Abstract: Various embodiments monitor system noise in a parallel computing system. In one embodiment, at least one set of system noise data is stored in a shared buffer during a first computation interval. The set of system noise data is detected during the first computation interval and is associated with at least one parallel thread in a plurality of parallel threads. Each thread in the plurality of parallel threads is a thread of a program. The set of system noise data is filtered during a second computation interval based on at least one filtering condition creating a filtered set of system noise data. The filtered set of system noise data is then stored.

Abstract: Various embodiments monitor system noise in a parallel computing system. In one embodiment, at least one set of system noise data is stored in a shared buffer during a first computation interval. The set of system noise data is detected during the first computation interval and is associated with at least one parallel thread in a plurality of parallel threads. Each thread in the plurality of parallel threads is a thread of a program. The set of system noise data is filtered during a second computation interval based on at least one filtering condition creating a filtered set of system noise data. The filtered set of system noise data is then stored.

Abstract: Various embodiments monitor system noise in a parallel computing system. In one embodiment, at least one set of system noise data is stored in a shared buffer during a first computation interval. The set of system noise data is detected during the first computation interval and is associated with at least one parallel thread in a plurality of parallel threads. Each thread in the plurality of parallel threads is a thread of a program. The set of system noise data is filtered during a second computation interval based on at least one filtering condition creating a filtered set of system noise data. The filtered set of system noise data is then stored.

Abstract: As disclosed herein, a method, executed by a computer, for analyzing a performance metric includes distributing link connection information for a network to each node in the network, pairing each node with another node in the network to provide paired nodes, testing a performance metric for the paired nodes, and generating a performance report for the network. The method may further include receiving link connection information for the network. The method may further include constructing a connectivity graph based on the received link connection information. The method can be used to test link performance within non-homogeneous networks. A computer system product and a computer program product corresponding to the method are also disclosed herein.

Abstract: As disclosed herein, a method, executed by a computer, for analyzing a performance metric includes distributing link connection information for a network to each node in the network, pairing each node with another node in the network to provide paired nodes, testing a performance metric for the paired nodes, and generating a performance report for the network. The method may further include receiving link connection information for the network. The method may further include constructing a connectivity graph based on the received link connection information. The method can be used to test link performance within non-homogeneous networks.

Abstract: Various embodiments monitor system noise in a parallel computing system. In one embodiment, at least one set of system noise data is stored in a shared buffer during a first computation interval. The set of system noise data is detected during the first computation interval and is associated with at least one parallel thread in a plurality of parallel threads. Each thread in the plurality of parallel threads is a thread of a program. The set of system noise data is filtered during a second computation interval based on at least one filtering condition creating a filtered set of system noise data. The filtered set of system noise data is then stored.

Abstract: A capability is provided for predicting the performance of a communications network, such as a multi-stage communications network. The predicted performance includes, in one aspect, predicting the amount of bandwidth that a server of the communications network could deliver with a full load on the network. In one aspect, the predictive capability includes determining an aggregate load for each server of the network being analyzed; performing normalization on the loads of the links used by the servers to obtain updated link and/or server bandwidths; and determining for each server, a minimum bandwidth over the links in the server's route.

Abstract: Various embodiments monitor system noise in a parallel computing system. In one embodiment, at least one set of system noise data is stored in a shared buffer during a first computation interval. The set of system noise data is detected during the first computation interval and is associated with at least one parallel thread in a plurality of parallel threads. Each thread in the plurality of parallel threads is a thread of a program. The set of system noise data is filtered during a second computation interval based on at least one filtering condition creating a filtered set of system noise data. The filtered set of system noise data is then stored.

Abstract: A capability is provided for predicting the performance of a communications network, such as a multi-stage communications network. The predicted performance includes, in one aspect, predicting the amount of bandwidth that a server of the communications network could deliver with a full load on the network. In one aspect, the predictive capability includes determining an aggregate load for each server of the network being analyzed; performing normalization on the loads of the links used by the servers to obtain updated link and/or server bandwidths; and determining for each server, a minimum bandwidth over the links in the server's route.

Abstract: A method, information processing system, and computer program product are provided for managing operating system interference on applications in a parallel processing system. A mapping of hardware multi-threading threads to at least one processing core is determined, and first and second sets of logical processors of the at least one processing core are determined. The first set includes at least one of the logical processors of the at least one processing core, and the second set includes at least one of a remainder of the logical processors of the at least one processing core. A processor schedules application tasks only on the logical processors of the first set of logical processors of the at least one processing core. Operating system interference events are scheduled only on the logical processors of the second set of logical processors of the at least one processing core.

Abstract: A method, information processing system, and computer program product are provided for managing operating system interference on applications in a parallel processing system. A mapping of hardware multi-threading threads to at least one processing core is determined, and first and second sets of logical processors of the at least one processing core are determined. The first set includes at least one of the logical processors of the at least one processing core, and the second set includes at least one of a remainder of the logical processors of the at least one processing core. A processor schedules application tasks only on the logical processors of the first set of logical processors of the at least one processing core. Operating system interference events are scheduled only on the logical processors of the second set of logical processors of the at least one processing core.

Abstract: A method, information processing system, and computer program product are provided for managing operating system interference on applications in a parallel processing system. A mapping of hardware multi-threading threads to at least one processing core is determined, and first and second sets of logical processors of the at least one processing core are determined. The first set includes at least one of the logical processors of the at least one processing core, and the second set includes at least one of a remainder of the logical processors of the at least one processing core. A processor schedules application tasks only on the logical processors of the first set of logical processors of the at least one processing core. Operating system interference events are scheduled only on the logical processors of the second set of logical processors of the at least one processing core.

Abstract: Network managers are operated in verification mode to facilitate error handling of communications networks. In verification mode, error reporting remains enabled, even for those components of a communications network reporting errors. A step-by-step procedure is provided for handling each type of error that is detected. Subsequent to handling any reported errors, the network manager is removed from verification mode and may be placed in production mode.