International Business Machines Corporation

Advanced Micro Devices Corporation

Richard Heller

Kristin Marie Hazlewood

Mark Joseph Cavage

Gary Dale Williams

Jason Murray Bell

Maciej Wiatr

Rolf Geilenkeuser

Processing data in a multithreaded application to alleviate impaired or substandard performance conditions is provided. Work items that are pending processing by the multithreaded application are placed into a data structure. The work items are processed by a plurality of threads within the multithreaded application in accordance with a first algorithm, e.g., first-in first-out (FIFO). A thread within the multithreaded application is configured apart from the plurality of threads such that it processes work items in accordance with a second algorithm that differs from the first algorithm, thereby avoiding the impairing condition. For example, the thread may process a pending work item only if it has a particular characteristic. The thread restricts its own processing of work items by intermittently evaluating workflow conditions for the plurality of threads; if the workflow conditions improve or are unimpaired, then the thread does not process any work items.

Alleviating denial-of-service conditions on a server

By maintaining a substantially constant total die power during the entire lifetime of sophisticated integrated circuits, the performance degradation may be reduced. Consequently, greatly reduced guard bands for parts classification may be used compared to conventional strategies in which significant performance degradation may occur when the integrated circuits are operated on the basis of a constant supply voltage.

Compensation of operating time-related degradation of operating speed by a constant total die power mode

A method is presented for processing data in a multithreaded application to alleviate impaired or substandard performance conditions. Work items that are pending processing by the multithreaded application are placed into a data structure. The work items are processed by a plurality of threads within the multithreaded application in accordance with a first algorithm, e.g., first-in first-out (FIFO). A thread within the multithreaded application is configured apart from the plurality of threads such that it processes work items in accordance with a second algorithm that differs from the first algorithm, thereby avoiding the impairing condition. For example, the thread may process a pending work item only if it has a particular characteristic. The thread restricts its own processing of work items by intermittently evaluating workflow conditions for the plurality of threads; if the workflow conditions improve or are unimpaired, then the thread does not process any work items.

Alleviate denial-of-service conditions on a server

Each LDIF entry of a directory tree is read, split to a domain of LDIF fragments (corresponding to backend servers) and written to each LDIF fragment. The split may be accomplished through a hash function, establishing, for that iteration of LDIF entry, a write file. The LDIF entry is appended to the write file. A subsequent LDIF entry is read. A corresponding LDIF fragment is determined, which need not be different from the LDIF fragment to which the first LDIF entry was written. The current LDIF entry is written to the currently selected write file. The process continues until all LDIF entries are exhausted from the directory tree. LDIF fragments are each copied to distinct backend servers, where, each LDIF fragment may be loaded into a distributed directory data structure.

Growing community of inventors

Austin, TX, United States of America

Richard Heller