Microsoft Technology Licensing, LLC

Intel Corporation

Microsoft Corporation

David S Callahan

Ali-Reza Adl-Tabatabai

Jan Stephen Gray

Gad Shlomo Sheaffer

Bratin Saha

Vadim Bassin

Martin Taillefer

John Joseph Duffy

Yang Ni

Arun Upadhyaya Kishan

Koichi Yamada

Landy Wang

Manish V Tangri

David L Detlefs

Burton Jordan Smith

Vance Palmer Morrison

Niklas Gustafsson

Brian M Grunkemeyer

Eric Dean Tribble

Aravind Seshadri

Mohanraj Dharmarajan

Mitchell G Lacey

Kirk M Brackebusch

Gurdev S Sethi

Vinod Grover

Paul F Ringseth

Srivaths Badrinath Copparam

Michael M Magruder

Shlomo Raikin

Yosseff Levanoni

Robert Geva

Edward George Essey

William Robert Messmer

Brian R Daugherty

Rick Molloy

Dave Detlefs

Sirius Kuttiyan

Marius Ionescu

Graefe Goetz

Vijay Narula

Vinod K Grover

Access point management is described. In an implementation, a method includes selecting one of a plurality of clients, in which each of the clients has a client profile which describes the client's experience with one or more of a plurality of access points. The selecting is performed based on one or more of the client profiles. A determination is made as to whether at least one of the access points utilized by the selected client should be replaced by another one of the access points. The determination includes establishing whether a difference between a connectivity score for the at least one access point and a connectivity score for the other access point exceeds an inertia value.

Access point management

In one embodiment, the present invention includes a method for receiving control in a kernel mode via a ring transition from a user thread during execution of an unbounded transactional memory (UTM) transaction, updating a state of a transaction status register (TSR) associated with the user thread and storing the TSR with a context of the user thread, and later restoring the context during a transition from the kernel mode to the user thread. In this way, the UTM transaction may continue on resumption of the user thread. Other embodiments are described and claimed.

Handling operating system (OS) transitions in an unbounded transactional memory (UTM) mode

A method and apparatus for accelerating a Software Transactional Memory (STM) system is herein described. A data object and metadata for the data object may each be associated with a filter, such as a hardware monitor or ephemerally held filter information. The filter is in a first, default state when no access, such as a read, from the data object has occurred during a pendancy of a transaction. Upon encountering a first access to the metadata, such as a first read, access barrier operations, such as logging of the metadata; setting a read monitor; or updating ephemeral filter information with an ephemeral/buffered store operation, are performed. Upon a subsequent/redundant access to the metadata, such as a second read, access barrier operations are elided to accelerate the subsequent access based on the filter being set to the second state to indicate a previous access occurred. Additionally, mapping of data objects to ephemeral information may be provided by software, such as through a pointer to the ephemeral information associated with the data object; an offset from a base address of the data object to the ephemeral information included associated with the data object; an index into a segment containing the ephemeral information associated with the data object; mapping the data object to the ephemeral information utilizing address arithmetic; and a hash that maps the data object to ephemeral information.

Using buffered stores or monitoring to filter redundant transactional accesses and mechanisms for mapping data to buffered metadata

In one embodiment, the present invention includes a method for executing a transactional memory (TM) transaction in a first thread, buffering a block of data in a first buffer of a cache memory of a processor, and acquiring a write monitor on the block to obtain ownership of the block at an encounter time in which data at a location of the block in the first buffer is updated. Other embodiments are described and claimed.

Mechanisms to accelerate transactions using buffered stores

A scheduler in a process of a computer system schedules tasks of a task group for concurrent execution by multiple execution contexts. The scheduler provides a mechanism that allows the task group to be cancelled by an arbitrary execution context or an asynchronous error state. When a task group is cancelled, the scheduler sets a cancel indicator in each execution context that is executing tasks corresponding to the cancelled task group and performs a cancellation process on each of the execution contexts where a cancel indicator is set. The scheduler also creates local aliases to allow task groups to be used without synchronization by execution contexts that are not directly bound to the task groups.

