Abstract: A cell library (21) is optimized for specific operating characteristics. A stimulus file (23) is divided into a number of simulation run files. The simulation run files (27) are distributed to more than one computer work station so that the simulation of the cell occurs in parallel. The netlist (24) of each cell is parameterized to allow the cell to be changed and resimulated to better meet the specific operating characteristics. A cost function (32) is provided which uses the results of a transistor level simulation to calculate the quality of the cell design relative to the specific operating characteristics. Simulated annealing (34) is used to generate new simulation parameter values from a cost generated from the cost function (32). The cell is resimulated a number of times to optimize for the specific operating characteristics and the best design is retained for a new cell library (39). The process is repeated for each cell of the cell library (21).

Abstract: The integrated circuit includes at least one conductance (6) which is adjustable by a digital control signal (17) which encodes steps in value for quantizing an exact value to within a fixed relative accuracy .DELTA.p. The conductance (6) includes elementary conductances which each define a step in value such that each elementary conductance is dimensioned so that a single level in value of the conductance (6) corresponds to two successive levels of the value of the digital control signal (17). If the result from the first of the two levels of value of the digital control signal (17) is a value which is lower, or respectively higher, than the exact value, the corresponding elementary conductance is enabled, or respectively disabled. Accordingly, the adjusted total value of the conductance (6) is equal to the exact value to within the same relative accuracy .DELTA.p, without oscillating between two values straddling this exact value.

Abstract: In order to efficiently recover from a processing error in a central processing trait (CPU) incorporating a cache memory and a basic processing unit, the BPU is provided in duplicate, and all BPU data manipulation operations are performed redundantly. After duplicate data has been obtained from the cache memory and manipulated by the duplicate BPUs, the outputs from the duplicate BPUs are placed on respective master (MRB) and slave (SRB) result busses which are coupled to the cache unit where the results are compared for identity. If the results are not identical, a local error signal is issued.