초록 ▼

Methods and apparatus are provided for eliminating hot spots on processor chips in a symmetric multiprocessor (SMP) computer system. Some operations, in particular, floating point multiply/add, repetitively utilize portions of a processor chip to the point that the average power of the affected port

Methods and apparatus are provided for eliminating hot spots on processor chips in a symmetric multiprocessor (SMP) computer system. Some operations, in particular, floating point multiply/add, repetitively utilize portions of a processor chip to the point that the average power of the affected portions exceeds cooling capabilities. The localized temperature of the affected portions can then exceed design limits. The current invention determines when a hot spot occurs and task swaps the task to another processor prior to the localized temperature becoming too hot. Moving of tasks to processors that have data affinity with the processor reporting a hot spot is considered. Further considerations include prioritizing unused processors and those processors that have not recently reported a hot spot.

대표청구항 ▼

What is claimed is: 1. A symmetric multiprocessor computer system comprising: a plurality of processors; an operating system capable of switching and swapping tasks among the processors; a detector to detect occurrence of hot spots in the plurality of processors; wherein the operating system respon

What is claimed is: 1. A symmetric multiprocessor computer system comprising: a plurality of processors; an operating system capable of switching and swapping tasks among the processors; a detector to detect occurrence of hot spots in the plurality of processors; wherein the operating system responds to the detection of a hot spot on a first processor by switching or swapping a task from the first processor to a second processor in the plurality of processors; wherein the detector is capable of measuring temperature differences between physically separate areas within one or more of the processors in the plurality of processors; wherein the detector utilizes a pattern of instructions executed over a predetermined time interval to identify the one or more hot spots; and wherein the detector comprises: a power token generator that utilizes an opcode of an instruction and produces a power token; an accumulator that sums the power tokens; a reset timer that resets the accumulator at predetermined intervals; and a comparator that changes the digital value of a thermal warning if the value in the accumulator exceeds a predetermined value; wherein each power token is a predetermined value for the power or power density of the associated opcode. 2. A symmetric multiprocessor computer system comprising: a plurality of processors; an operating system capable of switching and swapping tasks among the processors; a detector to detect occurrence of hot spots in the plurality of processors; wherein the operating system responds to the detection of a hot spot on a first processor by switching or swapping a task from the first processor to a second processor in the plurality of processors; wherein the detector is capable of measuring temperature differences between physically separate areas within one or more of the processors in the plurality of processors; wherein the detector utilizes a pattern of instructions executed over a predetermined time interval to identify the one or more hot spots; and wherein the detector comprises: a power token generator that utilizes an opcode of an instruction and produces a power token; an averager that produces an average of a last N power tokens; and a comparator that changes the digital value of a thermal warning if the average of the last N power token exceeds a predetermined value; wherein each power token is a predetermined value for the power or power density of the associated opcode. 3. A method of improving reliability in a symmetric multiprocessor computer system comprising the steps of: checking for a hot spot occurring on one or more processors in the symmetric multiprocessor computer system; and if a hot spot is detected on an instant processor in the symmetric multiprocessor computer system, moving a task executing on the instant processor to one of a number of candidate processors in the symmetric multiprocessor computer system; wherein the step of moving the task executing on the instant processor to one of a number of candidate processors further comprises the steps of: determining which candidate processors have data affinity with the instant processor; determining which candidate processors with data affinity with the instant processor are unutilized; determining which candidate processors have reported hot spots and when the hot spot on each candidate processor was reported; prioritizing processors with data affinity with the instant processor; further prioritizing unutilized processors with data affinity with the instant processor; and further prioritizing processors that have not recently reported hot spots. 4. The method of claim 3, further comprising the step of: idling, for a first predetermined amount of time, the task on the instant processor if all of the candidate processors have reported a hot spot within a preceding second predetermined amount of time. 5. A computer storage medium having instructions or programs which, when executed by a symmetric multiprocessor computer system, causes the computer system to perform a method of improving reliability in a symmetric multiprocessor computer system, comprising the steps of: checking for a hot spot occurring on one or more processors in the symmetric multiprocessor computer system; and if a hot spot is detected on an instant processor in the symmetric multiprocessor computer system, moving a task executing on the instant processor to one of a number of candidate processors in the symmetric multiprocessor computer system; wherein the step of moving the task executing on the instant processor to one of a number of candidate processors further comprises the steps of: determining which candidate processors have data affinity with the instant processor; determining which candidate processors with data affinity with the instant processor are unutilized; determining which candidate processors have reported hot spots and when the hot spot on each candidate processor was reported; prioritizing processors with data affinity with the instant processor; further prioritizing unutilized processors with data affinity with the instant processor; and further prioritizing processors that have not recently reported hot spots. 6. The computer-readable medium of claim 5, further comprising the step of: idling, for a first predetermined amount of time, the task on the instant processor if all of the candidate processors have reported a hot spot within a preceding second predetermined amount of time.