating to the defect magnitude of the second sample object defect to thereby generate a sample object lifetime corresponding to the second sample object defect.

31. A computer program product according to claim 24:

wherein said second computer-readable program code means comprises;

seventh computer-readable program code means for identifying cells from the circuit layout, each of the cells occupying an area of the circuit layout and having a reliability connectivity in the integrated circuit design;

eighth computer-readable program code means, responsive to said seventh computer-readable program code means, for identifying objects within each cell, each of the objects within a cell having a location in the cell and a reliability connectivity in the cell; and

wherein said third computer-readable program code means comprises computer-readable program code means for generating sample object defects for objects within the cells;

wherein said fifth computer-readable program code means comprises computer-readable program code means for generating sample object lifetimes for objects within the cells; and

wherein said sixth computer-readable program code means comprises;

ninth computer-readable program code means for generating sample cell lifetimes for cells from the sample object lifetimes for objects within the cells according to the reliability connectivity of the associated objects in the cells; and

tenth computer-readable program code means, responsive to said ninth computer-readable program code means, for generating a prediction of the reliability of integrated circuits produced by the integrated circuit fabrication process from the sample cell lifetimes according to the reliability connectivity of the associated cells in the integrated circuit design.

32. A method of predicting defect-related failures in integrated circuits produced by an integrated circuit fabrication process according to a circuit layout for an integrated circuit design, the circuit layout including a plurality of objects, each object having a location in the circuit layout and a reliability connectivity in the integrated circuit design, the method comprising the following steps which are performed in a data processing system:

generating sample object defects for the plurality of objects, each sample object defect representing a defect induced in an object by the integrated circuit fabrication process, each sample object defect having a defect magnitude associated therewith;

identifying a defect influence model for each sample object defect, each defect influence model relating the reliability connectivity of an object to the defect magnitude of a defect in the object;

generating a prediction of the reliability of instances of the plurality of objects having the sample object defects from the defect magnitudes associated with the sample object defects according to the corresponding identified defect influence models; and

generating a prediction of the reliability of integrated circuits produced by the integrated circuit fabrication process according to the circuit layout, from the prediction of the reliability of the instances of the plurality of objects according to the reliability connectivity of the associated objects in the integrated circuit design.

33. A method according to claim 32:

wherein said step of identifying a defect influence model comprises the step of identifying an accelerated life defect influence model relating the magnitude of a defect in an object to a lifetime for the object;

wherein said step of generating a prediction of the reliability of instances of the plurality of objects comprises the step of generating sample object lifetimes from the defect magnitudes associated with the sample object defects according to the corresponding identified accelerated life defect influence models; and

wherein said step of generating a prediction of the reliability of integrated circuits comprises the step of generating a prediction of the reliability of integrated circuits produced by the integrated circuit fabrication process according to the circuit layout from the sample object lifetimes according to the reliability connectivity of the associated objects in the integrated circuit design.

34. A method according to claim 32 wherein said step of identifying an accelerated life defect influence model comprises the step of identifying a log-linear regression model for at least one of the sample object defects, the log-linear regression model having the form

35. A computer apparatus for predicting defect-related failures in integrated circuits produced by an integrated circuit fabrication process according to a circuit layout for an integrated circuit design, the computer apparatus comprising:

means for storing the circuit layout;

means, responsive to said means for storing the circuit layout, for identifying objects in the circuit layout, each object having a location in the circuit layout and a reliability connectivity in the integrated circuit design;

means, responsive to said means for identifying objects, for generating sample object defects for the identified objects, each sample object defect representing a defect induced in an object by the integrated circuit fabrication process, each sample object defect having a defect magnitude associated therewith;

means, responsive to said means for generating sample object defects, for identifying an accelerated life defect influence model for each sample object defect, each accelerated life defect influence model relating the lifetime of an object to the defect magnitude of a defect in the object;

means, responsive to said means for generating sample object defects and to said means for identifying an accelerated life defect influence model, for generating sample object lifetimes from the defect magnitudes associated with the sample object defects according to the corresponding identified accelerate life defect influence models; and

means, responsive to said means for generating sample object lifetimes, for generating a prediction of the reliability of integrated circuits produced by the integrated circuit fabrication process according to the circuit layout from the sample object lifetimes according to the reliability connectivity of the associated objects in the integrated circuit design.

