Unisys Corporation

Peter John Hancock

Hans Christian Mikkelsen

Charles R Caldarale

Michael F Stanton

James W Willen

John Carter

Robert Morris

John A Miller

David R Johnson

James R McBreen

Robert Marion Malek

James F Torgerson

Lee B Hansen

Ian R Hepburn

Ronald S Tanning

Daniel J Lenz

Klaus G Dudda, Deceased

Robert T Sandelands

Lonny R Lebahn

Jerome J Witalka

A system and method for scaling the performance of a data processing system is disclosed. According to one method, a level of system performance is purchased for use with the data processing system. This purchased performance level is described in Millions of Instructions Per Second (MIPS) or a similar unit of measure. A system administrator selects which resources within the data processing system will be enabled, as well as how those resources will be configured. The enabled system resources will include one or more instruction processors. The performance of each of the enabled processors is then scaled to achieve the purchased system performance level. Performance scaling is performed in a manner that takes into account characteristics associated with the selected configuration as well as the system architecture.

System and method for scaling performance of a data processing system

Driving task dispatching by an operating system in a multiprocessor computer system is accomplished with a virtual switching queue which accepts highest priority tasks from multiple switching queues which in turn are accessed by instruction processors seeking tasks. Overload, plucking and stealing of tasks may be used to effectively balance load within the computer system. Dynamic reconfiguration used where regular queues lose underlying instruction processors.

Computer OS dispatcher operation with virtual switching queue and IP queues

Regardless of the switching queue organization, enhanced multiprocessor factor performance for software application group is accomplished via allocation of cache neighborhoods to software application groups that share memory space as identified by a user. Certain user requirements may be refused as unavailable to protect overall performance. Exclusivity and priority requirements of a user can also be honored by the operating system. Load balancing and dynamic reconfiguration features respect software application group dedications to the extent feasible.

Computer OS dispatcher operation with user controllable dedication

A network input/output processing system for sending and receiving messages between a large scale computer system and associated communications networks. Executive operating system services provide access to a control table, an input queue, and an output queue stored in the computer system's main memory. A network input/output processor responds to requests by application programs, through a communications program, for receiving input from and sending output to a network, concurrently with requests to communicate with directly attached peripheral devices such as disk drives, tape drives, and printers. The network input/output processor receives initialization, reset, and termination requests via the control table. Requests to receive input are received from the input queue. Input data is stored into buffers as directed by the input request. Requests to send output are received from the output queue. Output data is read from the buffers as directed by the output request. Executive operating system services provide for control of input data transfers and output data transfers. Special purpose Instruction Processor instructions provide the capability to build control programs for processing input and output messages used by the network input/output processor to effect message transfers, thereby minimizing host instruction pathlength for communications I/O. The system architecture minimizes internal data copy between processes by using transferable buffers as communications buffers.

System architecture for improved network input/output processing

An apparatus and method for managing a number of data processing resources to produce one or more independent and separate data processing partitions. Each partition is defined as a proper subset of the total inventory of available data processing resources. Each partition is controlled by a system control facility. All system control facilities communicate with a site configuration management system, which defines and redefines the composition of each partition. The site configuration management system can move resources from one partition to another as required. Preferably, the system control facilities and the site configuration management system are industry standard personal computers which communicate via a local area network.

Method and apparatus for performing system resource partitioning

A method of resolving conflicts when the index values of different binary bit designators are found to be equal is provided which is implementable in a virtual memory to real memory address translation scheme and alternately in a data base environment. Binary bit designators are used, each of which consists of a first compare segment, a second index segment and a third offset segment. When two indexes are found to be identical, similar groups of bits in the first segment are considered as subsidiary indexes and are utilized in sequence until all of the bits of said first segments have been used. Each time an index operation occurs a new table entry in a new table is identified, until a group of bits from the first segments of the different binary bit designators are found to be unequal. When this occurs, comparisons of the final table entries in the final table are undertaken to determine if each stored value in each final table entry is equal to the value represented by the associated first segment.

Paging system using extension tables for conflict resolution

Growing community of inventors

White Bear Lake, MN, United States of America

Peter John Hancock