International Business Machines Corporation

Robert Anthony Petrillo

Mark Robert Funk

Timothy Joseph Torzewski

Thomas Owen McKinley

Donald Arthur Morrison

Richard Miles Rocheleau

Gregory Alan Chaney

Robert Carl Seemann

Arthur Douglas Smet

Robert Paul Mech

Larry Keith McMains

Daniel Dean Bender

A method, apparatus, system, and signal-bearing medium that in an embodiment use a requested address for an intent seize and a processor associated with the intent seize to determine a hash table entry. If the requested address is not found in the hash table, all hash tables for all processors are updated to anchor to the requested address. Non-intent seizes use a hash table associated with a designated processor, regardless of whether the designated processor initiated the non-intent seize. In this way, in an embodiment modified, cache line interventions may eliminated for intent seizes.

Intent seizes in a multi-processor environment

A weak record locking scheme that, while enforcing serialization of record updates, permits a read to extract data from uncommitted updates. Specifically, records are locked in association with a process or thread performing a lock. When the records affected by a transaction are initially locked for reading or updating, they are locked at the process level. When a particular record is to be read or updated as part of a transaction, it is locked at the thread level. A read transaction that will accept uncommitted changes uses a weak lock to lock the record for reading. A weak lock may be made on a record so long as the record is not already locked for updating at the thread level. While a record is weakly locked for reading, it cannot be locked for updating at the thread level. A record may thus be weakly locked even though the record is locked for updating at the process level by the same process. Weak locking thus preserves serialization of updates, while reducing delays caused by locking conflicts.

Weak record locks in database query read processing

According to the present invention, pool allocation and process assignment mechanisms create process nodal affinity in a NUMA multiprocessor system for enhanced performance. The multiprocessor system includes multiple interconnected multiprocessing nodes that each contain one or more processors and a local main memory, the system main storage being distributed among the local main memories of the multiprocessing nodes in a NUMA architecture. A pool reservation mechanism reserves pools of memory space within the logical main storage, and the pool allocation mechanism allocates those pools to real pages in the local main-memory of multiprocessing nodes. Processes to be created on the multiprocessor are given an attribute that indicates an associated pool. Upon creation, the process assignment mechanism will only assign a process to a multiprocessing node that has been allocated the pool indicated by the process' attribute. This process nodal affinity increases accesses by the assigned process to local main storage of that node, thereby enhancing system performance.

Process assignment by nodal affinity in a myultiprocessor system having

Growing community of inventors

Lanesboro, MN, United States of America

Robert Anthony Petrillo