36. A computer apparatus according to claim 35 wherein said means for generating a prediction of the reliability of integrated circuits comprises:

means for generating sample integrated circuit lifetimes from the sample object lifetimes according to the reliability connectivity of the associated objects.

37. A computer apparatus according to claim 35 wherein said means for identifying an accelerated life defect influence model comprises means for identifying a log-linear regression model for at least one of the sample object defects, the log-linear regression model having the form

38. A computer apparatus according to claim 37:

wherein said means for identifying a log-linear regression model comprises means for identifying a deterministic object lifetime function relating the defect magnitude of a sample object defect to one object lifetime value; and

wherein said means for generating a sample object lifetime corresponding to the at least one sample object defect comprises means for computing a sample object lifetime corresponding to the at least one sample object defect from the defect magnitude of the at least one sample object defect according to the deterministic object lifetime function to thereby generate a sample object lifetime for the at least one sample object defect.

39. A computer apparatus according to claim 37:

wherein said means for identifying a log-linear regression model comprises means for identifying a log-linear object lifetime distribution relating the defect magnitude of a sample object defect to a plurality of object lifetime values; and

wherein said means for generating a sample object lifetime corresponding to the at least one sample object defect comprises means for sampling the log-linear object lifetime distribution to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

40. A computer apparatus according to claim 37:

wherein said means for identifying a log-linear regression model comprises means for identifying a predetermined generalized parametric object lifetime distribution function derived from intrinsic reliability data and having a distribution parameter related to the defect magnitude of a sample object defect by a distribution parameter function; and

wherein said means for generating a sample object lifetime corresponding to the at least one sample object defect comprises;

means for computing a value for the distribution parameter for the at least one sample object defect from the defect magnitude of the at least one sample object defect according to the distribution parameter function to thereby define a particularized object lifetime distribution function for the at least one sample object defect; and

means for sampling the particularized object lifetime distribution function to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

41. A computer apparatus according to claim 37:

wherein said means for identifying a log-linear regression model comprises means for identifying a predetermined generalized parametric object lifetime distribution function derived from intrinsic reliability data and having a distribution parameter related to the defect magnitude of a sample object defect and a conditional probability factor by a distribution parameter function, the conditional probability factor being independent of the predetermined generalized parametric object lifetime distribution function; and

wherein said means for generating a sample object lifetime corresponding to the at least one sample object defect comprises;

means for computing a value for the distribution parameter for the at least one sample object defect from the defect magnitude of the at least one sample object defect and the conditional probability factor according to the distribution parameter function to thereby define a particularized object lifetime distribution function for the at least one sample object defect; and

means for sampling the particularized object lifetime distribution function to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

42. A computer apparatus according to claim 35:

wherein said means for identifying an accelerated life defect influence model for each sample object defect comprises;

means for identifying a deterministic object lifetime function relating the defect magnitude of a first sample object defect to one object lifetime value, for the first sample object defect; and

means for identifying an object lifetime distribution relating the defect magnitude of a second sample object defect to a plurality of object lifetime values, for the second sample object defect; and

wherein said means for generating sample object lifetimes comprises;

means for computing an object lifetime value from the defect magnitude of the at least one sample object defect according to the deterministic object lifetime function to thereby generate a sample object lifetime corresponding to the first sample object defect; and

means for sampling the object lifetime relating to the defect magnitude of the second sample object defect to thereby generate a sample object lifetime corresponding to the second sample object defect.

43. A computer apparatus according to claim 35:

wherein said means for identifying objects from the circuit layout comprises;

means for identifying cells from the circuit layout, each of the cells occupying an area of the circuit layout and having a reliability connectivity in the integrated circuit design; and

means, responsive to said means for identifying cells, for identifying objects within each cell, each of the objects within a cell having a location in the cell and a reliability connectivity in the cell;

wherein said means for generating sample object defects comprises means for generating sample object defects for objects within the cells;

wherein said means for generating sample object lifetimes comprises means for generating sample object lifetimes for objects within the cells; and

wherein said means for generating a prediction of the reliability of integrated circuits comprises;

means for generating sample cell lifetimes for cells from the sample object lifetimes for objects within the cells according to the reliability connectivity of the associated objects in the cells; and

means, responsive to said means for generating sample cell lifetimes, for generating a prediction of the reliability of integrated circuits produced by the integrated circuit fabrication process from the sample cell lifetimes according to the reliability connectivity of the associated cells in the integrated circuit design.

