Abstract: A mechanism is provided for composite contention aware task scheduling. The mechanism performs task scheduling with shared resources in computer systems. A task is a group of instructions. A compute task is a group of compute instructions. A memory task, also referred to as a communication task, may be a group of load/store operations, for example. The mechanism performs composite contention-aware scheduling that considers the interaction among compute tasks, communication tasks, and application threads that include compute and communication tasks. The mechanism performs a composite of memory task throttling and application thread throttling.

Abstract: A mechanism is provided for composite contention aware task scheduling. The mechanism performs task scheduling with shared resources in computer systems. A task is a group of instructions. A compute task is a group of compute instructions. A memory task, also referred to as a communication task, may be a group of load/store operations, for example. The mechanism performs composite contention-aware scheduling that considers the interaction among compute tasks, communication tasks, and application threads that include compute and communication tasks. The mechanism performs a composite of memory task throttling and application thread throttling.

Abstract: Mechanisms are provided for performing approximate run-ahead computations. A first group of compute engines is selected to execute full computations on a full set of input data. A second group of compute engines is selected to execute computations on a sampled subset of the input data. A third group of compute engines is selected to compute a difference in computation results between first computation results generated by the first group of compute engines and second computation results generated by the second group of compute engines. The second group of compute engines is reconfigured based on the difference generated by the third group of compute engines.

Abstract: Mechanisms are provided for performing approximate run-ahead computations. A first group of compute engines is selected to execute full computations on a full set of input data. A second group of compute engines is selected to execute computations on a sampled subset of the input data. A third group of compute engines is selected to compute a difference in computation results between first computation results generated by the first group of compute engines and second computation results generated by the second group of compute engines. The second group of compute engines is reconfigured based on the difference generated by the third group of compute engines.

Abstract: A mechanism is provided for composite contention aware task scheduling. The mechanism performs task scheduling with shared resources in computer systems. A task is a group of instructions. A compute task is a group of compute instructions. A memory task, also referred to as a communication task, may be a group of load/store operations, for example. The mechanism performs composite contention-aware scheduling that considers the interaction among compute tasks, communication tasks, and application threads that include compute and communication tasks. The mechanism performs a composite of memory task throttling and application thread throttling.

Abstract: Mechanisms are provided for performing approximate run-ahead computations. A first group of compute engines is selected to execute full computations on a full set of input data. A second group of compute engines is selected to execute computations on a sampled subset of the input data. A third group of compute engines is selected to compute a difference in computation results between first computation results generated by the first group of compute engines and second computation results generated by the second group of compute engines. The second group of compute engines is reconfigured based on the difference generated by the third group of compute engines.

Abstract: Mechanisms are provided for performing approximate run-ahead computations. A first group of compute engines is selected to execute full computations on a full set of input data. A second group of compute engines is selected to execute computations on a sampled subset of the input data. A third group of compute engines is selected to compute a difference in computation results between first computation results generated by the first group of compute engines and second computation results generated by the second group of compute engines. The second group of compute engines is reconfigured based on the difference generated by the third group of compute engines.

Abstract: A mechanism is provided for composite contention aware task scheduling. The mechanism performs task scheduling with shared resources in computer systems. A task is a group of instructions. A compute task is a group of compute instructions. A memory task, also referred to as a communication task, may be a group of load/store operations, for example. The mechanism performs composite contention-aware scheduling that considers the interaction among compute tasks, communication tasks, and application threads that include compute and communication tasks. The mechanism performs a composite of memory task throttling and application thread throttling.