Semiconductor chip repair by stacking of a base semiconductor chip and a repair semiconductor chip
In one aspect, a method of enhancing semiconductor chip process variability and lifetime reliability through a three-dimensional (3D) integration applied to electronic packaging is disclosed. Also provided is an arrangement for implementing the inventive method. In another aspect, a method and on-chip controller are disclosed for enhancing semiconductor chip process variability and lifetime reliability through a three-dimensional (3D) integration applied to electronic packaging. Also provided is an on-chip reliability/variability controller arrangement for implementing the inventive method. In yet another aspect, base semiconductor chips, each comprising a plurality of chiplets, are manufactured and tested. For a base semiconductor chip having at least one non-functional chiplet, at least one repair semiconductor chiplet is vertically stacked. A functional multi-chip assembly is formed, which provides the same functionality as a base semiconductor chip in which all chiplets are functional.
1. A semiconductor system configured to enhance reliability or reduce performance variability comprising:
a primary layer including a first semiconductor chip that is a microprocessor chip, said microprocessor chip including at least one first critical block; and
an auxiliary layer including a second semiconductor chip attached to said first semiconductor chip through 3D integration, said second semiconductor chip including an on-chip reliability/variability controller and a set of configurable resources including at least one second critical block that is an exact replica of said at least one first critical block, wherein said on-chip reliability/variability controller includes:
a first means for detecting a need for replacing, or enhancing performance of, a resource in said first semiconductor chip;
a second means for scanning a pool of existing execution resources or memory resources to find an eligible replacement resource for, or an eligible performance-enhancing resource for, said resource in said first semiconductor chip;
a third means for configuring at least one element of said set of configurable resources to replace, or to enhance performance of, said resource in said first semiconductor chip;
at least one general logic block including a look-up table embodying preprogrammed recovery schemes for various faults.
2. The semiconductor system of claim 1 , further comprising a fourth means for deactivating or disconnecting at least one elements of said set of configurable resources.
3. The semiconductor system of claim 1 , wherein said set of configurable resources includes at least one of configurable memory resources, configurable memory resources, and configurable or programmable logic blocks.
4. The semiconductor system of claim 1 , wherein said third means includes a configurable interconnect and switch box.
5. The semiconductor system of claim 4 , wherein one of said at least one first critical block is located vertically on top of, or vertically below, one of said at least one second critical block.
6. The semiconductor system of claim 5 , wherein said configurable interconnect and switch box connects and disconnects said replicas of critical blocks.
7. The semiconductor system of claim 1 , wherein said on-chip reliability/variability controller is configured to perform regular checks on existing hardware in order to detect potential faults.
8. The semiconductor system of claim 1 , wherein each recovery scheme among said preprogrammed recovery schemes indicates how to recover from specific faults using existing redundancy in said auxiliary layer.
9. The semiconductor system of claim 8 , wherein said preprogrammed recovery schemes include a program for:
determining if an exact replica of said at least one critical block is available within said auxiliary layer; and
instructing, if no exact replica is available, said on-chip reliability/variability controller to use configurable hardware blocks to emulate a desired functionality.
10. The semiconductor system of claim 1 , wherein said preprogrammed recovery schemes compensate for changes in a configurable logic timing by adjusting a clock frequency in said primary layer and a clock frequency in said auxiliary layer.
11. The semiconductor system of claim 1 , wherein said on-chip reliability/variability controller is configured to select from said preprogrammed recovery schemes based on at least one condition selected from a power overhead of a recovery scheme, a current power saving mode, a frequency target for said primary and auxiliary layers, severity of fault, and current workload demand.
12. The semiconductor system of claim 1 , wherein said preprogrammed recovery schemes include information for replica availability, existence of multiple recovery options, specification of blocks to be used in recovery, target clock frequency during recovery, rerouting path for connecting redundancies, and power dissipation overhead for each preprogrammed recovery scheme.
13. A method of enhancing reliability or reducing performance variability of a semiconductor system, said method comprising:
attaching an auxiliary layer to a primary layer, wherein said primary layer includes a first semiconductor chip that is a microprocessor chip, said microprocessor chip including at least one first critical block, and said auxiliary layer includes a second semiconductor chip attached to said first semiconductor chip through 3D integration, said second semiconductor chip including an on-chip reliability/variability controller and a set of configurable resources including at least one second critical block that is an exact replica of said at least one first critical block;
detecting a need for replacing, or enhancing performance of, a resource in said first semiconductor chip;
scanning a pool of existing execution resources or memory resources to find an eligible replacement resource for, or an eligible performance-enhancing resource for, said resource in said first semiconductor chip; and
configuring at least one element of said set of configurable resources that replaces, or enhances performance of, said resource in said first semiconductor chip according to a look-up table in at least one general logic block, said look-up table embodying preprogrammed recovery schemes for various faults.
14. The method of claim 13 , further comprising deactivating or disconnecting at least one elements of said set of configurable resources.
15. The method of claim 13 , wherein said set of configurable resources includes at least one of configurable memory resources, configurable memory resources, and configurable or programmable logic blocks.
16. The method of claim 13 , wherein a configurable interconnect and switch box located in said second semiconductor chip configures said at least one element of said set of configurable resources.
17. The method of claim 16 , wherein said auxiliary layer and said primary layer are attached so that one of said at least one first critical block is located vertically on top of, or vertically below, one of said at least one second critical block.
18. The method of claim 17 , wherein said configurable interconnect and switch box connects and disconnects said replicas of critical blocks.
19. The method of claim 13 , wherein said on-chip reliability/variability controller performs regular checks on existing hardware in order to detect potential faults.
20. The method of claim 13 , wherein each recovery scheme among said preprogrammed recovery schemes indicates how to recover from specific faults using existing redundancy in said auxiliary layer.
21. The method of claim 20 , further comprising:
determining if an exact replica of said at least one critical block is available within said auxiliary layer; and
instructing, if no exact replica is available, said on-chip reliability/variability controller to use configurable hardware blocks to emulate a desired functionality.
22. The method of claim 13 , wherein said preprogrammed recovery schemes compensate for changes in a configurable logic timing by adjusting a clock frequency in said primary layer and a clock frequency in said auxiliary layer.
23. The method of claim 13 , wherein said on-chip reliability/variability controller is configured to select from said preprogrammed recovery schemes based on at least one condition selected from a power overhead of a recovery scheme, a current power saving mode, a frequency target for said primary and auxiliary layers, severity of fault, and current workload demand.
24. The method of claim 13 , wherein said preprogrammed recovery schemes include information for replica availability, existence of multiple recovery options, specification of blocks to be used in recovery, target clock frequency during recovery, rerouting path for connecting redundancies, and power dissipation overhead for each preprogrammed recovery scheme.