[특허]Methods and apparatus for initiating and resynchronizing multi-cycle SIMD instructions

Altera Corporation

검색어에 아래의 연산자를 사용하시면 더 정확한 검색결과를 얻을 수 있습니다.
검색연산자

검색연산자	기능	검색시 예
()	우선순위가 가장 높은 연산자	예1) (나노 (기계 \| machine))
공백	두 개의 검색어(식)을 모두 포함하고 있는 문서 검색	예1) (나노 기계) 예2) 나노 장영실
\|	두 개의 검색어(식) 중 하나 이상 포함하고 있는 문서 검색	예1) (줄기세포 \| 면역) 예2) 줄기세포 \| 장영실
!	NOT 이후에 있는 검색어가 포함된 문서는 제외	예1) (황금 !백금) 예2) !image
*	검색어의 *란에 0개 이상의 임의의 문자가 포함된 문서 검색	예) semi*
""	따옴표 내의 구문과 완전히 일치하는 문서만 검색	예) "Transform and Quantization"

과학기술 지식인프라 ScienceON

특허 상세정보

MyON담기

내보내기

Methods and apparatus for initiating and resynchronizing multi-cycle SIMD instructions

국가/구분	United States(US) Patent 등록
국제특허분류(IPC7판)	G06F-009/302
미국특허분류(USC)	712/024
출원번호	US-0778303 (2007-07-16)
등록번호	US-7506137 (2009-03-17)
발명자 / 주소	Pechanek,Gerald George Strube,David Wolff,Edward A. Barry,Edwin Franklin Morris,Grayson Busboom,Carl Donald Schneider,Dale Edward
출원인 / 주소	Altera Corporation
대리인 / 주소	Priest & Goldstein, PLLC
인용정보	피인용 횟수 : 1 인용 특허 : 2

초록
▼

Techniques for adding more complex instructions and their attendant multi-cycle execution units with a single instruction multiple data stream (SIMD) very long instruction word (VLIW) processing framework are described. In one aspect, an initiation mechanism also acts as a resynchronization mechanism to read the results of multi-cycle execution. This multi-purpose mechanism operates with a short instruction word (SIW) issue of the multi-cycle instruction, in a sequence processor (SP) alone, with a VLIW, and across all processing elements (PEs) individua...

