Juniper Networks, Inc.

Intel Corporation

Netlogic Microsystems, Inc

Rmi Corporation

Akeana, Inc.

Broadcom Corporation

Ricardo Ramirez

Robert G Hathaway

David T Hass

Nazar A Zaidi

Fred Gruner

Elango Ganesan

Ramesh Panwar

Michael Freed

Abbas Rashid

Frederick R Gruner

Mark Bryers

Yen C Lee

Sankar Ramanoorthi

Mark Vilas

John Phillips

Yuhong Zhou

Gregory G Spurrier

Chau Anh Ngoc Nguyen

Gaurav Singh

Srivatsan Srinivasan

Hai Ngoc Nguyen

Wei-Hsiang Chen

Michael J Morrison

Mike Morrison

Chau Ahn Ngoc Nguyen

Ramesh Penwar

Abhijit Sil

Disclosed techniques enable processors that are capable of performing a wide range of vector operations. A processor can support multiple types of instructions. The instructions can include one or more operands, and the one or more operands can include different data types. An A-type instruction can have dependencies on a B-type instruction. An A-type instruction includes a vector instruction. A B-type instruction includes an integer instruction or a floating-point instruction. A datapath is provided to enable intermediate results from a B-type instruction to be supplied to the A-type instruction on which it depends, without utilizing register file resources, such as general-purpose register (GPR) register resources. Vector instruction performance is thereby enabled without the additional resources used with GPR register access.

Mixed-sourced dependency control for vector instructions

A network content service apparatus includes a set of compute elements adapted to perform a set of network services; and a switching fabric coupling compute elements in said set of compute elements. The set of network services includes firewall protection, Network Address Translation, Internet Protocol forwarding, bandwidth management, Secure Sockets Layer operations, Web caching, Web switching, and virtual private networking. Code operable on the compute elements enables the network services, and the compute elements are provided on blades which further include at least one input/output port.

Content service aggregation system

A system, method, and computer program product are provided for performing a register renaming operation utilizing hardware which operates in at least two modes. In operation, hardware is operated in at least two modes including a first mode for operating the hardware using a logical register of a first bit width and a second mode for operating the hardware using a logical register of a second bit width. The first bit width is twice a width of the second bit width. Additionally, a register renaming operation is performed, including renaming at least one logical register to at least one physical register of the first bit width, utilizing the hardware.

System and method for performing a register renaming operation utilizing hardware which is capable of operating in at least two modes utilizing registers of multiple widths

A system for speculative branch predictions. An embodiment of the system includes branch prediction logic, fetch logic, and branch identification logic. The branch prediction logic is configured to predict a branch path for a branch in an instruction stream. The fetch logic is coupled to the branch prediction logic. The fetch logic is configured to speculatively fetch an instruction corresponding to the predicted branch path. The branch identification logic is coupled to the branch prediction logic and the fetch logic. The branch identification logic is configured to mark the speculatively fetched instruction with a branch identifier using a universal branch identification format. The universal branch identification format includes a bit value at a bit position corresponding to the predicted branch path.

Universal branch identifier for invalidation of speculative instructions

An advanced processor comprises a plurality of multithreaded processor cores each having a data cache and instruction cache. A data switch interconnect is coupled to each of the processor cores and configured to pass information among the processor cores. A messaging network is coupled to each of the processor cores and a plurality of communication ports. In one aspect of an embodiment of the invention, the data switch interconnect is coupled to each of the processor cores by its respective data cache, and the messaging network is coupled to each of the processor cores by its respective message station. Advantages of the invention include the ability to provide high bandwidth communications between computer systems and memory in an efficient and cost-effective manner.

Advanced processor with use of bridges on a data movement ring for optimal redirection of memory and I/O traffic

A coprocessor transfers data between media access controllers and a set of cache memory without accessing main memory. The coprocessor includes a reception media access controller that receives data from a network and a transmission media access controller that transmits data to a network. A streaming output data transfer engine in the coprocessor transfers data from the reception media access controller to cache memory. A streaming input data transfer engine in the coprocessor transfers data from cache memory to the transmission media access controller. The coprocessor's data transfer engines transfer data between cache memory and the media access controllers in a single data transfer operation&#x2014;eliminating the need to store data in an intermediary memory location between the cache memory and data transfer engines. In one implementation, the coprocessor is employed in a compute engine that performs different network services, including but not limited to: 1) virtual private networking; 2) secure sockets layer processing; 3) web caching; 4) hypertext mark-up language compression; 5) virus checking; 6) firewall support; and 7) web switching.

