IPC분류정보
국가/구분 |
United States(US) Patent
등록
|
국제특허분류(IPC7판) |
|
출원번호 |
UP-0696502
(2007-04-04)
|
등록번호 |
US-7765382
(2010-08-13)
|
발명자
/ 주소 |
|
출원인 / 주소 |
|
대리인 / 주소 |
Allen, Dyer, Doppelt, Milbrath & Gilchrist, P.A.
|
인용정보 |
피인용 횟수 :
4 인용 특허 :
5 |
초록
▼
A semiconductor device includes a plurality of processing clusters that operate synchronously internally and arranged in a M×N matrix. Each processing cluster is formed as a plurality of processing elements and clocked buses that interconnect the processing elements within each processing clust
A semiconductor device includes a plurality of processing clusters that operate synchronously internally and arranged in a M×N matrix. Each processing cluster is formed as a plurality of processing elements and clocked buses that interconnect the processing elements within each processing cluster. A self-synchronous cluster wrapper is operative with the processing elements such that each processing cluster forms a programmable module. Self-synchronous global and local buses interconnect the processing clusters for communicating externally. An input/output circuit interconnects the global and local buses.
대표청구항
▼
That which is claimed is: 1. A semiconductor device, comprising: a plurality of processing clusters that each operate synchronously internally and arranged in a M×N matrix, each processing cluster comprising a memory and a plurality of reconfigurable processing elements each formed as an arith
That which is claimed is: 1. A semiconductor device, comprising: a plurality of processing clusters that each operate synchronously internally and arranged in a M×N matrix, each processing cluster comprising a memory and a plurality of reconfigurable processing elements each formed as an arithmetic logic unit, multiplier, or multiplier/accumulation circuit and clocked buses interconnecting the memory and processing elements within each processing cluster and an asynchronous cluster wrapper as interface circuitry and operative with said processing elements and memory; asynchronous, bidirectional and self-synchronous global and inter cluster local buses interconnecting each asynchronous cluster wrapper through which data, addresses and programs are transferred with the memory and said processing elements; and an input/output circuit interconnecting global and inter cluster local buses and comprising a clocked Boolean logic self-synchronous wrapper circuit coupled to the asynchronous global buses through which address information, programs and data are carried and coupled to the inter cluster local buses through which data is carried and wherein said processing clusters, global and inter cluster local buses and input/output circuit are configured to operate together to propagate single word reconfiguration commands on wavefront array flows. 2. The semiconductor device according to claim 1, and further comprising a plurality of functional modules interconnecting said input/output circuit and processing clusters to form an ASIC (Application Specific Integrated Circuit). 3. The semiconductor device according to claim 2, and further comprising an input/output ring having input/output ports interconnecting said functional modules. 4. The semiconductor device according to claim 1, wherein each processing cluster is programmable using a Hardware Description Language (HDL) or software. 5. The semiconductor device according to claim 1, wherein a processing cluster comprises a RISC microcontroller or state machine. 6. The semiconductor device according to claim 1, wherein said buses are programmable using a Hardware Description Language (HDL). 7. The semiconductor device according to claim 1, wherein said processing clusters and processing elements comprise circuits configured in Clock Boolean Logic. 8. An Application Specific Integrated Circuit (ASIC), comprising: a plurality of processing clusters that each operate synchronously internally and arranged in a M×N matrix and forming a programmable fabric, each processing cluster comprising a memory and a plurality of reconfigurable processing elements and each formed as an arithmetic logic unit, multiplier or multiplier/accumulator circuit and clocked buses interconnecting the memory and processing elements and an asynchronous cluster wrapper as interface circuitry and operative with said processing elements and memory; a plurality of functional modules forming an ASIC structure into which said programmable fabric is embedded; asynchronous, bi-directional global and self-synchronous inter cluster local buses interconnecting each asynchronous cluster wrapper through which data, addresses and programs are transferred with the memory and said processing elements; and an input/output circuit interconnecting the functional modules and global and inter cluster local buses and comprising a clocked Boolean logic self-synchronous wrapper circuit coupled to the asynchronous global buses through which address information, programs and data are carried and coupled to the inter cluster local buses through which data is carried and wherein said processing clusters, global and inter cluster local buses and input/output circuit are configured to operate together to propagate single word reconfiguration commands on wavefront array flows. 9. The ASIC according to claim 8, and further comprising an input/output ring having input/output ports interconnecting said functional modules. 10. A method of forming a semiconductor device, which comprises: implementing in a M×N matrix a plurality of processing clusters that each operate synchronously internally, wherein each processing cluster comprises a memory and a plurality of reconfigurable processing elements each formed as an arithmetic logic unit, multiplier, or multiplier/accumulator circuit and clocked buses interconnecting the memory and processing elements within each processing cluster and an asynchronous cluster wrapper as interface circuitry and operative with said processing elements and memory; interconnecting said asynchronous cluster wrapper with asynchronous, bidirectional and self-synchronous global and inter cluster local buses such that data, addresses and programs are transferred with the memory and processing elements; and interconnecting each global and inter cluster local buses with an interconnect circuit and comprising a clocked Boolean logic self-synchronous wrapper circuit coupled to the asynchronous global buses through which address information, programs and data are carried and coupled to the inter cluster local buses through which data is carried and wherein said processing clusters, global and inter cluster local buses and input/output circuit are configured to operate together to propagate single word reconfiguration commands on wavefront array flows. 11. The method according to claim 10, which further comprises interconnecting functional modules with the processing clusters to form an ASIC. 12. A semiconductor device, comprising: a plurality of processing clusters that each operate synchronously internally and arranged in a M×N matrix, each processing cluster comprising a memory and a plurality of reconfigurable processing elements each formed as a logic unit, multiplier or multiplier/accumulator circuit and clocked buses interconnecting the memory and processing elements within each processing cluster an asynchronous cluster wrapper as interface circuitry and operative with said processing elements and memory; asynchronous, bidirectional and self-synchronous global and inter cluster local buses interconnecting each asynchronous cluster wrapper through which data, addresses and programs are transferred with the memory and said processing elements wherein data travels on wavefronts separated by nulls; and an input/output circuit interconnecting global and inter cluster local buses and comprising a clocked Boolean logic self-synchronous wrapper circuit coupled to the asynchronous global buses through which address information, programs and data are carried and coupled to the inter cluster local buses through which data is carried and wherein said processing clusters, global and inter cluster local buses and input/output circuit are configured to operate together to propagate single word reconfiguration commands on wavefront array flows. 13. The semiconductor device according to claim 12, and further comprising a data, address and instruction control line on each bus. 14. The semiconductor device according to claim 13, and further comprising a zero and one data line on each bus. 15. The semiconductor device according to claim 12, wherein each processing cluster comprises a memory element operable with said processing elements.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.