대표
청구항
▼

We claim: 1. An apparatus for executing multi-cycle instructions comprising: a first execution module for executing a first multi-cycle instruction to generate a first execution module result at the completion of a predetermined number of cycles, wherein the predetermined number of cycles is based upon the complexity of the function specified by the first multi-cycle instruction; a first instruction result register coupled to the first execution module for storing the first execution module result; and a register file coupled to the first instruction result register for storing the contents of the first instruction result register in the register file at a target register location specified by the first multi-cycle instruction prior to generating the first execution module result. 2. The apparatus of claim 1 wherein upon receiving a second multi-cycle instruction after the first execution module results have been generated, the first execution module results are stored in the register file at a second target register location specified by the second multi-cycle instruction prior to generating the results for the second multi-cycle instruction. 3. The apparatus of claim 1 wherein upon receiving a second multi-cycle instruction prior to generating the first execution module results, the first execution module stops execution of the first multi-cycle instruction and resynchronizes the first execution module to initiate execution of the second multi-cycle instruction. 4. The apparatus of claim 1 wherein the predetermined number of cycles is three or more cycles. 5. The apparatus of claim 1 further comprising: a second execution module for executing a second instruction in less than or equal to the predetermined number of cycles to generate a second execution module result, the second execution module operating in parallel with the first execution module, the second instruction specifying a target register location in the register file for storing the second execution module results at the end of executing the second instruction. 6. The apparatus of claim 1 further comprising: a flag register coupled to the first execution module for storing first execution module flags generated by the first execution module upon the completion of executing the first multi-cycle instruction. 7. The apparatus of claim 6 further comprising: arithmetic flags and condition flags coupled to the flag register to provide previously set arithmetic and condition information, wherein a multi-cycle instruction is conditionally executed based upon the state of the condition flags. 8. The apparatus of claim 6 wherein the first execution module is an integer divide unit for executing a multi-cycle integer divide instruction to generate an integer divide result as the first execution module result and generate divide arithmetic flags as the first execution module flags. 9. The apparatus of claim 6 wherein the first execution module is a floating point divide (FDIV) unit for executing a multi-cycle FDIV instruction to generate a FDIV result as the first execution module result and generate FDIV arithmetic flags as the first execution module flags. 10. The apparatus of claim 6 wherein the first execution module is a floating point square root (FSQRT) unit for executing a multi-cycle FSQRT instruction to generate a FSQRT result as the first execution module result and generate FSQRT arithmetic flags as the first execution module flags. 11. The apparatus of claim 6 wherein the first execution module is a floating point reciprocal (FRCP) unit for executing a multi-cycle FRCP instruction to generate a FRCP result as the first execution module result and generate FRCP arithmetic flags as the first execution module flags. 12. The apparatus of claim 6 wherein the first execution module is a floating point reciprocal square root (FRSQRT) unit for executing a multi-cycle FRSQRT instruction to generate a FRSQRT result as the first execution module result and generate FRSQRT arithmetic flags as the first execution module flags. 13. The apparatus of claim 6 wherein the first execution module supports execution of one or more of the following multi-cycle instructions: floating point divide (FDIV), floating point square root (FSQRT), floating point reciprocal (FRCP), and floating point reciprocal square root (FRSQRT). 14. A method for executing multi-cycle instructions comprising: executing a first multi-cycle instruction in a first execution module to generate a first execution module result at the completion of a predetermined number of cycles, wherein the predetermined number of cycles is based upon the complexity of the function specified by the first multi-cycle instruction; storing the first execution module result in a first instruction result register coupled to the first execution module; and storing the contents of the first instruction result register in a register file coupled to the first instruction result register at a target register location specified by the first multi-cycle instruction prior to generating the first execution module result. 15. The method of claim 14 further comprising: receiving, after the first execution module results have been generated, a second multi-cycle instruction that specifies a second target registers location; storing the first execution module results in the register file at the second target register location prior to generating results for the second multi-cycle instruction; and executing the second multi-cycle instruction on the first execution module. 16. The method of claim 14 further comprising: receiving a second multi-cycle instruction prior to generating the first execution module results; stopping execution of the first multi-cycle instruction on the first execution module; and resynchronizing the first execution module to initiate execution of the second multi-cycle instruction. 17. The method of claim 14 further comprising: storing first execution module flags generated by the first execution module upon the completion of executing the first multi-cycle instruction in a flag register coupled to the first execution module. 18. The method of claim 14 wherein the first execution module is an integer divide unit for executing a multi-cycle integer divide instruction to generate an integer divide result as the first execution module result and generate divide arithmetic flags as the first execution module flags. 19. The method of claim 14 wherein the-first execution module is a floating point divide (FDIV) unit for executing a multi-cycle FDIV instruction to generate a FDIV result as the first execution module result and generate FDIV arithmetic flags as the first execution module flags. 20. A method of executing multi-cycle floating point instructions comprising: executing a first multi-cycle floating point (FP) instruction in a first FP execution module to generate a first FP execution module result and arithmetic flags at the completion of a first number of cycles, wherein the first number of cycles is based upon the complexity of the floating point function specified by the FP instruction; storing the first FP execution module result in a first FP instruction result register coupled to the first FP execution module and the arithmetic flags in a flag register coupled to the first execution FP execution module; and storing the contents of the first FP instruction result register in a register file coupled to the first FP instruction result register at a target register location specified by the first multi-cycle FP instruction prior to generating the first FP execution module result. 21. The method of claim 20 further comprising: executing a second FP instruction in a second FP execution module to generate a second FP execution module result and arithmetic flags at the completion of a second number of cycles, wherein the second number of cycles is based on the complexity of the FP function specified by the second FP instruction and is independent of the first number of cycles; and storing the second FP execution module result in a register file coupled to the second FP execution module at a target register location specified by the second multi-cycle FP instruction at completion of generating the second FP execution module result. 22. The method of claim 20 further comprising: copying the contents of the first FP instruction result register in a register file coupled to the first FP instruction result register at a target register location specified by a copy instruction.

