Sensitivity-based complex statistical modeling for random on-chip variation
The invention provides a method for performing statistical static timing analysis using a novel on-chip variation model, referred to as Sensitivity-based Complex Statistical On-Chip Variation (SCS-OCV). SCS-OCV introduces complex variation concept to resolve the blocking technical issue of combining local random variations, enabling accurate calculation of statistical variations with correlations, such as common-path pessimism removal (CPPR). SCS-OCV proposes practical statistical min/max operations for random variations that can guarantee pessimism at nominal and targeted N-sigma corner, and extends the method to handle complex variations, enabling graph-based full arrival/required time propagation under variable compaction. SCS-OCV provides a statistical corner evaluation method for complex random variables that can transform vector-based parametric timing information to the single-value corner-based timing report, and based on the method derives equations to bridge POCV/SSTA with LOCV. This significantly reduces the learning curve and increases the usage of the technology, being more easily adopted by the industry.
1. A computer-implemented method of statistical static timing analysis (SSTA) comprising:
receiving, by a computer, information describing a circuit, the information comprising:
a first input node, a second input node, and an output node, such that there is
a first path from the first input node to the output node, and
a second path from the second input node to the output node,
the first path and the second path converging at the output node,
each path associated with a parametric delay represented as a nominal delay value and a standard deviation value, the standard deviation value representing a timing impact of local random variation;
performing statistical static timing analysis (SSTA) based on on-chip variation (OCV) model, the SSTA comprising, determining a parametric delay at the output node based on a statistical maximum of parametric delay through the first path and parametric delay through the second path, wherein the statistical maximum preserves N sigma corner delay values, and determining the statistical maximum comprises:
determining a nominal delay value of the parametric delay at the output node based on a maximum of:
nominal delay value of the parametric delay through the first path, and
nominal delay of the parametric delay through the second path; and
determining a standard deviation value of the parametric delay at the output node, comprising:
determining a first value as a maximum of:
a weighted sum of nominal delay value and standard deviation value of the parametric delay through the first path, and
a weighted sum of nominal delay value and standard deviation value of the parametric delay through the second path;
determining a second value as a maximum of:
the nominal delay value of the parametric delay through the first path, and
the nominal delay value of the parametric delay through the second path; and
determining the difference between the first value and the second value; and
standard deviation value of the parametric delay through the first path, and
standard deviation value of the parametric delay through the second path; and
storing the nominal delay and the standard deviation value of the parametric delay for the output node.
2. The method of claim 1 , wherein responsive to a request to determine the parametric delay for the output node for an N sigma corner:
the weighted sum of nominal delay value and standard deviation value is based on the value of N.
3. The method of claim 1 , wherein responsive to a request to determine the parametric delay for the output node for an N sigma corner the weight of the standard deviation value used for computing the weighted sum for each path is based on the value of N.
4. The method of claim 1 , further comprising:
determining an arrival time for the output node based on the parametric delay of the output node.
5. The method of claim 1 , further comprising:
determining a required time for the output node based on the parametric delay of the output node.
6. The method of claim 1 , further comprising:
determining a slack time for the output node based on the parametric delay of the output node.
7. The method of claim 1 , further comprising using the parametric delay for the output node determined based on the statistical maximum in a path-based timing analysis of the circuit.
8. The method of claim 1 , further comprising using the parametric delay for the output node determined based on the statistical maximum in a graph-based full chip timing analysis.