44. A computer apparatus for predicting defect-related failures in integrated circuits produced by an integrated circuit fabrication process according to a circuit layout for an integrated circuit design, the apparatus comprising:

means for storing the circuit layout;

means, responsive to said means for storing the circuit layout, for identifying objects in the circuit layout, each object having an object type, a location and a reliability connectivity in the integrated circuit design;

means for storing a process flaw distribution associated with the integrated circuit fabrication process, the process flaw distribution representing a statistical distribution of flaws produced by the integrated circuit fabrication process;

means, responsive to said means for storing a process flaw distribution, for generating sample flaws according to the process flaw distribution, each of the sample flaws having a location with respect to the circuit layout;

means, responsive to said means for identifying objects and to said means for generating sample flaws, for correlating the locations of the sample flaws with the locations of the identified objects to thereby generate sample object defects, each sample object defect representing a defect induced in an object by a sample flaw, each sample object defect having a defect magnitude associated therewith;

means, responsive to said means for generating sample object defects, for identifying an accelerated life defect influence model for each sample object defect, each accelerated life defect influence model representing an influence of the defect magnitude of a defect on the lifetime of an object;

means, responsive to said means for generating sample object defects and to said means for identifying a log-linear accelerated life defect influence model, for generating sample object lifetimes for the identified objects from the defect magnitudes associated with sample object defects according to the corresponding identified accelerated life defect influence models; and

means, responsive to said means for generating sample object lifetimes, for generating a prediction of the reliability of integrated circuits produced by the integrated circuit fabrication process according to the circuit layout, from the sample object lifetimes according to the reliability connectivity of the associated identified objects in the integrated circuit design.

45. A computer apparatus according to claim 44 wherein said means for generating sample flaws comprises means for sampling the process flaw distribution.

46. A computer apparatus according to claim 44 wherein the process flaw distribution comprises at least at least one of a process flaw magnitude distribution, a process flaw spatial distribution, and a process flaw quantity distribution.

47. A computer apparatus according to claim 44 wherein the process flaw distribution comprises a process flaw magnitude distribution, a process flaw spatial distribution, and a process flaw quantity distribution.

48. A computer apparatus according to claim 44:

wherein said means for identifying objects comprises means for identifying objects having a critical area associated therewith;

wherein said means for generating sample flaws comprises means for generating sample flaws having a location and an area; and

wherein said means for correlating the locations of sample flaws with the locations of the identified objects comprises means for generating a sample object defect if a critical area of an object intersects the area of a sample flaw.

49. A computer apparatus according to claim 44 wherein said means for identifying an accelerated life defect influence model comprises means for identifying a log-linear regression model for at least one of the sample object defects, the log-linear regression model having the form

50. A computer apparatus according to claim 49:

wherein said means for identifying a log-linear regression model comprises means for identifying a deterministic object lifetime function relating the defect magnitude of a sample object defect to one object lifetime value; and

wherein said means for generating a sample object lifetime corresponding to the at least one sample object defect comprises means for computing a sample object lifetime corresponding to the at least one sample object defect from the defect magnitude of the at least one sample object defect according to the deterministic object lifetime function to thereby generate a sample object lifetime for the at least one sample object defect.

51. A computer apparatus according to claim 49:

wherein said means for identifying a log-linear regression model comprises means for identifying a log-linear object lifetime distribution relating the defect magnitude of a sample object defect to a plurality of object lifetime values; and

wherein said means for generating a sample object lifetime corresponding to the at least one sample object defect comprises means for sampling the log-linear object lifetime distribution to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

52. A computer apparatus according to claim 49:

wherein said means for identifying a log-linear regression model comprises means for identifying a predetermined generalized parametric object lifetime distribution function derived from intrinsic reliability data and having a distribution parameter related to the defect magnitude of a sample object defect by a distribution parameter function; and

wherein said means for generating a sample object lifetime corresponding to the at least one sample object defect comprises;

means for computing a value for the distribution parameter for the at least one sample object defect from the defect magnitude of the at least one sample object defect according to the distribution parameter function to thereby define a particularized object lifetime distribution function for the at least one sample object defect; and

means for sampling the particularized object lifetime distribution function to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

