Xilinx, Inc.

Synopsys, Inc.

Sankaranarayanan Srinivasan

Raymond Kong

Kamal Chaudhary

Anirban Rahut

Daniel J Downs

Richard Yachyang Sun

John J Laurence

Sridhar Krishnamurthy

Brian Philofsky

Etienne Lepercq

Vinod Kumar Nakkala

Avinash Anantharamu

Senthilkumar Thoravi Rajavel

Pierre Clement

Saibal Ghosh

Sashikala Oblisetty

Sudip K Nag

Srinivasan Dasasathyan

Dinesh D Gaitonde

Sandor S Kalman

Amit Singh

W Story Leavesley, Iii

Walter A Manaker, Jr

Krishnan Anandh

Douglas P Wieland

Nicholas A Mezei

George L McHugh

Damon McCormick

Benoit Payette

David A Ewing

A method includes storing a base model generated using base data and receiving training data generated by compiling circuit designs. The method also includes generating, using the training data, a tuned model and generating, using the training data and the base data, a hybrid model. The method further includes receiving a selected cost function and biasing the base model, the tuned model, and the hybrid model using the selected cost function.

Machine-learning enhanced compiler

A method includes generating a netlist for a circuit design and predicting, by applying a first machine learning model to the netlist, a first compile time for the circuit design. The method also includes predicting, by applying a second machine learning model to the netlist, a first place and route strategy based on the first compile time. The method further includes adjusting a logic of the circuit design in accordance with the first place and route strategy.

A computer-implemented method of implementing a circuit design within a programmable logic device can include selecting at least one circuit element of the circuit design. The selected circuit element can be converted to a latch. A timing analysis can be performed upon the circuit design after conversion of the selected circuit element to a latch. A determination can be made by a computer as to whether the timing of the circuit design improves and the conversion of the selected circuit element to a latch can be accepted when the timing of the circuit design improves. The circuit design can be output.

Latch based optimization during implementation of circuit designs for programmable logic devices

A design hierarchy based on an implementation set abstraction of a user design for an integrated circuit design includes a plurality of nodes and a definition for each of the nodes in the plurality of nodes that describes the type of elements contained in each node and the hierarchy defined by each of the nodes. Each node can include at least one implementation element of the design and the at least one implementation element can be selected among the group including a set of logical elements, a set of placed elements, and a set of placed and routed elements.

Method and arrangement providing for implementation granularity using implementation sets

Circuit design fitting for an integrated circuit is described. A mapped design for the circuit design is obtained. A first placement of the mapped design in association with an integrated circuit is performed. Circuit blocks are marked associated with the integrated circuit with control set identifiers. A circuit object is associated with a control set identifier. A site for placement of the first circuit object is located. The site is associated with a circuit resource block, which is associated with circuit resource blocks of the integrated circuit. Nearest neighbor circuit resource blocks with respect to the circuit resource block are acquired. The nearest neighbor circuit resource blocks of the circuit resource block are categorized in response to statuses. The circuit object is placed in a nearest neighbor of the nearest neighbor circuit resource blocks of the circuit resource block for a second placement.

Circuit design fitting

Various approaches for improving clock speed for a circuit design. In one embodiment, a graph having nodes and edges that represent the circuit design is generated. The nodes represent flip-flops of the design, the edges represent couplings of data inputs and outputs of the flip-flops, and the edges have associated delay values for respective durations of signal delays of the couplings. A smallest period is determined for which subtracting each delay value from the smallest period and associating the difference with the associated edge does not create a negative cycle in the graph. A path in the graph is selected, the path including selected flip-flops and connecting edges. The circuit design is modified by replacing the selected flip-flops with latches, and the smallest period is output.

Clock speed for a digital circuit

A computer-implemented method of implementing a circuit design within a programmable logic device can include selecting at least one circuit element of the circuit design. The selected circuit element can be converted to a latch. A timing analysis can be performed upon the circuit design after conversion of the selected circuit element to a latch. A determination can be made as to whether the timing of the circuit design improves and the conversion of the selected circuit element to a latch can be accepted when the timing of the circuit design improves. The circuit design can be output.

Approaches for merging replicate logic blocks of a circuit design. Groups of replicate logic blocks in a placed circuit design are determined. For the replicate logic blocks in each group, a determination is made whether or not to merge replicate logic blocks in a subset of the replicate logic blocks into a respective single replacement logic block for the subset. In response to determining to merge the replicate logic blocks in the subset, the replicate logic blocks in the subset are replaced in the circuit design with the respective replacement logic block. The circuit design having the replacement logic block is stored in a memory by a processor executing the process.

Merging of equivalent logic blocks in a circuit design

Circuit placement for increasing circuit packing density for an integrated circuit is described. A design is synthesized and mapped. Components of the design are placed to provide a first placed design. A congestion density map is generated for the first placed design. A congestion region in the congestion density map is identified and targeted for determining if the first placed design has a control set conflict. A first circuit object associated with the control set conflict is selected and either re-placed or re-synthesized to at least diminish the control set conflict.

Placement driven control set resynthesis

A method of implementing a circuit design for a target device can include assigning load pins of a high fanout signal of a placed circuit design into a plurality of windows according to a location of each load pin on the target device. A source of the high fanout signal can be replicated, wherein each window is associated with a source of the high fanout signal. For each source of the high fanout signal, the source can be connected to load pins of the window associated with the source and the source can be placed within the window associated with the source. The placed circuit design can be output.

Fanout-optimization during physical synthesis for placed circuit designs

Various approaches for incrementally updating the timing of an implementation of an electronic circuit design are disclosed. In one approach, a subset timing graph is selected from a primary timing graph. Alternative subset timing graphs are generated that are functionally equivalent and structurally different with respect to the selected subset timing graph. For each of the alternative timing graphs, a respective timing metric is determined. The determined timing metrics and a timing metric for the selected subset timing graph are compared. An alternative timing graph is selected in response to the comparison. Structurally different portions of the selected one of the one or more alternative timing graphs are verified with regard to the design constraints. The structurally different portions are stored to the primary timing graph.

Efficient timing graph update for dynamic netlist changes

A method of placing a circuit design can include selecting one or more candidate mobile nodes from a plurality of overlapped nodes of the circuit design and determining a gain region for each candidate mobile node. The method also can include assigning the candidate mobile node to a site within a gain region according to a cost function. The gain region is associated with the candidate mobile node. The method further can include iteratively selecting and assigning candidate mobile nodes according to a measure of overlap for the circuit design.

Cost-based performance driven legalization technique for placement in logic designs

Method and apparatus for implementing a pre-implemented circuit design for a programmable logic device is described. In one example, a definition of the pre-implemented circuit design is obtained (). The definition includes a first physical implementation and a first logical implementation. A second logical implementation is produced () for an instance of the pre-implemented circuit design using the first logical implementation. A second physical implementation is produced () for then instance of the pre-implemented circuit design using the first physical implementation.

Method and apparatus for implementing a pre-implemented circuit design for a programmable logic device

A method of designing an integrated circuit using implementation directives for flow control can include the step of loading a design along with specified constraints, creating at least one instance of an data structure formed from a partial netlist, and decomposing at least one set of high level rules into simple implementation directives. The method can further include the steps of selectively attaching the simple implementation directives to the data structure, implementing a task manager which queries a data structure node to create a list of tasks to be performed on the data structure, and executing the list of tasks using a generic flow engine.

Growing community of inventors

San Jose, CA, United States of America

Sankaranarayanan Srinivasan