Abstract: A technique for performing barrier synchronization among a plurality of program threads. More particularly, at least one embodiment of the invention keeps track of completed tasks associated with a number of program threads using bits within a barrier register that can be updated and reassigned without incurring the amount of bus traffic as in the prior art.

Abstract: Systems and methods for enabling Reliability, Availability & Serviceability features after launching a secure environment under the control of LaGrande Technology (LT), or comparable security technology, without compromising security are provided. In one embodiment, the method comprises adding at least one specific capability to a processor to enable at least one of CPU hot-plug, CPU migration, CPU hot removal and capacity on demand.

Abstract: A method is described in which, in response to notice of a configuration event yet to happen within a network that is part of a link-based computing system, a component within said link based computing system: a) identifies networking configuration information changes to be made by components within the link-based computing system; and, b) sends instances of program code to each one of the components. Each instance of program code is to be executed by a specific component that it was sent to. Each instance of program code is customized to implement the particular one or more networking configuration information changes to be made at the specific component it was sent to.

Abstract: A discussion of improving integrated device deterministic response to test vectors. For example, limiting the transmission delay for an integrated device's response within known bounds by synchronizing an initialization training sequence to a reset deassertion. Specifically, the proposal facilitates response determinism from the DUT by synchronizing training sequences and subsequently synchronizing flit transmission to reset assertion as sampled by reference clock.

Abstract: Systems and methods for enabling Reliability, Availability & Serviceability features after launching a secure environment under the control of LaGrande Technology (LT), or comparable security technology, without compromising security are provided. In one embodiment, the method comprises adding at least one specific capability to a processor to enable at least one of CPU hot-plug, CPU migration, CPU hot removal and capacity on demand.

Abstract: Methods and apparatuses for firm partitioning of a computing platform.

Abstract: A method is described in which, in response to notice of a configuration event yet to happen within a network that is part of a link-based computing system, a component within said link based computing system: a) identifies networking configuration information changes to be made by components within the link-based computing system; and, b) sends instances of program code to each one of the components. Each instance of program code is to be executed by a specific component that it was sent to. Each instance of program code is customized to implement the particular one or more networking configuration information changes to be made at the specific component it was sent to.

Abstract: A discussion of improving integrated device deterministic response to test vectors. For example, limiting the transmission delay for an integrated device's response within known bounds by synchronizing an initialization training sequence to a reset deassertion. Specifically, the proposal facilitates response determinism from the DUT by synchronizing training sequences and subsequently synchronizing flit transmission to reset assertion as sampled by reference clock.

Abstract: A technique for performing barrier synchronization among a plurality of program threads. More particularly, at least one embodiment of the invention keeps track of completed tasks associated with a number of program threads using bits within a barrier register that can be updated and reassigned without incurring the amount of bus traffic as in the prior art.

Abstract: An embodiment of the present invention is a technique to perform error correction using a trial-and-error method. A syndrome generator provides a generation of a data syndrome of a data word modified according to a selection of at least one of error correcting parameters. The data word is associated with at least one transaction performed on a unit. A controller controls iterating the generation of the data syndrome.

Abstract: Embodiments of the invention provide an algorithm for dividing a link into one or more reduced-width links. For one embodiment of the invention, a multiplexing scheme is employed to effect a bit transmission order required by a particular cyclic redundancy check. The multiplexed output bits are then swizzled on-chip to reduce on-board routing congestion.

Abstract: In a computing system a method for performing a multiplication of a first multiplicand and a second multiplicand is presented. The computing system includes a plurality of registers, an instruction decoder, an arithmetic logic unit, and a preshifter. The first multiplicand is divided into a plurality of equal length sections. Each section includes "n" bits, where "n" is an integer greater than one. The second multiplicand is placed in a first register from the plurality of registers. A second register from the plurality of registers is cleared to zero. For each section from the plurality of sections, starting with a first section containing high order bits of the first multiplication and proceeding to a last section of the first multiplicand containing low order bits of the first multiplicand the following three substeps. First, when the low order bit of a current section is a "1", the contents of the first register are added to the contents of the second register via the arithmetic logic unit.

Abstract: An instruction is presented to the cache; the instruction includes a cache control specifier which identifies a type of data being requested. Based on the cache control specifier, one of a plurality of replacement schemes is selected for swapping a data block out of the cache.

Abstract: A low overhead way for insuring that only routines of sufficient privilege can execute on a secured page of memory in an hierarchial computer system, and for raising the privilege level of a low privilege process in an orderly and secure way is presented. This is done through the execution of a single "gateway" branch instruction standing between a procedure call by a lower privileged routine, such as a user program, and an operating system itself.

Abstract: Memory integrity is maintained in a system with a hierarchical memory using a set of explicit cache control instructions. The caches in the system have two status flags, a valid bit and a dirty bit, with each block of information stored. The operating system executes selected cache control instructions to ensure memory integrity whenever there is a possibility that integrity could be compromised.

Abstract: A memory circuit for interconnection to a computer including several memory banks, each bank including memory for the storage of information for the total address space addressable by the data processor. The memory circuit further includes a bank selection circuit connected to the data processor for receiving data representing a selected one of the memory banks. The memory circuit further includes a memory access circuit that determines from the bank selection circuit which one of the memory banks has been selected and provides alternating access between the selected memory bank and a specific memory bank in accordance with a timing signal from the data processor. The specific data bank includes display information and is accessed by the data processor during each interval when information is being output to the display.