Abstract: A mechanism is provided for post timing layout modification for performance. The mechanism selectively applies layout modification based on timing analysis at the path level. The mechanism applies stress only to transistors that are in a setup critical path without applying stress to transistors in hold critical paths. The mechanism may use a method to apply stress to improve performance of a transistor in a setup critical path, as long as the stress does not also improve performance of a neighboring transistor in a hold critical path. In some instances, the mechanism may apply stress to improve performance of a transistor in a setup critical path while simultaneously degrading performance of a transistor in a hold critical path.

Abstract: A mechanism is provided for post timing layout modification for performance. The mechanism selectively applies layout modification based on timing analysis at the path level. The mechanism applies stress only to transistors that are in a setup critical path without applying stress to transistors in hold critical paths. The mechanism may use a method to apply stress to improve performance of a transistor in a setup critical path, as long as the stress does not also improve performance of a neighboring transistor in a hold critical path. In some instances, the mechanism may apply stress to improve performance of a transistor in a setup critical path while simultaneously degrading performance of a transistor in a hold critical path.

Abstract: A method of routing an interconnect metal layer of an integrated circuit, wherein single-width nets are replicated and routed in parallel to reduce the total resistance on the net; wide wires are decomposed into a several single-width wires routed in parallel to improve uniformity of metal interconnect routing and therefore manufacturability of metal interconnect layers. The decomposition step is performed during a preliminary wire route after initial physical placement. Access to pin shapes is ensured through a branching and a recombination of the parallel single-width wires. Separate wire segments are rejoined at the source and sink of the net. The parallel wire segments do not change the logic behavior of the circuit.

Abstract: A method of routing an interconnect metal layer of an integrated circuit, wherein single-width nets are replicated and routed in parallel to reduce the total resistance on the net; wide wires are decomposed into a several single-width wires routed in parallel to improve uniformity of metal interconnect routing and therefore manufacturability of metal interconnect layers. The decomposition step is performed during a preliminary wire route after initial physical placement. Access to pin shapes is ensured through a branching and a recombination of the parallel single-width wires. Separate wire segments are rejoined at the source and sink of the net. The parallel wire segments do not change the logic behavior of the circuit.