Xilinx, Inc.

Satish B Sivaswamy

Garik Mkrtchyan

Nitin Deshmukh

Mehrdad Eslami Dehkordi

Sreesan Venkatakrishnan

Brian Guttag

Jindrich Zejda

Ilya K Ganusov

Guenter Stenz

Grigor S Gasparyan

Benjamin S Devlin

Chun Zhang

Parivallal Kannan

Walter A Manaker, Jr

Xiaojian Yang

Frederic Revenu

Atul Srinivasan

Paul D Kundarewich

Nicholas A Mezei

Meghraj Kalase

Amish Pandya

Frank Mueller

Karthic P

Srinivasan Dasasathyan

Jinny Singh

Thomas O Satter

Bhasker Pinninti

Ashot Shakhkyan

Vishal Tripathi

Routing a circuit design includes generating a graph of the circuit design where each connected component is represented as a vertex, generating a routing solution for the circuit design by routing packet-switched nets so that the packet-switched nets of a same connected component do not overlap, and, for each routing resource that is shared by packet-switched nets of different connected components, indicating the shared routing resource on the graph by adding an edge. Cycle detection may be performed on the graph. For each cycle detected on the graph, the cycle may be broken by deleting the edge from the graph and ripping-up a portion of the routing solution corresponding to the deleted edge. The circuit design, or portion thereof, for which the routing solution was ripped up may be re-routed using an increased cost for a shared routing resource freed from the ripping-up.

Deadlock detection and prevention for routing packet-switched nets in electronic systems

Methods and systems for selecting between single-process and multi-process implementation flows involve identifying features of a circuit design by a design tool. A classification model is applied to the features. The classification model indicates whether an implementation flow on the circuit design is likely to have a runtime within a first range of runtimes or a runtime within a second range of runtimes. The implementation flow is executed by the design tool in a single process in response to the classification model indicating the implementation flow on the circuit design is likely to have a runtime within the first range of runtimes. The implementation flow is executed by the design tool in a plurality of processes in response to the classification model indicating the implementation flow on the circuit design is likely to have a runtime within the second range of runtimes.

Optimizing use of computer resources in implementing circuit designs through machine learning

Examples described herein provide a non-transitory computer-readable medium storing instructions, which when executed on one or more processors, cause the one or more processors to perform operations. The operations include generating a plurality of child processes according to a number of a plurality of partitions in an integrated circuit (IC) design for an IC die, each of the plurality of child processes corresponding to and assigned to a respective one of the plurality of partitions. The operations include transmitting each of the plurality of partitions to a respective one of the plurality of child processes for routing, each of the plurality of partitions comprising a placement of components for the IC design. The operations include receiving a plurality of routings from the plurality of child processes. The operations include merging the plurality of routings into a global routing for the IC design by assembling together to form a global routing.

Distributed parallel processing routing

A design tool determines features of a circuit design and applies a first model to the features. The first model indicates a predicted value of a metric based on the plurality of features. The design tool applies an explanation model to the features, and the explanation model indicates levels of contributions by the features to the predicted value of the metric, respectively. The design tool selects a feature of the plurality of features based on the respective levels of contributions and looks up a recipe associated with the feature in a database having possible features associated with recipes. The design tool processes the circuit design according to the recipe into implementation data that is suitable for making an integrated circuit (IC).

Predicting a performance metric based on features of a circuit design and explaining marginal contributions of the features to the prediction

Implementing an application can include generating, from the application, a compact data flow graph (DFG) including load nodes, inserting, in the compact DFG, a plurality of virtual buffer nodes (VBNs) for each of a plurality of buffers of a data processing engine (DPE) array to be allocated to nets of the application, and, forming groups of one or more load nodes of the compact DFG based on shared buffer requirements of the loads on a per net basis. Virtual driver nodes (VDNs) that map to drivers of nets can be added to the compact DFG, where each group of the compact DFG is driven by a dedicated VDN. Connections between VDNs and load nodes through selected ones of the VBNs are created according to a plurality of constraints. The plurality of buffers are allocated to the nets based on the compact DFG as connected.

