Altera Corporation

Mark Jervis

Steven Perry

Gregor Nixon

Zhengjun Pan

Gihan De Silva

Volker Mauer

Colman Cheung

Philippe Marti

Efficient matrix operations circuitry is based on combining a matrix decomposition and a forward substitution operations to share the same processing overhead. A dual multiplier circuit selectively applies complex multiplication operations to a first and second input vectors for computing a conjugate dot product vector or a non-conjugate dot product vector. The conjugate dot product vector corresponds to the matrix decomposition operation for triangulating an input matrix to generate an element of a triangulated matrix. The non-conjugate dot product vector corresponds to a forward substitution operation for determining an element of a forward substitution vector from the triangulated matrix.

Matrix operations in an integrated circuit device

While debugging, a user chooses an incremental recompile. Internal signals of interest are selected and output pins are optionally reserved. An incremental recompile of the compiled design includes compiling a routing from each internal signal to an output pin. The technology-mapped netlist and placing and routing information corresponding to an original compiled design are saved into a database during full compilation. During debugging, an incremental compiler retrieves this information to build the original routing netlist. The database building, logic synthesis and technology mapping stages may be skipped. New connections are added, fitted to the device, and then the final routing netlist is output into a programming output file (POF) in a form suitable for programming the PLD. The user views the internal signals at the output pins chosen. The user may iterate through this process many times in order to debug the PLD. The debugging assignments may be deleted.

Chip debugging using incremental recompilation

While debugging, a user chooses an incremental recompile. Internal signals of interest and output pins are selected, and a number of additional registers are chosen to insert in the path of each internal signal. A clock is selected for the registers. An incremental recompile of the compiled design compiles a routing from each internal signal to an output pin via the added registers. The database building and logic synthesis stages are skipped. The post-fitting logical netlist and routing netlist are retrieved. The new registers are created and the internal signal is connected to the output pin atom in the logical netlist. The fitter places and routes the connections to create a new routing netlist and then the new routing netlist is output into a programming output file (POF) in a form suitable for programming the PLD. The original routing netlist is undisturbed. The user views the internal signals at the output pins chosen. The user may iterate through this process many times in order to debug the PLD. The debugging assignments may be deleted.

Chip debugging using incremental recompilation and register insertion

Growing community of inventors

Windsor, United Kingdom

Mark Jervis