Abstract: A method and system for performing diagnostic testing to speed the computer boot process. A boot process is initiated and an error counter value is read in any of memory, input/output, central processing, networking, mass storage, or other computing subsystems. The error counter values are compared to subsystem error thresholds. The method includes identifying subsets of subsystems with error counters exceeding error thresholds and then, performing diagnostic tests only on this subset of subsystems as part of the boot process. The error counter may be a correctable error counter that is incremented by an operating system error handler as it isolates subsystem errors. The method includes identifying subsystems in service less than a predefined time threshold by comparing a value stored in a power-on hours field in each subsystem to time thresholds, and including these modules in the tested subset.

Abstract: A method for evaluating portable electronic devices includes performing a first benchmark on a baseline device for a predetermined time interval, resulting in a first dataset representing work performed by the baseline device. The method further includes performing a second benchmark on a test device for the predetermined time interval resulting in a second dataset representing work performed by the test device. Finally, the test device is determined to be initially acceptable if the total amount of work performed by the test device and the total amount of work performed by the baseline device differ by less than an acceptance threshold, and is determined to be finally acceptable by performing a quotient test on the data.

Abstract: A test apparatus including a data interface configured to couple with at least one of a test device and a baseline device, and a computing device configured to perform a method including performing a first benchmark on a baseline device for a predetermined time interval, resulting in a first dataset representing work performed by the baseline devices, performing a second benchmark on a test device for the predetermined time interval resulting in a second dataset representing work performed by the test device, and using a heuristic including a number of tests to determine whether the test device has an acceptable level of performance relative to the baseline device.