International Business Machines Corporation

Ronald Edward Newhart

Michael Jay Shapiro

Louis Bennie Capps, Jr

Anand Haridass

Robert H Bell, Jr

Glenn Graham Daves

Thomas Edward Cook

Jean Audet

Michael A Paolini

Joanne Ferris

Michael Richard Ouellette

Naresh Nayar

Bernadette Ann Pierson

Thomas J Dewkett

James Douglas Jordan

Warren D Dyckman

A multiprocessor system having plural heterogeneous processing units schedules instruction sets for execution on a selected of the processing units by matching workload processing characteristics of processing units and the instruction sets. To establish an instruction set's processing characteristics, the homogeneous instruction set is executed on each of the plural processing units with one or more performance metrics tracked at each of the processing units to determine which processing unit most efficiently executes the instruction set. Instruction set workload processing characteristics are stored for reference in scheduling subsequent execution of the instruction set.

Multicore processor and method of use that configures core functions based on executing instructions

A processor has multiple cores with each core having an associated function to support processor operations. The functions performed by the cores are selectively altered to improve processor operations by balancing the resources applied for each function. For example, each core comprises a field programmable array that is selectively and dynamically programmed to perform a function, such as a floating point function or a fixed point function, based on the number of operations that use each function. As another example, a processor is built with a greater number of cores than can be simultaneously powered, each core associated with a function, so that cores having functions with lower utilization are selectively powered down.

Multicore processor and method of use that adapts core functions based on workload execution

A multiprocessor system which includes automatic workload distribution. As threads execute in the multiprocessor system, an operating system or hypervisor continuously learns the execution characteristics of the threads and saves the information in thread-specific control blocks. The execution characteristics are used to generate thread performance data. As the thread executes, the operating system continuously uses the performance data to steer the thread to a core that will execute the workload most efficiently.

Method for autonomic workload distribution on a multicore processor

A computer implemented method, data processing system, and computer usable code are provided for burn-in testing of a multiprocessor. A process identifies a power management data set for a plurality of processor cores associated with the multiprocessor. The process selects one or more of the plurality of processor cores to form a selected set of processor cores based upon the power management data set. The process initiates a burn-in test across the selected set of processor cores. In response to a determination that all processor cores in the plurality of processor cores have not been selected, the process repeats the above selecting and initiating steps until all the processor cores have been selected.

Uniform power density across processor cores at burn-in

A system and method to optimize multi-core microprocessor performance using voltage offsets is presented. A multi-core device tests each of its processor cores in order to identify each processor core's optimum supply voltage. In turn, the device configures voltage offset networks for each processor core based upon each processor core's identified optimum supply voltage. As a result, the offset voltages produced by the voltage offset networks are subtracted from the multi-core device's main voltage, which results in the voltage offset networks supplying optimum supply voltages to each processor core. The voltage offset networks may include fuses to generate a fixed voltage offset, or the voltage offset networks may include a control circuit to dynamically adjust voltage offsets during the multi-core device's operation.

System and method to optimize multi-core microprocessor performance using voltage offsets

A reduced number of voltage regulator modules provides a reduced number of supply voltages to the package. The package includes a voltage plane for each of the voltage regulator modules. Each core or other component on the die is tied to a switch on the package, and each switch is electrically connected to all of the voltage planes. A wafer-level test determines a voltage that optimizes performance of each core or other component. Given these voltage values, an engineer may determine voltage settings for the voltage regulator modules and which cores are to be connected to which voltage regulator modules. A database stores voltage setting data, such as the optimal voltage for each component, switch values, or voltage settings for each voltage regulator module. An engineering wire may permanently set each switch to customize the voltage supply to each core or other component.

Power grid structure to optimize performance of a multiple core processor

An apparatus and method for providing a multi-core integrated circuit chip that reduces the cost of the package and board while optimizing performance of the cores for use with a single voltage plane. The apparatus and method of the illustrative embodiments make use of a dynamic burn-in technique that optimizes all of the cores on the chip to run at peak performance at a single voltage. Each core is burned-in with a customized burn-in voltage that provides uniform power and performance across the whole chip. This results in a higher burn-in yield and lower overall power in the integrated circuit chip. The optimization of the cores to run at peak performance at a single voltage is achieved through use of the negative bias temperature instability affects on the cores imparted by the burn-in voltages applied.

Growing community of inventors

Essex Junction, VT, United States of America

Ronald Edward Newhart