53. A computer apparatus according to claim 49:

wherein said means for identifying a log-linear regression model comprises means for identifying a predetermined generalized parametric object lifetime distribution function derived from intrinsic reliability data and having a distribution parameter related to the defect magnitude of a sample object defect and a conditional probability factor by a distribution parameter function, the conditional probability factor being independent of the predetermined generalized parametric object lifetime distribution factor; and

wherein said means for generating a sample object lifetime corresponding to the at least one sample object defect comprises;

means for computing a value for the distribution parameter for the at least one sample object defect from the defect magnitude of the at least one sample object defect and the conditional probability factor according to the distribution parameter function to thereby define a particularized object lifetime distribution function for the at least one sample object defect; and

means for sampling the particularized object lifetime distribution function to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

54. A computer apparatus according to claim 44:

wherein said means for identifying an accelerated life defect influence model for each sample object defect comprises;

means for identifying a deterministic object lifetime function relating the defect magnitude of a first sample object defect to one object lifetime value, for the first sample object defect; and

means for identifying an object lifetime distribution relating the defect magnitude of a second sample object defect to a plurality of object lifetime values, for the second sample object defect; and

wherein said means for generating sample object lifetimes comprises;

means for computing an object lifetime value from the defect magnitude of the at least one sample object defect according to the deterministic object lifetime function to thereby generate a sample object lifetime corresponding to the first sample object defect; and

means for sampling the object lifetime relating to the defect magnitude of the second sample object defect to thereby generate a sample object lifetime corresponding to the second sample object defect.

55. A computer apparatus according to claim 44:

wherein said means for identifying objects from the circuit layout comprises;

means for identifying cells from the circuit layout, each of the cells occupying an area of the circuit layout and having a reliability connectivity in the integrated circuit design; and

means, responsive to said means for identifying cells, for identifying objects within each cell, each of the objects within a cell having a location in the cell and a reliability connectivity in the cell;

wherein said means for generating sample object defects comprises means for generating sample object defects for objects within the cells;

wherein said means for generating sample object lifetimes comprises means for generating sample object lifetimes for objects within the cells; and

wherein said means for generating a prediction of the reliability of integrated circuits comprises;

means for generating sample cell lifetimes for cells from the sample object lifetimes for objects within the cells according to the reliability connectivity of the associated objects in the cells; and

means, responsive to said means for generating sample cell lifetimes, for generating a prediction of the reliability of integrated circuits produced by the integrated circuit fabrication process from the sample cell lifetimes according to the reliability connectivity of the associated cells in the integrated circuit design.

59. A computer program product according to claim 51, wherein the log-linear regression model comprises a log-linear object lifetime distribution relating the defect magnitude of a sample object defect to a plurality of object lifetime values, and wherein said second computer-readable program code means comprises:

computer-readable program code means for sampling the log-linear object lifetime distribution to thereby generate a sample object lifetime corresponding to the sample object defect.

56. A computer program product for use with a computer-based reliability simulator to produce a sample object lifetime for an object in a circuit layout of an integrated circuit design for a sample object defect in the object according to an accelerated life defect influence model for the sample object defect, the sample object defect having a defect magnitude, the accelerated life defect influence model relating the lifetime of the object to the defect magnitude of the sample object defect, the computer program product comprising:

a computer-readable storage medium having computer-readable program code means embodied in said medium, said computer-readable program code means comprising;

computer-readable program code means for generating a sample object lifetime from the stored defect magnitude according to the accelerated life defect influence model.

57. A computer program product according to claim 56 wherein the accelerated life defect influence model is a log-linear regression model having the form

58. A computer program product according to claim 57, wherein the log-linear regression model comprises a deterministic object lifetime function relating the defect magnitude of a sample object defect to one object lifetime value, and wherein said second computer-readable program code means comprises:

computer-readable program code means for computing a sample object lifetime corresponding to the sample object defect from the defect magnitude of the sample object defect according to the deterministic object lifetime function to thereby generate a sample object lifetime for the sample object defect.

