International Business Machines Corporation

David Paul Grove

David Francis Bacon

Stephen Fink

Perry S Cheng

Peter Francis Sweeney

Matthew R Arnold

Michael James Hind

Joshua S Auerbach

Perry Sze-Din Cheng

Wei Zhang

Martin J Hirzel

Daniel John Spoonhower

Martin T Vechev

John Whaley

Perry Chang

A quality-directed adaptive analytic retraining is provided. Training example data with which to retrain a machine learning model that has been previously trained is received. The training example data is stored in a memory. The machine learning model is evaluated at least by running the machine learning model with the training example data. A normalized quality measure may be determined based on the evaluating. Whether to retrain the machine learning model is determined at least based on the normalized quality measure. Responsive to determining that the machine learning model is to be retrained, the machine learning model is retrained.

Quality-directed adaptive analytic retraining

Techniques are disclosed for schedule management. By way of example, a method for managing performance of tasks of a thread associated with a processor comprises the following steps. A request to execute a task of a first task type within the thread is received. A determination is made whether the processor is currently executing a critical section of a task of a second task type within the thread. When it is determined that the processor is not executing a critical section of the second task type within the thread, the task of the first task type is executed within the thread. When it is determined that the processor is executing a critical section of the first task type within the thread, a determination is made whether the request for execution of the task of the first task type within the thread is deferrable based on a prior execution of one or more units of the first task type. The first task type may be an overhead task type such as a garbage collection task type, and the second task type may be an application task type.

Lock deferral for real-time garbage collection

A computer readable storage medium for associating a phase with an activation of a computer program that supports garbage collection include: a plurality of stacks, each stack including at least one stack frame that includes an activation count; and a processor with logic for performing steps of: zeroing the activation count whenever the program creates a new stack frame and after garbage collection is performed; determining whether an interval has transpired during program execution; examining each stack frame's content and incrementing the activation count for each frame of the stacks once the interval has transpired; detecting the phase whose activation count is non-zero and associating the phase with the activation; and ensuring that when the phase ends, an action is immediately performed.

Detecting program phases with periodic call-stack sampling during garbage collection

Techniques are disclosed for schedule management. By way of example, a method for managing performance of tasks in threads associated with at least one processor comprises the following steps. One or more units of a first task type are executed. A count of the one or more units of the first task type executed is maintained. The count represents one or more credits accumulated by the processor for executing the one or more units of a first task type. One or more units of a second task type are executed. During execution of the one or more units of a second task type, a request to execute at least one further unit of the first task type is received. The amount of credits in the count is checked. When it is determined that there is sufficient credit in the count, the request to execute the at least one further unit of the first task type is forgone, and execution of the one or more units of the second task type continues. When it is determined that there is insufficient credit in the count, the at least one further unit of the first task type is executed. The first task type may be an overhead task type such as a garbage collection task type, and the second task type may be an application task type.

Scheduling for real-time garbage collection

A method of performing data and pointer compression includes, in a buffer which is formed between a processor and a level one cache and stores plural tags and full-word values associated with the tags, when the buffer is presented with an address, breaking the address into a line number which indexes a set of the full-word values, and a tag which is used as a key to determine whether a value in the set of full-word values includes a value associated with the presented address, if a tag in the presented address matches a tag in the buffer, returning a full-word value in the buffer which is associated with the tag, and storing the returned full-word value in a destination register of an instruction which originated the presented address, and if a tag in the presented address does not match a tag in the buffer, generating a fault and branching control to a pre-defined handler.

System and method for managing data

A method (and system) for executing a task includes executing a first task in which garbage is collected using a garbage collector, and executing a second task, the garbage collector being preemptable (e.g., preempted) by the second task.

Method and system for executing a task and medium storing a program therefor

A system for associating a phase with an activation of a computer program that supports garbage collection include: a plurality of stacks, each stack including at least one stack frame that includes an activation count; and a processor with logic for performing steps of: zeroing the activation count whenever the program creates a new stack frame and after garbage collection is performed; determining whether an interval has transpired during program execution; examining each stack frame's content and incrementing the activation count for each frame of the stacks once the interval has transpired; detecting the phase whose activation count is non-zero and associating the phase with the activation; and ensuring that when the phase ends, an action is immediately performed. Detecting phases in a running computer program, creates an activation count associated with each stack frame. The activation count is zeroed whenever a new frame is created in a stack and incremented for each frame encountered during periodic intervals. A phase is detected with an activation whose activation count is non-zero.