Application implementation and buffer allocation for a data processing engine array

Synthetizing a hardware description language code into a netlist comprising loads and a multi-clock buffer (MBUF). The MBUF receives a global clocking signal and generates a first and a second related clocking signals. The loads are grouped into a first and a second groups receiving the first and the second clocking signals respectively. A first/second clock modifying leaf are placed between a common node and the first/group groups respectively, wherein the common node is positioned closer in proximity to the first/second groups in comparison to a clock source generating the global clocking signal. The first/second clock modifying leaves receive a least divided clocking signal from the MBUF and generate the first/second clocking signals respectively. The least divided clocking signal is routed from the MBUF to the first/second clock modifying leaves. The first/second clocking signals are routed from the first/second clock modifying leaves to the first/second group respectively.

Synchronous clock domain crossing skew optimization and multi-clock buffer (MBUFG)

Embodiments herein describe techniques for preventing a stall when transmitting data between a producer and a consumer in the same integrated circuit (IC). A stall can occur when there is a split point and a convergence point between the producer and consumer. To prevent the stall, the embodiments herein adjust the latencies of one of the paths (or both paths) such that a maximum latency of the shorter path is greater than, or equal to, the minimum latency of the longer path. When this condition is met, this means the shortest path has sufficient buffers (e.g., a sufficient number of FIFOs and registers) to queue/store packets along its length so that a packet can travel along the longer path and reach the convergence point before the buffers in the shortest path are completely full (or just become completely full).

Optimizing hardware design throughput by latency aware balancing of re-convergent paths

Performing partition wire assignment for routing a multi-partition circuit design can include performing, using computer hardware, a global assignment phase by clustering a plurality of super-long lines (SLLs) into a plurality of SLL bins, clustering loads of nets of a circuit design into a plurality of load clusters, and assigning the plurality of SLL bins to the plurality of load clusters. For each SLL bin, a detailed assignment phase can be performed wherein each net having a load cluster assigned to the SLL bin is assigned one or more particular SLLs of the SLL bin using the computer hardware.

Partition wire assignment for routing multi-partition circuit designs

Some examples described herein relate to routing in routing elements (e.g., switches). In an example, a design system includes a processor and a memory, storing instruction code, coupled to the processor. The processor is configured to execute the instruction code to model a communication network among switches interconnected in an array of data processing engines (DPEs), generate routes for an application on the modeled communication network, and translate the routes to a file. Each DPE includes a hardened processor core, a memory module, and one or more of the switches. Each switch includes an input or output port that is capable of being shared by multiple routes. Port(s) of each switch are modeled as respective node(s). Generating the routes includes using an A* algorithm that includes a congestion costing function based on a capacity of respective nodes in the modeled communication network and a cumulative demand for the respective nodes.

Data processing engine (DPE) array routing

Disclosed approaches for fixing a hold time violation in a circuit design include determining a first hold budget that is an amount to fix a first hold time violation on a first path of the circuit design. For each connection of a first plurality of connections on the first path, a respective projected setup slack of the connection in allocating the first hold budget to fixing the first hold time violation on the connection is determined. For each connection of the first plurality of connections, a respective connection hold budget based on the first hold budget and the respective projected setup slack is determined. Each connection of the first plurality of connections is adjusted according to the respective connection hold budget.

Fixing hold time violations using hold time budgets and slacks of setup times

Approaches for processing a circuit design include determining pin slack values for pins of the circuit elements in the circuit design. A processor selects a subset of endpoints based on pin slack values of the endpoints being in a critical slack range and determines startpoints of the circuit design that are in respective critical fanin cones. For each endpoint of the subset, the processor determines an arrival time from each startpoint in the respective critical fanin cone and determines for each startpoint-endpoint pair, a respective set of constraint values as a function of the respective arrival time from the startpoint. The processor generates a graph in the memory circuit from the startpoint-endpoint pairs. First nodes in the graph represent the startpoints and second nodes in the graph represent the endpoints, and values in the respective set of constraint values are associated with edges that connect the nodes.

Growing community of inventors

Fremont, CA, United States of America

Satish B Sivaswamy