Abstract: A robust printed circuit board (PCB) that includes at least two power layers that are used in providing power to components connected to the PCB. The power layers may be a power plane layer and a ground plane layer. The power plane layer is situated such that its edge is pulled back a second distance from the planar edge of the PCB. The ground plane layer is situated such that its edge is pulled back a first distance from the planar edge of the PCB. The second distance and the first distance are different, and as a result, the planar edges of the power plane layer and the ground plane layer respectively do not coincide.

Abstract: One embodiment provides an error detection method wherein single-bit errors in a memory module are detected and identified as being a random error or a repeat error. Each identified random error and each identified repeat error occurring in a time interval is counted. An alert is generated in response to a number of identified random errors reaching a random-error threshold or a number of identified repeat errors reaching a repeat-error threshold during the predefined interval. The repeat-error threshold is set lower than the random-error threshold. A hashing process may be applied to the memory address of each detected error to map the location of the error in the memory system to a corresponding location in an electronic table.

Abstract: A robust printed circuit board (PCB) that includes at least two power layers that are used in providing power to components connected to the PCB. The power layers may be a power plane layer and a ground plane layer. The power plane layer is situated such that its edge is pulled back a second distance from the planar edge of the PCB. The ground plane layer is situated such that its edge is pulled back a first distance from the planar edge of the PCB. The second distance and the first distance are different, and as a result, the planar edges of the power plane layer and the ground plane layer respectively do not coincide.

Abstract: One embodiment provides an error detection method wherein single-bit errors in a memory module are detected and identified as being a random error or a repeat error. Each identified random error and each identified repeat error occurring in a time interval is counted. An alert is generated in response to a number of identified random errors reaching a random-error threshold or a number of identified repeat errors reaching a repeat-error threshold during the predefined interval. The repeat-error threshold is set lower than the random-error threshold. A hashing process may be applied to the memory address of each detected error to map the location of the error in the memory system to a corresponding location in an electronic table.

Abstract: Methods, apparatus, and products are disclosed for preemptive thermal management for a computing system based on cache performance, the computing system having a processor, cached computer memory operatively coupled to the processor, and a processor cache operatively coupled to the processor, the processor cache capable of storing a subset of memory contents of the cached computer memory, that include: attempting, by the processor, to retrieve portions of the memory contents of the cached computer memory from the processor cache, resulting in cache misses for the processor cache; tracking, by the processor, cache miss statistics for the processor cache in the computing system, the cache miss statistics describing the cache misses for the processor cache; and administering a thermal management device for the computing system in dependence upon the cache miss statistics, the thermal management device operatively coupled to the processor and capable of managing temperature for the computing system.

Abstract: The present invention is directed to the detection of faulty CPU heat sink coupling during system power-up. A method in accordance with an embodiment of the present invention includes: monitoring a slope of a CPU temperature rise from initial system power-up; determining if the slope of the CPU temperature rise exceeds an expected value; and in the case that the slope of the CPU temperature rise exceeds the expected value, indicating an existence of a possible fault (PFA) related to a heat sink coupled to the CPU.

Abstract: Power is allocated to blades based on an estimate of the actual power they are expected to use rather than their maximum-power draw-value. To protect against situations where the estimated actual-power draw-value is exceeded, a hardware comparator monitors the blade system load against a predetermined threshold value set by a management module (MM) based on user input. If this threshold value is exceeded, a throttle latch is triggered, based on a signal from a service processor monitoring the blade system load. The output of this latch directly engages throttling. The service processor also monitors the output of the latch and communicates information regarding the throttling to the MM for evaluation.

Abstract: Methods, apparatus, and products are disclosed for preemptive thermal management for a computing system based on cache performance, the computing system having a processor, cached computer memory operatively coupled to the processor, and a processor cache operatively coupled to the processor, the processor cache capable of storing a subset of memory contents of the cached computer memory, that include: attempting, by the processor, to retrieve portions of the memory contents of the cached computer memory from the processor cache, resulting in cache misses for the processor cache; tracking, by the processor, cache miss statistics for the processor cache in the computing system, the cache miss statistics describing the cache misses for the processor cache; and administering a thermal management device for the computing system in dependence upon the cache miss statistics, the thermal management device operatively coupled to the processor and capable of managing temperature for the computing system.

Abstract: A method and system are disclosed to enable and control power reduction in a blade/chassis system. A “maximum power reduction” attribute is stored in the VPD of the blade (or can otherwise be input to or retrieved or calculated by the management entity). The management module of the chassis in which the blades and power supplies are located uses this information to manage the power reduction of blades when the system is operating in an over-subscription mode and a power supply fails. If throttling is required, the system knows the amount of power reduction available for each blade and controls the throttling by spreading it out among the blades in the system so that, ideally, no blade will cease operation altogether.

Abstract: The present invention is directed to the detection of faulty CPU heat sink coupling during system power-up. A method in accordance with an embodiment of the present invention includes: monitoring a slope of a CPU temperature rise from initial system power-up; determining if the slope of the CPU temperature rise exceeds an expected value; and in the case that the slope of the CPU temperature rise exceeds the expected value, indicating an existence of a possible fault (PFA) related to a heat sink coupled to the CPU.