Transferring data between cache memory and a media access controller

A compute engine includes a central processing unit coupled to a coprocessor. The coprocessor includes a sequencer coupled to a set of application engines for performing operations assigned to the compute engine. The sequencer is coupled to application engines through a set of data, enable, and control interfaces. An arbiter couples the sequencer and application engines to memory. Alternatively, the coprocessor may include multiple sequencers, with each sequencer being coupled to a different set of application engines. One set of application engines includes a media access controller for communicating with a network and a data transfer engine coupling the media access controller to the arbiter. In one implementation, application engines facilitate different network services, including but not limited to: 1) virtual private networking; 2) secure sockets layer processing; 3) web caching; 4) hypertext mark-up language compression; 5) virus checking; 6) firewall support; and 7) web switching.

Compute engine employing a coprocessor

A system includes a plurality of processing clusters and a snoop controller. A first processing cluster in the plurality of processing clusters includes a first tier cache memory coupled to a second tier cache memory. The system employs a store-create operation to obtain sole ownership of a full cache line memory location for the first processing cluster, without retrieving the memory location from other processing clusters. The system issues the store-create operation for the memory location to the first tier cache. The first tier cache forwards a memory request including the store-create operation command to the second tier cache. The second tier cache determines whether the second tier cache has sole ownership of the memory location. If the second tier cache does not have sole ownership of the memory location, ownership of the memory location is relinquished by the other processing clusters with any ownership of the memory location.

Managing ownership of a full cache line using a store-create operation

A system includes a memory, a sequencer, and a set of application engines in communication with the sequencer and memory. The set of application engines includes a streaming input engine for retrieving data from the memory and supplying the data to the set of application engines. In one embodiment, the streaming input engine includes a fetch engine with a memory opcode output and address output for accessing cache memory. The streaming input engine also includes an alignment circuit for buffering and aligning data retrieved from the memory. The alignment circuit includes a data buffer, register, byte selector, and shifter. The data buffer stores data accessed by the fetch engine. The register stores old data from the data buffer's output when the buffer sources new data. The byte selector selects data from the data buffer and the register. The shifter receives data selected by the byte selector and shifts a number of the bytes onto an output for retrieval by an output register coupled to the set of application engines.

Streaming input engine facilitating data transfers between application engines and memory

A multi-processor includes multiple processing clusters for performing assigned applications. Each cluster includes a set of compute engines, with each compute engine coupled to a set of cache memory. A compute engine includes a central processing unit and a coprocessor with a set of application engines. The central processing unit and coprocessor are coupled to the compute engine's associated cache memory. The sets of cache memory within a cluster are also coupled to one another.

Processing cluster having multiple compute engines and shared tier one caches

A system includes a memory, a sequencer, and a set of application engines in communication with the sequencer and memory. The set of application engines includes a streaming output engine with a storage engine, alignment circuit, and data buffer. The storage engine includes a memory opcode output and memory address output in communication with the memory. The storage engine employs these outputs to access the memory by supplying memory transaction opcodes and memory addresses. The alignment circuit receives data from other application engines in the set of application engines. In operation, the alignment circuit aligns data transfers from an application engine into a data word. The data buffer stores data words from the alignment circuit and transfers them to locations accessed in the memory by the storage engine.

Streaming output engine facilitating data transfers between application engines and memory

The present invention is a method and apparatus for processing full exceptions in a partial parallel processor operating on parallel operands which form into N groups. The method comprising: (a) generating P partial exception states for P partial exceptions from the partial parallel processor operating on the N groups of the parallel operands; the P partial exceptions correspond to the full exceptions; and (b) handling the P partial exceptions based on the P partial exception states.

Growing community of inventors

Sunnyvale, CA, United States of America

Ricardo Ramirez