Microsoft Technology Licensing, LLC

Danny Chen

Steven J Pratschner

Hoi Huu Vo

Jay Krell

Allan James Murphy

Galen Clyde Hunt

James D Laflen

HoYuen Chau

Andrew Heth Farrier

Vijaykumar K Aski

Vijay Aski

Chris Lauren

Colin Mical Francis

Colin M Francis

Joe Chau

Eric M Vaughn

This document relates to memory access profiling. One example relates to a method or technique that can include obtaining samples collected when executing an application, the samples comprising sampled register values that were present in one or more registers of a processor when the samples were collected. The method or technique can also include identifying sampled instructions of the application that were executing when the samples were collected and other instructions of the application. The method or technique can also include evaluating the sampled instructions and one or more of the other instructions using the sampled register values to identify memory accesses by the application. The method or technique can also include outputting the identified memory accesses.

Employing sampled register values to infer memory accesses by an application

The discussion relates to automatically providing information about what code sequences contribute to a length of time a program takes to execute. One example can collect context switch and ready thread event tracing data from a program over a period of interest and identify time blocks of program threads from the period of interest. The example can distinguish individual time blocks that contribute to execution time for the period of interest from other individual time blocks that do not contribute to the execution time.

Automatic code path determination

A computing device is provided, including a processor configured to execute an application-under-test including a plurality of tasks. Each task may be executed in one or more task instances. The processor may determine respective performance data for the one or more task instances of each task. The processor may output, for display on a display, a graphical user interface (GUI) including a statistical representation of the performance data. The processor may receive, at the GUI, a selection of a task executed in a plurality of selected task instances in the application-under-test. The selected task instances may be executed in selected task execution time intervals that are at least partially non-contiguous in time. The processor may generate an aggregated view of the corresponding performance data for the selected task instances aggregated over the selected task execution time intervals. The processor may output the aggregated view for display at the GUI.

Aggregated view of task performance data

High-performance tracing can be achieved for an input program having a plurality of instructions. Techniques such as executable instruction transcription can enable execution of a plurality of instructions at a time via a run buffer. Execution information can be extracted via run buffer execution. Fidelity of execution can be preserved by executing instructions on the target processor. Other features, such as an executable extraction instruction ensemble, branch interpretation, and relative address compensation can be implemented. High quality instruction tracing can thus be achieved without the usual performance penalties.

High-performance processor instruction tracing

Methods for dynamically instrumenting a program while the program is executing are described. In some embodiments, profiling hooks may be selectively inserted into and removed from a program while the program is running. The hooks may gather profiling information, such as the frequency and duration of function calls, for a selected set of functions. The hooks may be inserted into the program without requiring a special build or modifications to the binary by modifying machine-level instructions for the program stored in system memory. The ability to selectively insert instrumentation into the machine-level instructions stored in the system memory allows a set of functions to be selected during execution of the program and hooks for each function of the set of functions to be dynamically inserted or removed during execution of the program to precisely capture profiling information for the set of functions.

Function-level dynamic instrumentation

Systems, methods and computer program products that provide a framework for the creation, editing, manipulation and use of model-based, multidimensional analysis services (MAS) cubes and using substitute dimensions in such cubes are disclosed. To permit a user to obtain better and automatic access to business intelligence, a method of generating a model-based MAS cube comprises creating a data source comprising a data warehouse in the memory via the processor, creating a data source view providing a dimension, a fact and an outrigger from the created data source, and creating the MAS cube comprising at least one measure group. Using substitute dimensions comprises finding all relevant substitutions for a measure group, creating a table for the measure group in the data source view, adding a property as the primary key of the substitute dimension and generating a query containing an inner join logical link between the substitute and original dimension.

Using dimension substitutions in OLAP cubes

A mapping is received and stored that maps elements of a data warehouse to types of a type system implemented by a data source. Program code is generated that performs a transform of data retrieved from a data source based on the mapping. Generation of the program code may include generating program code for performing a dimension transform based on the mapping, generating program code for performing a fact transform based on the mapping, and generating program code for performing an outrigger transform based on the mapping. The generated program code may then be executed to transform the data retrieved from the data source prior to loading into the data warehouse.

Growing community of inventors

Woodinville, WA, United States of America

Danny Chen