60. A computer program product according to claim 57, wherein the log-linear regression model comprises a predetermined generalized parametric object lifetime distribution function derived from intrinsic reliability data and having a distribution parameter related to the defect magnitude of a sample object defect by a distribution parameter function, and wherein said second computer-readable program code means comprises:

computer-readable program code means for computing a value for the distribution parameter for the sample object defect from the defect magnitude of the sample object defect according to the distribution parameter function to thereby define a particularized object lifetime distribution function for the sample object defect; and

computer-readable program code means for sampling the particularized object lifetime distribution function to thereby generate a sample object lifetime corresponding to the sample object defect.

61. A computer program product according to claim 57, wherein the log-linear regression model is a predetermined generalized parametric object lifetime distribution function derived from intrinsic reliability data and having a distribution parameter related to the defect magnitude of a sample object defect and a conditional probability factor by a distribution parameter function, the conditional probability factor being independent of the predetermined generalized object lifetime distribution function;

and wherein said second computer-readable program code means comprises;

computer-readable program code means for computing a value for the distribution parameter for the sample object defect from the defect magnitude of the sample object defect and the conditional probability factor according to the distribution parameter function to thereby define a particularized object lifetime distribution function for the sample object defect; and

computer-readable program code means for sampling the particularized object lifetime distribution function to thereby generate a sample object lifetime corresponding to the sample object defect.

62. A computer program product, comprising:

a computer-readable storage medium having computer-readable program code means embodied in said medium, said computer-readable program code means comprising;

first computer-readable program code means for storing a circuit layout for an integrated circuit design;

second computer-readable program code means, responsive to said first computer-readable program code means, for identifying objects in the circuit layout, each object having a location in the circuit layout and a reliability connectivity in the integrated circuit design;

third computer-readable program code means for storing a process flaw distribution associated with the integrated circuit fabrication process, the process flaw distribution representing a statistical distribution of flaws produced by the integrated circuit fabrication process;

fourth computer-readable program code means, responsive to said third computer-readable program code means, for generating sample flaws according to the process flaw distribution, each of the sample flaws having a location with respect to the circuit layout;

fifth computer-readable program code means, responsive to said second and fourth computer-readable program code means, for correlating the locations of the sample flaws with the locations of the identified objects to thereby generate sample object defects, each sample object defect representing a defect induced in an object by an integrated circuit fabrication process producing integrated circuits according to the circuit layout, each sample object defect having a defect magnitude associated therewith;

sixth computer-readable program code means, responsive to said fifth computer-readable computer program code means, for identifying an accelerated life defect influence model for each sample object defect, each accelerated life defect influence model relating the lifetime of an object to the defect magnitude of a defect in the object;

seventh computer-readable program code means, responsive to said fifth and sixth computer-readable program code means, for generating sample object lifetimes from the defect magnitudes associated with the sample object defects according to the corresponding identified accelerated life defect influence models; and

eighth computer-readable program code means, responsive to seventy computer-readable program code means, for generating a prediction of the reliability of integrated circuits produced by the integrated circuit fabrication process according to the circuit layout from the sample object lifetimes according to the reliability connectivity of the associated objects in the integrated circuit design.

63. A computer program product according to claim 62:

wherein said sixth computer-readable program code means comprises;

computer-readable program code means for identifying a deterministic object lifetime function relating the defect magnitude of a first sample object defect to one object lifetime value, for the first sample object defect; and

computer-readable program code means for identifying an object lifetime distribution relating the defect magnitude of a second sample object defect to a plurality of object lifetime values, for the second sample object defect; and

wherein said seventh computer-readable program code means comprises;

computer-readable program code means for computing an object lifetime value from the defect magnitude of the at least one sample object defect according to the deterministic object lifetime function to thereby generate a sample object lifetime corresponding to the first sample object defect; and

computer-readable program code means for sampling the object lifetime relating to the defect magnitude of the second sample object defect to thereby generate a sample object lifetime corresponding to the second sample object defect.

