Fujitsu Corporation

Zenasis Technologies, Inc.

Rajeev Murgai

Subodh M Reddy

William Warren Walker

Takashi Miyoshi

Takeshi Horie

Debashis Bhattacharya

Vamsi Boppana

Hongyu Chen

Rabindra Roy

Gustavo R Wilke

Yinghua Li

Mehdi B Tahoori

In one embodiment, a method for computing jitter in a clock tree includes dividing a clock tree into a plurality of stages and computing jitter in one or more of the stages according to a model of at least a portion of a circuit associated with the clock tree. The model includes a representation of each source of jitter in the circuit. The method also includes, to compute jitter associated with a path or a pair of paths in the clock tree, statistically combining the jitter in each of the stages of the path or the pair of paths in the clock tree with each other. In one embodiment, to efficiently compute jitter and to achieve zero clock skew, a method synthesizes a symmetrical clock tree of a circuit in which corresponding stages in all paths from a root of the clock tree to sinks of the clock tree exhibit approximate electrical equivalence to each other.

Estimating jitter in a clock tree of a circuit and synthesizing a jitter-aware and skew-aware clock tree

A method of analyzing timing uncertainty involves creating an accurate model of one or more circuit elements of a mesh circuit residing within a window that covers a subset of the mesh circuit. An approximate model of one or more circuit elements of the mesh circuit residing outside of the window is also created. Monte Carlo simulations are performed on the combination of the accurate model and the approximate model to determine a plurality of timing values, wherein each run of the Monte Carlo simulation varies one or more parameters potentially affecting the operation of the mesh circuit. An uncertainty associated with the circuit elements is determined, based at least in part on the plurality of timing values. One embodiment considers clock as the signal whose timing uncertainty can be determined. Other embodiments model and simulate the global drive circuit that drives the mesh circuit separately from the mesh circuit to take into account common path correlations in the drive circuit.

Analyzing timing uncertainty in mesh-based architectures

A method is provided and includes accessing a description of a chip, which includes sequential elements and a clock mesh. Items used include: the description of the chip, the information for modeling the sequential elements and interconnections, and the set of parameters of the clock mesh. Additionally, the method includes determining a plurality of original window locations covering the clock mesh. Further, for each original window location, the method includes expanding the original window location in one or more directions to generate a larger window location; generating a mesh simulation model inside the larger window location; simulating the mesh simulation model; measuring clock timing for the sequential elements in the original window location based on the simulation of the mesh simulation model; and collecting timing information on all the sequential elements on the chip based on the measured clock timing for the sequential elements in the original window locations.

System and method for providing an improved sliding window scheme for clock mesh analysis

In one embodiment, a method includes accessing a description of a chip including multiple sequential elements and a clock mesh, information for modeling the sequential elements and interconnections, and a set of parameters of the clock mesh. The method also includes, using the description of the chip, the information for modeling the sequential elements and interconnections, and the set of parameters of the clock mesh, determining multiple original window locations covering the clock mesh. Each window location includes one or more of the sequential elements on the chip. The method also includes, for each original window location, expanding the original window location in one or more directions to generate a larger window location and generating a mesh simulation model including a detailed model inside the larger window location and an approximate model outside the larger window location, simulating the mesh simulation model, and measuring clock timing for the sequential elements in the window location based on the mesh simulation model. The method also includes collecting timing information on the sequential elements on the chip based on the measured clock timing for the sequential elements in the original window locations.

Border-enhanced sliding window scheme (SWS) for determining clock timing in a mesh-based clock architecture

In one embodiment, a method includes accessing a description of a chip including multiple sequential elements and a clock mesh, information for modeling the sequential elements and interconnections, and a set of parameters of the clock mesh. The method also includes, using the description of the chip, the information for modeling the sequential elements and interconnections, and the set of parameters of the clock mesh, determining multiple window locations covering the clock mesh. Each window location includes one or more of the sequential elements on the chip. The method also includes, for each window location, generating a mesh simulation model including a detailed model inside the window location and an approximate model outside the window location, simulating the mesh simulation model, and measuring clock timing for the sequential elements in the window location based on the mesh simulation model. The method also includes collecting timing information on the sequential elements on the chip based on the measured clock timing for the sequential elements in the window locations.

