Adaptive recovery for parallel reactive power throttling

A system, wherein the system includes a global clock signal that informs each of a plurality of compute nodes of the global time for the computing system, the system comprising:

• a computer processor;
• and
• a memory containing a program that, when executed on the computer processor, performs an operation for managing a parallel computing system that comprises the plurality of compute nodes, comprising:
• monitoring a first parameter associated with at least one of the plurality of compute nodes, wherein the first parameter comprises at least one of:
• a measured temperature, a measured current, and a measured power consumption;
• upon determining that the first parameter reaches or exceeds a first threshold for the first parameter, transmitting a global interrupt to at least two of the plurality of compute nodes, wherein the global interrupt includes a predefined time, wherein the at least two of the plurality of compute nodes are configured to reduce power consumption upon the global time matching the predefined time of the global interrupt;
• after transmitting the global interrupt, determining whether second parameters respectively associated with each of the at least two compute nodes reach or fall below a second threshold for the second parameters, wherein the second parameters comprise at least one of:
• a measured temperature, a measured current, and a measured power consumption;
• and
• after all the second parameters reach or fall below the second threshold, cancelling the global interrupt for the compute nodes in the operational group.

A computer program product for managing a parallel computing system that comprises a plurality of compute nodes, wherein the computing system comprises a global clock signal that informs each of the compute nodes of the global time for the computing system, the computer program product comprising:

• a computer-readable storage medium having computer-readable program code embodied therewith, the computer-readable program code configured to:
• monitor a first parameter of the first compute node and the first parameter of the second compute node, wherein the first parameter comprises at least one of:
• a measured temperature, a measured current, and a measured power consumption, and wherein the plurality of compute nodes in the parallel computing system are coupled for data communications;
• upon determining that one of the first parameter of the first compute node and the first parameter of the second compute node reaches or exceeds a first threshold for the first parameter, transmit a global interrupt to the first and second compute nodes, wherein the global interrupt includes a predefined time, wherein the at least two of the plurality of compute nodes are configured to reduce power consumption upon the global time matching the predefined time of the global interrupt;
• after transmitting the global interrupt, determine whether a second parameter of the first compute node and the second parameter of the second compute node fall below a second threshold for the second parameter, wherein the second parameter comprises at least one of:
• a measured temperature, a measured current, and a measured power consumption;
• and
• only after both the second parameter of the first compute node and the second parameter of the second compute node are at or below the second threshold, cancel the global interrupt for both the first compute node and the second compute node.