Computer system performance estimator and layout configurator

• US 7,836,314 B2
• Filed: 08/21/2006
• Issued: 11/16/2010
• Est. Priority Date: 08/21/2006
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Chan, Hoi Y., Mok, Lawrence S., Kamath, Vinod, Chieu, Trieu C.
• Primary Examiner(s): Lee; Thomas
• Assistant Examiner(s): KINSEY, BRANDON MICHAEL
• Application Number:
  US11/506,876
• Publication Number:
  US 20080046766A1
• Time in Patent Office: 1,548 Days
• Field of Search: 713/300, 700/28, 700/90
• US Class Current: 713/300
• CPC Class Codes: G06F 1/206 comprising thermal management
• CPC Class Codes: H05K 7/20836 Thermal management, e.g. server temperature control

Abstract

A method, system and computer readable medium for maximizing the performance of a computer system that includes at least one computing unit. Temperature and location data for each computing unit is received by a server unit and the location of each computing unit within a given environment is reevaluated and revised to maximize the overall performance of the computer system.

Claims

1. A method for optimizing an overall performance of a computing system located within a given environment, the computing system comprising at least one computing unit, the method comprising:

querying each of the at least one computing units for data, wherein the data comprises at least temperature data corresponding to at least one sub-unit of each computing unit;

determining an overall performance value indicative of the relative performance of the computing system, wherein the overall performance value is based on weighted performance indices of each of the at least one computing units;

receiving location data indicative of the relative locations of each of the at least one computing units within the given environment;

querying each sub-unit of each computing unit for respective sub-unit data, wherein the sub-unit data includes temperature data and a list of running tasks performed by the sub-unit;

calculating a respective performance value for each of the at least one computing units, wherein each respective performance value includes a weighted performance index based on a comparison of said list of running tasks with a benchmark, wherein said calculating of said respective performance value includes comparing central processing unit and memory usages of said running tasks with said benchmark; and

calculating a revised location within the given environment for at least one of the computing units, wherein the revised locations are based on the received location data and the temperature data.

2. The method of claim 1, further comprising:

determining whether all of the at least one sub-units are running at peak performance;

determining at least one cause for any sub-units that are not running at peak performance.

3. The method of claim 2, wherein the at least one cause comprises at least one of air inlet temperature, component health status, power consumption and computational load.

4. The method of claim 1, wherein the location data indicative of the relative locations of each of the at least one computing units within the given environment comprises unique data words received from each computing unit and wherein each unique data word uniquely identifies a respective location within the given environment.

5. The method of claim 1, further comprising:

receiving temperature data from a hard disk drive sub-unit of a computing unit, wherein the temperature data from the hard disk drive sub-unit is indicative of at least one of an internal temperature of the hard disk drive sub-unit, the temperature outside the hard disk drive sub-unit and the temperature on the cover of the hard disk drive sub-unit;

determining at least one of whether any of the hard disk drive sub-unit temperatures exceed a predetermined threshold and whether an error rate associated with the hard disk drive sub-unit exceeds another predetermined threshold; and

calculating at least one of a revised motor speed and a revised desired error rate associated with the hard disk drive sub-unit based on the temperature data received from the hard disk drive sub-unit.

6. The method of claim 1, further comprising:

receiving temperature data from a memory device sub-unit of a computing unit, wherein the temperature data from the memory device sub-unit is indicative of a temperature of the memory device sub-unit;

determining whether the temperature data received from the memory device sub-unit exceeds a predetermined threshold; and

if the temperature data received from the memory device sub-unit exceeds a predetermined threshold, calculating an updated read/write request rate.

7. The method of claim 1, further comprising:

receiving data from a processor sub-unit of the computing unit, wherein the data from the processor sub-unit comprises at least one of status data corresponding to the processor sub-unit, a fan speed and an air inlet temperature corresponding to the processor sub-unit;

determining whether the processor sub-unit is in a throttling condition.

8. The method of claim 7, further comprising measuring a performance condition of the processor sub-unit.

9. The method of claim 7, further comprising:

determining an amount of power drawn by the processor sub-unit; and

comparing the determined amount of power drawn by the processor sub-unit to a predefined threshold.

10. The method of claim 9, further comprising:

calculating a revised clock frequency for the processor sub-unit;

calculating a revised supply voltage for the processor sub-unit; and

sending the revised clock frequency and revised supply voltage to a hardware management controller.

11. The method of claim 1, further comprising displaying the performance value and temperature data associated with at least one of the computing units.

12. The method of claim 1, further comprising reallocating at least one task performed by at least one sub-unit of a first computing unit to a respective sub-unit of a second computing unit, wherein said reallocation is based at least partially on temperature data associated with the first and second computing units.

13. A computer system comprising:

at least one computing unit located within a given environment, each computing unit comprising at least one sub-unit;

a server operatively connected to each of said computing units, wherein said server is operable to receive location data, performance data and temperature data from each of said computing units and further operable to determine an ideal location for each of said computing units based on the location data, performance data and temperature data received from each computing unit;

means for querying each sub-unit of each computing unit for respective sub-unit data, wherein the sub-unit data includes temperature data and a list of running tasks performed by the sub-unit;

means for calculating a respective performance value for each of the at least one computing units, wherein each respective performance value includes a weighted performance index based on a comparison of said list of running tasks with a benchmark, wherein said calculating of said respective performance value includes comparing central processing unit and memory usages of said running tasks with said benchmark; and

means for determining an overall performance value indicative of a relative performance of the computer system, wherein the overall performance value is based on weighted performance indices of each of the at least one computing units.

14. The computer system claimed in claim 13, further comprising:

means for querying each of the at least one computing units for the location data, performance data and temperature data; and

means for receiving location data indicative of the relative locations of each of the at least one computing units within the given environment.

15. A computer program product for maximizing computational performance of a computer system comprised of at least one computing unit located within a given location, the computer program product comprising:

a computer readable medium;

first program instruction means for querying each of the at least one computing units for data, wherein the data comprises at least temperature data corresponding to at least one sub-unit of each computing unit;

second program instruction means for determining an overall performance value indicative of a relative performance of the computing system, wherein the overall performance value is based on weighted performance indices of each of the at least one computing units;

third program instruction means for receiving location data indicative of the relative locations of each of the at least one computing units within the given environment;

fourth program instruction means for querying each sub-unit of each computing unit for respective sub-unit data, wherein the sub-unit data includes temperature data and a list of running tasks performed by the sub-unit;

fifth program instruction means for calculating a respective performance value for each of the at least one computing units, wherein each respective performance value includes a weighted performance index based on a comparison of said list of running tasks with a benchmark, wherein said calculating of said respective performance value includes comparing central processing unit and memory usages of said running tasks with said benchmark; and

sixth program instruction means for determining whether all of the at least one sub-units are running at peak performance.

16. The computer program product of claim 15, further comprising:

seventh program instruction means for determining at least one cause for any subunits that are not running at peak performance.