Cancellation mechanism for cancellable tasks including stolen task and descendent of stolen tasks from the cancellable taskgroup

Local storage may be allocated for each processing resource in a process of a computer system. Each processing resource may be virtualized and may have a one-to-one or a many-to-one correspondence with with physical processors. The contents of each local storage persist across various execution contexts that are executed by a corresponding processing resource. Each local storage may be accessed without synchronization (e.g., locks) by each execution context that is executed on a corresponding processing resource. The local storages provide the ability to segment data and store and access the data without synchronization. The local storages may be used to implement lock-free techniques such as a generalized reduction where a set of values is combined through an associative operator.

Persistent local storage for processor resources

Capacity management is described. In an implementation, a method includes executing a module on a computing device to monitor use of a product during a measurement interval to determine a maximum capacity of the product used during the measurement interval and whether a capacity failure point is reached during the measurement interval. A learned capacity limit is set based on the monitoring for determining whether one or more of a plurality of clients, if any, are to receive a list which references at least the monitored product, wherein the learned capacity limit is set such that when the capacity failure point is not reached, the learned capacity limit is set according to the maximum capacity.

Access point capacity management

Private or shared read-only memory regions. One embodiment may be practiced in a computing environment including a plurality of agents. A method includes acts for declaring one or more memory regions private to a particular agent or shared read only amongst agents by having software utilize processor level instructions to specify to hardware the private or shared read only memory address regions. The method includes an agent executing a processor level instruction to specify one or more memory regions as private to the agent or shared read-only amongst a plurality of agents. As a result of an agent executing a processor level instruction to specify one or more memory regions as private to the agent or shared read-only amongst a plurality of agents, a hardware component monitoring the one or more memory regions for conflicting accesses or prevents conflicting accesses on the one or more memory regions.

Private memory regions and coherence optimizations

A method and apparatus for utilizing hardware mechanisms of a transactional memory system is herein described. Various embodiments relate to software-based filtering of operations from read and write barriers and read isolation barriers during transactional execution. Other embodiments relate to software-implemented read barrier processing to accelerate strong atomicity. Other embodiments are also described and claimed.

Software filtering in a transactional memory system

Operating system virtual memory management for hardware transactional memory. A system includes an operating system deciding to unmap a first virtual page. As a result, the operating system removes the mapping of the first virtual page to the first physical page from the virtual memory page table. As a result, the operating system performs an action to discard transactional memory hardware state for at least the first physical page. Embodiments may further suspend hardware transactions in kernel mode. Embodiments may further perform soft page fault handling without aborting a hardware transaction, resuming the hardware transaction upon return to user mode, and even successfully committing the hardware transaction.

Operating system virtual memory management for hardware transactional memory

Service provisioning tasks can be performed to provision or deprovision services for users, reconcile overlapping services, and apply effective service properties. A service can be provisioned by receiving a request to provision the service for a user, reconciling new service properties with current service properties, from one or more service instances, to determine effective service properties, and applying the effective service properties as properties to use for the service for the user. A service can also be provisioned by receiving a request to provision the service for a user, storing a new service instance, and reconciling the new service instance with one or more previously stored service instances for the same service to determine effective service properties. A service provisioning system can perform provisioning operations for one or more internal and external service providers.

Syndication of multiple service instances

A method includes receiving control in a kernel mode via a ring transition from a user thread during execution of an unbounded transactional memory (UTM) transaction, updating a state of a transaction status register (TSR) associated with the user thread and storing the TSR with a context of the user thread, and later restoring the context during a transition from the kernel mode to the user thread. In this way, the UTM transaction may continue on resumption of the user thread.

Growing community of inventors

Mukilteo, WA, United States of America

David S Callahan