64. A computer program product according to claim 62:

wherein said second computer-readable program code means comprises;

ninth computer-readable program code means for identifying cells from the circuit layout, each of the cells occupying an area of the circuit layout and having a reliability connectivity in the integrated circuit design;

tenth computer-readable program code means, responsive to said seventh computer-readable program code means, for identifying objects within each cell, each of the objects within a cell having a location in the cell and a reliability connectivity in the cell; and

wherein said fifth computer-readable program code means comprises computer-readable program code means for generating sample object defects for objects within the cells;

wherein said seventh computer-readable program code means comprises computer-readable program code means for generating sample object lifetimes for objects within the cells; and

wherein said eighth computer-readable program code means comprises;

eleventh computer-readable program code means for generating sample cell lifetimes for cells from the sample object lifetimes for objects within the cells according to the reliability connectivity of the associated objects in the cells; and

twelfth computer-readable program code means, responsive to said eleventh computer-readable program code means, for generating a prediction of the reliability of integrated circuits produced by the integrated circuit fabrication process from the sample cell lifetimes according to the reliability connectivity of the associated cells in the integrated circuit design.

65. A computer program product according to claim 62 wherein said fourth computer-readable program code means comprises computer-readable program code means for sampling the process flaw distribution.

66. A computer program product according to claim 62 wherein the process flaw distribution comprises at least at least one of a process flaw magnitude distribution, a process flaw spatial distribution, and a process flaw quantity distribution.

67. A computer program product according to claim 62 wherein the process flaw distribution comprises a process flaw magnitude distribution, a process flaw spatial distribution, and a process flaw quantity distribution.

68. A computer program product according to claim 62:

wherein said second computer-readable program code means comprises computer-readable program code means for identifying objects having a critical area associated therewith;

wherein said fourth computer-readable program code means comprises computer-readable program code means for generating sample flaws having a location and an area; and

wherein said fifth computer-readable program code means comprises computer-readable program code means for generating a sample object defect if a critical area of an object intersects the area of a sample flaw.

69. A computer program product according to claim 62 wherein said sixth computer-readable program code means comprises ninth computer-readable program code means for identifying a log-linear regression model for at least one of the sample object defects, the log-linear regression model having the form

70. A computer program product according to claim 69:

wherein said ninth computer-readable program code means comprises computer-readable program code means for identifying a deterministic object lifetime function relating the defect magnitude of a sample object defect to one object lifetime value; and

wherein said tenth computer-readable program code means comprises computer-readable program code means for computing a sample object lifetime corresponding to the at least one sample object defect from the defect magnitude of the at least one sample object defect according to the deterministic object lifetime function to thereby generate a sample object lifetime for the at least one sample object defect.

71. A computer program product according to claim 69:

wherein said ninth computer-readable program code means comprises computer-readable program code means for identifying a log-linear object lifetime distribution relating the defect magnitude of a sample object defect to a plurality of object lifetime values; and

wherein said tenth computer-readable program code means comprises computer-readable program code means for sampling the log-linear object lifetime distribution to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

72. A computer program product according to claim 69:

wherein said ninth computer-readable program code means comprises computer-readable program code means for identifying a predetermined generalized parametric object lifetime distribution function derived from intrinsic reliability data and having a distribution parameter related to the defect magnitude of a sample object defect by a distribution parameter function; and

wherein said tenth computer-readable program code means comprises;

computer-readable program code means for computing a value for the distribution parameter for the at least one sample object defect from the defect magnitude of the at least one sample object defect according to the distribution parameter function to thereby define a particularized object lifetime distribution function for the at least one sample object defect; and

computer-readable program code means for sampling the particularized object lifetime distribution function to thereby generate a sample object lifetime corresponding to the at least one sample object defect.

73. A computer program product according to claim 69:

wherein said ninth computer-readable program code means comprises computer-readable program code means for identifying a predetermined generalized parametric object lifetime distribution function derived from intrinsic reliability data and having a distribution parameter related to the defect magnitude of a sample object defect and a conditional probability factor by a distribution parameter function, the conditional probability factor being independent of the predetermined generalized parametric object lifetime distribution function; and

wherein said tenth computer-readable program code means comprises;

computer-readable program code means for computing a value for the distribution parameter for the at least one sample object defect from the defect magnitude of the at least one sample object defect and the conditional probability factor according to the distribution parameter function to thereby define a particularized object lifetime distribution function for the at least one sample object defect; and

computer-readable program code means for sampling the particularized object lifetime distribution function to thereby generate a sample object lifetime corresponding to the at least one sample object defect.