Determining soft error infliction probability

A computer-implemented method comprising:

• obtaining a representation of a circuit, wherein the circuit comprises nodes, wherein the nodes comprise at least one critical node;
• obtaining a trace, wherein the trace comprises recorded values of the nodes in a plurality of cycles;
• determining, by a processor, a Soft Error Infliction Probability (SEIP) of a node, wherein the SEIP is a value representing a probability that a Single Event Upset (SEU) effecting the node in a cycle will inflict a soft error by propagating through the circuit to the at least one critical node, wherein said determining comprises simulating a propagation of the SEU from the cycle to consecutive cycles, wherein the consecutive cycles are successive cycles after the cycle, wherein said simulating utilizes values from the trace which are associated with the consecutive cycles, wherein said simulating is performed until reaching at least one threshold selected from the group consisting of:
• a maximum number of inverted values within a simulated cycle;
• a maximum number of inverted values during said simulating;
• and a maximum number of simulated cycles during said simulating;
• and wherein in response to reaching the threshold, approximating that the SEU inflicts a soft error;
• and
• outputting the SEIP of the node.

A computer-implemented method comprising:

• obtaining a representation of a circuit, wherein the circuit comprises nodes, wherein the nodes comprise at least one critical node;
• obtaining a trace, wherein the trace comprises recorded values of the nodes in a plurality of cycles;
• determining, by a processor, a Soft Error Infliction Probability (SEIP) of a node, wherein the SEIP is a value representing a probability that a Single Event Upset (SEU) effecting the node in a cycle will inflict a soft error by propagating through the circuit to the at least one critical node, wherein said determining comprises simulating a propagation of the SEU from the cycle to consecutive cycles, wherein the consecutive cycles are successive cycles after the cycle, wherein said simulating utilizes values from the trace which are associated with the consecutive cycles, wherein said simulating further comprises repeatedly computing values of the nodes in a second cycle based on a first cycle and based on the trace, wherein said computing the values of the nodes in the second cycle comprises:
• having a first portion of the nodes whose values in the first cycle are inverted with respect to recorded values of the trace;
• determining a second portion of the nodes whose values in the second cycle are influenced by the values of the first portion in the first cycle;
• computing values of the second portion using recorded values of the trace in the first cycle for any node not in the first portion and using inverted values of recorded values of the trace in the first cycle for any node in the first portion;
• and
• comparing values of the second portion to identify a third portion of the nodes whose values in the second cycle are inverted with respect to recorded values of the trace;
• and
• outputting the SEIP of the node.

A computer-implemented method comprising:

• obtaining a representation of a circuit, wherein the circuit comprises nodes, wherein the nodes comprise at least one critical node;
• obtaining a trace, wherein the trace comprises recorded values of the nodes in a plurality of cycles;
• determining, by a processor, a Soft Error Infliction Probability (SEIP) of a node, wherein the SEIP is a value representing a probability that a Single Event Upset (SEU) effecting the node in a cycle will inflict a soft error by propagating through the circuit to the at least one critical node, wherein said determining comprises simulating a propagation of the SEU from the cycle to consecutive cycles, wherein the consecutive cycles are successive cycles after the cycle, wherein said simulating utilizes values from the trace which are associated with the consecutive cycles, wherein said simulating the propagation of the SEU is preformed until the SEU propagates to a second node having an SEIP above a threshold value, and wherein in response to the SEU propagating to the second node, approximating that the SEU inflicts a soft error;
• and
• outputting the SEIP of the node.

A computer-implemented method comprising:

• obtaining a representation of a circuit, wherein the circuit comprises nodes, wherein the nodes comprise at least one critical node;
• obtaining a trace, wherein the trace comprises recorded values of the nodes in a plurality of cycles;
• determining, by a processor, a Soft Error Infliction Probability (SEIP) of a node, wherein the SEIP is a value representing a probability that a Single Event Upset (SEU) effecting the node in a cycle will inflict a soft error by propagating through the circuit to the at least one critical node, wherein said determining comprises simulating a propagation of the SEU from the cycle to consecutive cycles, wherein the consecutive cycles are successive cycles after the cycle, wherein said simulating utilizes values from the trace which are associated with the consecutive cycles, wherein said determining the SEIP of the node further comprises:
• for each cycle of a plurality of cycles in the trace:
• simulating an SEU to the node in the cycle;
• and
• simulating propagation of the SEU to determine whether the SEU inflicts a soft error;
• and
• computing the SEIP of the node by using the formula N/M, wherein N is a total number of soft errors inflicted by the simulated SEUs, wherein M is a total number of simulated SEUs to the node;
• and
• outputting the SEIP of the node.

A computer-implemented method comprising:

• obtaining a representation of a circuit, wherein the circuit comprises nodes, wherein the nodes comprise at least one critical node;
• obtaining a trace, wherein the trace comprises recorded values of the nodes in a plurality of cycles;
• determining, by a processor, a Soft Error Infliction Probability (SEIP) of a node, wherein the SEIP is a value representing a probability that a Single Event Upset (SEU) effecting the node in a cycle will inflict a soft error by propagating through the circuit to the at least one critical node, wherein said determining comprises simulating a propagation of the SEU from the cycle to consecutive cycles, wherein the consecutive cycles are successive cycles after the cycle, wherein said simulating utilizes values from the trace which are associated with the consecutive cycles, wherein said simulating is performed by computing, in each cycle, values only to a subset of the nodes and using recorded values from the trace for other nodes not in the subset, wherein the subset of the nodes includes the nodes whose value is potentially effected by nodes whose simulated values in a previous cycle are different than corresponding recorded values in the trace;
• and
• outputting the SDP of the node.