Strategic placement of jobs for spatial elasticity in a high-performance computing environment

Patent metadata

Field	Value
Publication number	US-9317328-B2
Application number	US-201314029925-A
Country	US
Kind code	B2
Filing date	Sep 18, 2013
Priority date	May 24, 2013
Publication date	Apr 19, 2016
Grant date	Apr 19, 2016

Abstract

Accepting a job having a job size representing a number or quantity of processors; computing an expected size, and a standard deviation in size, for the accepted job; adding the expected size to the standard deviation in size to determine a sum; comparing the sum to a number or quantity of available clusters at each of a plurality of non-leaf nodes of a tree representing a high-performance computing environment; and when the number or quantity of available clusters is more than the sum at a sub-tree of the tree and, going down one level further in the sub-tree, the number of available clusters is less than the sum, selecting the sub-tree for the accepted job such that the accepted job is placed on one or more clusters associated with the selected sub-tree.

First claim

What is claimed is:

1. An apparatus comprising at least one processor and at least one memory including computer program code, wherein the at least one processor, in response to execution of the computer program code, is configured to cause the apparatus to perform at least the following:
allocating a quantity of processors for processing a job in a high-performance computing environment to provide an initial allocation, wherein the initial allocation of the quantity of processors for the job is allocated to at least one of: a first area of processors which is more densely occupied than a second area of processors based on a first probability indicative of a probability that the job will have an increased computing demand during the run time for the job; and a second area of processors which is more sparsely occupied than a first area of processors based on a second probability indicative of a probability that the job will have decreased computing demand during the run time for the job; and
modifying the initial allocation during a run time of the job by performing one or more of: increasing the allocated quantity of processors in response to an increased computing demand for the job by allocating one or more free processors from a pool of free processors which are configured to meet the increased computing demand for the job; or decreasing the allocated quantity of processors in response to a decreased computing demand for the job by reallocating one or more processors from the allocated quantity of processors to the pool of free processors
wherein the allocated one or more free processors configured to meet the increased demand for the job are within a vicinity of the quantity of processors used to provide the initial allocation.

2. The apparatus of claim 1 wherein the initial allocation is stored in a memory in a cloud environment.

3. The apparatus of claim 1 wherein the modified initial allocation is stored in a memory in a cloud environment.

4. An apparatus comprising at least one processor and at least one memory including computer program code, wherein the at least one processor, in response to execution of the computer program code, is configured to cause the apparatus to perform at least the following:
accepting a job having a job size representing a number or quantity of processors to be initially allocated for the job;
computing an expected size, and a standard deviation in size, for the accepted job;
adding the expected size to the standard deviation in size to determine a sum;
comparing the sum to a number of available clusters at each of a plurality of non-leaf nodes of a tree representing a high-performance computing environment; and
in response to the number of available clusters being more than the sum at a sub-tree of the tree and, going down one level further in the sub-tree, the number of available clusters being less than the sum, selecting the sub-tree for the accepted job such that the accepted job is placed on one or more clusters associated with the selected sub-tree,
wherein the computing the expected size, and the standard deviation in size, is performed by at least one of: using a first probability indicative of a probability that the job will be extended in size during a run time for the job; and using a second probability indicative of a probability that the job will be decreased in size during the run time for the job.

5. The apparatus of claim 4 wherein the computing the expected size, and the standard deviation in size, for the accepted job, is performed by using the first probability indicative of the probability that the job will be extended in size during the run time for the job.

6. The apparatus of claim 5 wherein the computing the expected size, and the standard deviation in size, for the accepted job, is performed by using the second probability indicative of the probability that the job will be decreased in size during the run time for the job.

7. The apparatus of claim 4 wherein the comparing is performed using a depth first search procedure.

8. The apparatus of claim 7 wherein the depth first search procedure is performed by traversing a left sub-tree at each of the plurality of non-leaf nodes prior to traversing a right sub-tree at each of the plurality of non-leaf nodes.

9. The apparatus of claim 7 wherein the depth first search procedure is performed by traversing a right sub-tree at each of the plurality of non-leaf nodes prior to traversing a left sub-tree at each of the plurality of non-leaf nodes.

Assignees

• IBM

Inventors

• Jamjoom Hani Talal
• Kumar Dinesh
• Shae Zon-Yin

Classifications

• G06F9/5061 Primary

  Partitioning or combining of resources · CPC title

• G06F9/505

  considering the load · CPC title

• G06F9/5027

  the resource being a machine, e.g. CPUs, Servers, Terminals · CPC title

• G06F9/50 Primary

  Allocation of resources, e.g. of the central processing unit [CPU] · CPC title

• G06F9/5094

  where the allocation takes into account power or heat criteria (power management in computers in general G06F1/3203; thermal management in computers in general G06F1/206) · CPC title