International Business Machines Corporation

Hemlata Gupta

Kerim Kalafala

David James Hathaway

Ronald Dennis Rose

Manish Verma

Naiju K Abdul

Debjit Sinha

Alex Rubin

Chandramouli Visweswariah

James Douglas Warnock

Stephen Gerard Shuma

Michael H Wood

Vasant B Rao

Adil Bhanji

Jennifer E Basile

Jeremy J Leitzen

Jeffrey G Hemmett

Alexander Joel Suess

Gregory M Schaeffer

Peter Feldmann

Soroush Abbaspour

Chad Andrew Marquart

Jin Hu

Chaitanya Ravindra Peddawad

Revanta Banerji

Frank Borkam

To increase the efficiency of electronic design automation, in a putative electronic logic circuit design, at least one transparent latch is identified as a candidate for slack stealing. An initial timing slack, available for stealing, and associated with the at least one transparent latch, is determined. Responsive to a determination that the initial timing slack available for stealing is insufficient, it is determined whether the initial timing slack available for stealing is on a feedback path. If so, responsive to determining that the initial timing slack available for stealing is on the feedback path, the initial timing slack available for stealing is replaced with a next worse slack.

Feedback-aware slack stealing across transparent latches empowering performance optimization of digital integrated circuits

A method, system, and compute program product use a generalized macro or a generalized macro timing abstract for timing analysis in a specific timing context. The method includes setting up a timer, and determining a divide ratio of each external clock divider of one or more external clock dividers associated with the generalized macro or the generalized macro timing abstract programmatically as a function of another value. The method also includes performing the timing analysis using the divide ratios of the one or more external clock dividers. Obtaining a physical implementation of an integrated circuit is based on the timing analysis.

Programmable clock division methodology with in-context frequency checking

A system to improve performance of a semiconductor chip design includes a hierarchical analysis module that determines a hierarchical arrangement of the semiconductor chip design. The hierarchical arrangement includes a plurality of arcs located at different levels internal to the semiconductor chip design. The different levels include a macro level, a unit level and a core level. The system further includes a timing/load analysis module that determines first timing characteristics of at least one first arc in the macro level based on a first load applied to the at least one first arc. The system further determines second timing characteristics of at least one second arc in at least one of the unit level and the core level based on the first timing characteristics, with a portion of the second timing characteristics determined irrespective of the first load.

Accurate statistical timing for boundary gates of hierarchical timing models

A method, system and computer program product perform timing analysis of an integrated circuit design with callback-based constraint processing for clock domain independence. A timing graph representation of the integrated circuit design includes nodes interconnected by edges. Loading timing abstracts representing the nodes of the timing graph precedes obtaining a timing result based on propagating timing values and associated timing tags from an input to an output of the integrated circuit design and processing timing constraints at one or more of the nodes as callbacks. Each timing tag indicates a clock domain. After applying a design change, one or more modified timing tags that are added or changed as a result of the design change are determined. The timing constraints associated with the modified timing tags are processed as callbacks, and the timing result are re-computed.

Callback based constraint processing for clock domain independence

According to one exemplary embodiment, a method for parallel processing a network of nodes having at least one ordering constraint and at least one conflict constraint is provided. The method may include breaking a plurality of loops caused by the at least one ordering constraint. The method may also include determining a node order based on the at least one ordering constraint. The method may then include determining a conflict order based on the at least one conflict constraint, whereby no new loops are created in the network. The method may further include performing parallel processing of the network of nodes based on the node order and the conflict order.

Efficient parallel processing of a network with conflict constraints between nodes

A method, system, and compute program product use a generalized macro or a generalized macro timing abstract for a timing analysis in a specific timing context. The method includes setting up a timer, and determining a divide ratio of each external clock divider of one or more external clock dividers associated with the generalized macro or the generalized macro timing abstract programmatically as a function of another value. The method also includes performing the timing analysis using the divide ratios of the one or more external clock dividers. Obtaining a physical implementation of an integrated circuit is based on the timing analysis.

A computer program product for improved modeling of differential circuits is provided. The computer program product includes a computer readable storage medium having program instructions embodied therewith. The program instructions are readable and executable by a processing circuit to cause the processing circuit to represent a configuration of a differential circuit on a defined space with representations of single-ended inputs and outputs disposed as differential input and output pairs along borders of the defined space, respectively, for each differential input and output pair, introduce an internal input or output differential node to feed from or to feed a corresponding differential input or output pair within the borders, respectively, with the internal input and output differential nodes being connectable and perform timing calculations with respect to input and output differential nodes.

Deterministic and statistical timing modeling for differential circuits

Determining static timing analysis margin on non-controlling inputs of clock shaping and other digital circuits using reverse merge timing includes: selecting one or more circuits within the logic design having a plurality of inputs and using reverse merge; identifying a controlling input of the selected circuit from among this plurality of inputs; and determining for at least one non-controlling input of the circuit, a timing value that may be used to drive design optimization based on the difference between arrival times of the controlling and non-controlling inputs.

Method for improving static timing analysis and optimizing circuits using reverse merge

A method for accurately performing a timing, power, and noise analysis by pre-processing the characterization points of the available libraries, storing time consuming parts of the analysis and utilizing the pre-processed information during active runs to calculate the attributes at a desired PVT point. The PVT space is preferably sub-divided into triangular or rectangular regions, preferably obtained using Delaunay triangulation. In one embodiment, the invention performs an up-front pre-processing step on the characterized libraries to compute the static portion of the interpolation function that is independent of the specific instance; and a coefficient matrix that allows for interpolation of specific instances.

Method for supporting multiple libraries characterized at different process, voltage and temperature points

A method for loading checkpoint timing in an environment where the boundary arrival times, slews, required arrival times, or loads differ from the checkpoint run. A timing checkpoint file generated for one or more hierarchical modules, during which each input is assigned a unique phase tag. The association of unique phase tags allows subsequent restart analyses to efficiently adjust the checkpoint timing in relation to the restart timing environment. In the restart run, one or more such checkpoint files is read, during which an initial propagation of arrival, required arrivals and slew times are performed, followed by a local re-update based on adjusted arrival times and the required arrival times. Finally, if multiple hierarchical modules are updated, a global recalculation of timing values is performed based on a slack change threshold in order to determine whether any new timing failures have been introduced.

Growing community of inventors

Hopewell Junction, NY, United States of America

Hemlata Gupta