Sliding window scheme (SWS) for determining clock timing in a mesh-based clock architecture

In one embodiment, a method for crosstalk-aware timing analysis includes accessing a design of a circuit and identifying critical paths in the design. Each critical path includes one or more victim interconnects and one or more cells. The method includes identifying potential aggressor interconnects associated with each victim interconnect and, for each victim interconnect, extracting parasitics of the victim interconnect and the potential aggressor interconnects associated with the victim interconnect. The method includes computing timing windows of the potential aggressor interconnects and computing a first timing of each cell and each victim interconnect on each critical path. The method also includes, for each critical path, generating timing waveforms of the potential aggressor interconnects, traversing the critical path from a start point on the critical path to an end point on the critical path, and, computing a second timing of each cell and each victim interconnect on the critical path according to a traversal of the critical path.

Crosstalk-aware timing analysis

In one embodiment, a method for computing current in a digital circuit based on an accurate current model for library cells includes accessing a cell library, for each cell in the cell library corresponding to a cell in a digital circuit, generating multiple waveforms of current drawn by the cell from a power supply according to one or more predetermined values of one or more input parameters of the cell, analyzing the digital circuit to determine one or more actual values of the input parameters of each cell in the digital circuit, for each of the cells in the digital circuit, generating a current waveform according to the determined actual values of the input parameters and a waveform of current drawn by the cell from the power supply generated by the characterization module corresponding to the determined actual values of the input parameters, and summing the current waveforms of the cells in the digital circuit to generate a waveform of current drawn by the digital circuit from the power supply for use in a pattern-dependent methodology (PDM) or a patten-independent methodology (PIM).

Computing current in a digital circuit based on an accurate current model for library cells

In one embodiment, a method for analyzing substrate noise includes applying a static timing analysis (STA) algorithm to a description of a digital circuit. Application of the STA algorithm generates timing information on one or more gates in the digital circuit. The method also includes applying a current waveform generation (CWG) algorithm to the description of the digital circuit, the timing information on one or more gates in the digital circuit, and a description of switching activity in the digital circuit. Application of the CWG algorithm generates a current waveform. The method also includes generating a reduced model (RM) of the digital circuit for simulation according to the description of the digital circuit, the current waveform, and a model of a package associated with the digital circuit. Simulation of the RM of the digital circuit generates an indication of noise in a substrate associated with the digital circuit.

Analyzing substrate noise

In one embodiment, a method for layout-driven, area-constrained design optimization includes accessing a design and a layout of the design. The design includes one or more gates and one or more nets coupling the gates to each other. The layout includes blocks that partition a chip area of the design. Each block includes one or more of the gates. The layout also includes a global routing of the nets. The method also includes performing a first timing analysis of the design and the layout and updating the design and the layout. The method also includes performing a second timing analysis of the design and the layout. The second timing analysis takes into account the updates to the design and the layout. The method also includes, if one or more results of the second timing analysis indicate that the design does not meet one or more predetermined design goals and indicate at least a predetermined amount of progress toward one or more of the design goals relative to the one or more results of the first timing analysis, further updating the design and the layout.

Layout-driven, area-constrained design optimization

The present invention pertains to an automated method for designing a integrated circuit (IC) design-specific cell, the method includes the steps of receiving a design specification for the design-specific cell, mapping a transistor-level representation of the design-specific cell, wherein the mapping is based on at least one, but perhaps plural design specifications, and evaluating the transistor-level representation of the design-specific cell for satisfaction of the design specification.

Process for automated generation of design-specific complex functional blocks to improve quality of synthesized digital integrated circuits in CMOS using altering process

Growing community of inventors

Santa Clara, CA, United States of America

Rajeev Murgai