摘要:
A multi-port memory controller (MPMC) can be parameterized to selectively connect to different memory configurations. In particular, a programmable device that is combined with a DRAM in a die-stacked distributed memory in a single chip is provided with the programmable device forming the MPMC. The programmable device is parameterized to form a memory controller that can either aggregate or segment memory controller components to control different DRAM memory banks either together or separately. The aggregation or segmentation of the memory devices can be configured dynamically during operation of the programmable device.
摘要:
Methods and apparatus for implementing a stacked memory-programmable integrated circuit system-in-package are described. An aspect of the invention relates to a semiconductor device. A first integrated circuit (IC) die is provided having an array of tiles that form a programmable fabric of a programmable integrated circuit. A second IC die is stacked on the first IC die and connected therewith via inter-die connections. The second IC die includes banks of memory coupled to input/output (IO) data pins. The inter-die connections couple the IO data pins to the programmable fabric such that all of the banks of memory are accessible in parallel.
摘要:
A configurable transactional memory synchronizes transactions from clients. The configurable transactional memory includes a memory buffer and a transactional buffer. The memory buffer includes allocation control and storage, and the allocation control is configurable to selectively allocate the storage between a transactional buffer and a data buffer for the data words. The transactional buffer stores state indicating each combination of a data word and a client for which the data word is referenced by a write access in the transaction in progress from the client. The transactional arbiter generates the completion status for the transaction in progress from each client. The completion status is either committed for no collision or aborted for a collision. A collision is an access that references a data word of the transaction from the client following a write access that references the data word of another transaction in progress from another client.
摘要:
Programmable architecture for implementing a message processing system using an integrated circuit is described. In an example, specification data is received that includes attributes of the memory system. A logical description of the memory system is generated in response to the specification data. The logical description defines a memory component and a memory-interconnection component. A physical description of the memory system is generated in response to the logical description. The physical description includes memory circuitry associated with the integrated circuit defined by the memory component. The memory circuitry includes an interconnection topology defined by the memory interconnection component.
摘要:
A method and apparatus that uses device defects as an identifier. Data is written to memory of an integrated circuit. Defects are identified based upon the writing of the data. An identifier for the IC is then derived using the identification of the defects.
摘要:
Memory apparatus for a message processing system and method of providing the same is described. In one example, a message processing system (200) includes a set of n processing elements (202) for processing messages, where n is an integer greater than zero. A set of m memories (204) is provided for storing the messages, where m is an integer greater than zero. Multiplexing logic (206) is provided for coupling each of the processing elements to each of the memories. Control logic (208) is provided for driving the multiplexing logic to provide access to each of the memories among the processing elements in accordance with a gated module-n schedule.
摘要:
Programmable architecture for implementing a message processing system using an integrated circuit is described. In an example, configurable logic of an integrated circuit is configured to have a plurality of thread circuits and a memory. Messages are received to the integrated circuit for storage in the memory. The memory is accessed with the plurality of threads to concurrently process a plurality of the messages.
摘要:
A method is provided for generating a hardware description language (HDL) specification of a network packet processor from a textual language specification of the processing of network packets by the processor. The processor includes a look-ahead stage, an operation stage, an insert/remove stage, and an interleave stage. The textual language specification identifies the ports of the processor. The textual language specification includes formats for the type or types of the incoming and outgoing network packets. Each format includes the fields of the type of network packet. The textual language specification includes a procedure for each input port and for each type of incoming network packet received at the input port. Each procedure includes one or more actions for modifying the fields of a type of network packet as a function of state data and/or the fields of the type of network packet.
摘要:
A method and system for concurrent data processing, and an integrated circuit having programmable logic therefor, are described. A multi-threaded application is parsed into respective threads. Data value variables, data operators, data processing order of execution, and data result variables are identified from the threads. A code listing is generated associated with each of the threads for the data value variables, the data operators, the data processing order of execution, and the data result variables identified. Source and destination address information is associated with the data value variables and the data result variables. The source and destination address information is ordered to preserve the data processing order of execution. A configuration bitstream is generated for instantiating thread-specific processors in programmable logic, the thread-specific processors associated with the threads each having at least a portion of the data operators.
摘要:
A configurable transactional memory synchronizes transactions from clients. The configurable transactional memory includes a memory buffer and a transactional buffer. The memory buffer includes allocation control and storage, and the allocation control is configurable to selectively allocate the storage between a transactional buffer and a data buffer for the data words. The transactional buffer stores state indicating each combination of a data word and a client for which the data word is referenced by a write access in the transaction in progress from the client. The transactional arbiter generates the completion status for the transaction in progress from each client. The completion status is either committed for no collision or aborted for a collision. A collision is an access that references a data word of the transaction from the client following a write access that references the data word of another transaction in progress from another client.