Reliability-aware runtime optimal processor configuration

A method, comprising:

• integrating soft error data and hard error data into a single metric within a metric module;
• performing a coarse-grained actuation by adjusting a number of on-cores of a processor based on operating system level metrics and sensor data;
• recomputing the single metric to determine a recomputed single metric;
• and
• performing a fine-grained actuation by adjusting the processor voltage based on the operating system level metrics and the sensor data, wherein the metric module uses a statistical dimensionality reduction technique to convert a combined input of the soft error data and the hard error data into a reduced dimensional array.

A non-transitory computer program product, comprising:

• a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a balanced reliability metric module to cause the balanced reliability metric module to:
• integrate soft error data and hard error data into a single metric;
• adjust at least one of a voltage of a processor and a number of on-cores to optimize overall reliability of the processor;
• and
• convert a combined input of the soft error data and the hard error data into a reduced dimensional array using a statistical dimensionality reduction technique.