System for detecting program phases with periodic call-stack sampling during garbage collection

A system includes a processor for executing a collector program to perform a method (e.g., a method of collection). The method includes using an object model during a collection phase that is different than an object model used during program execution. The processor may also perform a method including assigning a hash code to at least some objects, and consulting a structure that maintains a mapping of objects to hashcode values to determine said hashcode for one of said objects. The processor may also perform a method including storing a class pointer and garbage collector state information in a single word, and accessing said class pointer by masking out non-class bits.

Method and program for space-efficient representation of objects in a garbage-collected system

Disclosed are a partial heap garbage collector, and a partial heap garbage collection method, that during collection checks the time remaining (or equivalently the time taken so far, or the work done so far) to complete the partial heap collection. In a preferred embodiment, the partial heap collection is guaranteed to complete within a fixed time interval. For example, this guarantee may be obtained by applying a worst-case execution time (WCET) and sizing a nursery so that complete evacuation of the nursery can always be achieved on time. As an alternative, a technique, referred to as syncopation, may be used to allow generational collection to be used despite variance in program behavior over the short time scales in which a nursery can be collected. Syncopation may be accomplished via allocation control or via collection control.

Generational real-time garbage collection

A system and method for managing data includes executing a set of instructions which are used for operating on compressed data and another set of instructions (e.g., different instructions) which are used for operating on uncompressed data.

A system and method for detecting phases in a running computer program, creates an activation count associated with each stack frame. The activation count is zeroed whenever a new frame is created in a stack and incremented for each frame encountered during periodic intervals. A phase is detected with an activation whose activation count is non-zero.

A method for implementing an object model for an object-oriented programming language. Also contemplated is a method whereby some object state is materialized directly in those objects deemed likely to use such a state, but is externalized for those objects deemed unlikely to use the state.

Space-efficient object models for object-oriented programming languages

A system and method for characterizing runtime behavior of a computer program executing in an execution environment, the method comprising: identifying one or more instances of yield points in a program to be executed, each yield point indicating a potential sampling operation during program execution; during program execution, in response to an identified yield point instance, ascertaining a state of the execution environment for indicating whether a sampling operation is to be performed; and, when the state of the execution environment indicates a sampling operation, recording relevant information for characterizing behavior of the execution environment. Relevant information for characterizing program behavior includes frequencies of methods executed in the program, and calling context associated with methods called by the program. Different mechanisms are provided for determining the sampling condition including the setting of a trigger bit by a runtime system, or, determining a sampling operations based on a fixed percentage of all executed yield points taken.

System and method for characterizing program behavior by sampling at selected program points

A sampling-based system and method for adaptively optimizing a computer program executing in an execution environment that comprises one or more compiler devices for providing various levels of program optimization. The system comprises a runtime measurements sub-system for monitoring execution of the computer program to be optimized, the monitoring including obtaining raw profile data samples and characterizing the raw profile data; a controller device for receiving the characterized raw profile data from the runtime measurements sub-system and analyzing the data for determining whether a level of program optimization for the executing program is to be performed by a compiler device, the controller generating a compilation plan in accordance with a determined level of optimization; and, a recompilation sub-system for receiving a compilation plan from the controller and invoking a compiler device for performing the level of program optimization of the executing program in accordance with the compilation plan.

System and method for adaptively optimizing program execution by sampling at selected program points

A method for characterizing runtime behavior of a computer program executing in an execution environment comprising: generating a call stack runtime data structure for tracking methods currently active in an executing program thread, an active method on the call stack is represented by a frame; determining condition for sampling an executing program to determine current program behavior; and, upon determination of a sampling condition, the sampling including examining at least one frame in the call stack in response to evaluate context of one or more methods represented in the call stack, the at least one frame in the call stack providing context relating to an executing program's calling structure.

Growing community of inventors

Ridgefield, CT, United States of America

David Paul Grove