이 특허에 인용된 특허 (2)

이 특허를 인용한 특허 피인용횟수: 1

Jordan, Trevor Aaron; Palma, III, Peter Anthony. Magazine-shaped gun cleaning kit. USP201309D689692.

IPC 상위 출원인

내보내기

상세정보
내보내기 구분	파일저장 인쇄 메일전송
구성항목	관리번호, 국가코드, 자료구분, 상태, 출원번호, 출원일자, 공개번호, 공개일자, 등록번호, 등록일자, 발명명칭(한글), 발명명칭(영문), 출원인(한글), 출원인(영문), 출원인코드, 대표IPC 관리번호, 국가코드, 자료구분, 상태, 출원번호, 출원일자, 공개번호, 공개일자, 공고번호, 공고일자, 등록번호, 등록일자, 발명명칭(한글), 발명명칭(영문), 출원인(한글), 출원인(영문), 출원인코드, 대표출원인, 출원인국적, 출원인주소, 발명자, 발명자E, 발명자코드, 발명자주소, 발명자 우편번호, 발명자국적, 대표IPC, IPC코드, 요약, 미국특허분류, 대리인주소, 대리인코드, 대리인(한글), 대리인(영문), 국제공개일자, 국제공개번호, 국제출원일자, 국제출원번호, 우선권, 우선권주장일, 우선권국가, 우선권출원번호, 원출원일자, 원출원번호, 지정국, Citing Patents, Cited Patents
저장형식	Text(ASCII format) Excel format PIAS분석(.xls)
메일정보	받는사람 (필수) @ 보내는사람 (선택) @ 제목 내용 KISTI 검색결과 이메일 서비스
안내	> 총 건의 자료가 검색되었습니다. > 다운받으실 자료의 인덱스를 입력하세요. (1-10,000) > 검색결과의 순서대로 최대 10,000건 까지 다운로드가 가능합니다. > 데이타가 많을 경우 속도가 느려질 수 있습니다.(최대 2~3분 소요) > 다운로드 파일은 UTF-8 형태로 저장됩니다. 파일의 내용이 제대로 보이지 않을실 때는 웹브라우저 상단의 보기 -> 인코딩 -> 자동선택 여부를 확인하십시오 > ~ > Text(ASCII format) Excel format

과학기술 지식인프라 ScienceON

ScienceON Chatbot

특허 상세정보

Methods and apparatus for initiating and resynchronizing multi-cycle SIMD instructions

초록
▼

대표
청구항
▼

이 특허에 인용된 특허 (2)

이 특허를 인용한 특허 피인용횟수: 1

IPC 상위 출원인

과제정보

상세정보조회

원문조회

IPC 상세정보(version : 201201)

연구과제 성과물(0)

참여연구원의 다른 문헌(0)

처리중입니다.

잠시만 기다려주세요.

ScienceON Chatbot

특허 상세정보

Methods and apparatus for initiating and resynchronizing multi-cycle SIMD instructions

초록 ▼

대표청구항 ▼

이 특허에 인용된 특허 (2)

이 특허를 인용한 특허 피인용횟수: 1

IPC 상위 출원인

과제정보

상세정보조회

원문조회

IPC 상세정보(version : 201201)

연구과제 성과물(0)

참여연구원의 다른 문헌(0)

처리중입니다.

잠시만 기다려주세요.

초록
▼

대표
청구항
▼