Work Files Saved Searches
   My Account                                                  Search:   Quick/Number   Boolean   Advanced   Derwent    Help   


 The Delphion Integrated View

  Buy Now:   Buy PDF- 18pp  PDF  |   File History  |   Other choices   
  Tools:  Citation Link  |  Add to Work File:    
  View:  Expand Details   |  INPADOC   |  Jump to: 
  Go to:  Derwent  
 Email this to a friend  Email this to a friend 
       
Title: US6381691: Method and apparatus for reordering memory operations along multiple execution paths in a processor
[ Derwent Title ]


Country: US United States of America

View Images High
Resolution

 Low
 Resolution

 
18 pages

 
Inventor: Altman, Erik; Danbury, CT
Gschwind, Michael K.; Danbury, CT

Assignee: International Business Machines Corporation, Armonk, NY
other patents from INTERNATIONAL BUSINESS MACHINES CORPORATION (280070) (approx. 44,393)
 News, Profiles, Stocks and More about this company

Published / Filed: 2002-04-30 / 1999-08-13

Application Number: US1999000374255

IPC Code: Advanced: G06F 9/38;
Core: more...
IPC-7: G06F 9/312;

ECLA Code: G06F9/38D4; G06F9/38E2; G06F9/38E4; G06F9/38F2; G06F9/38H2;

U.S. Class: Current: 712/236; 712/216; 712/E09.048; 712/E09.05; 712/E09.053; 712/E09.056; 712/E09.061;
Original: 712/236; 912/216;

Field of Search: 712/236,216,23 717/006

Priority Number:
1999-08-13  US1999000374255

Abstract:     A method is provided for scheduling instructions for execution along multiple paths in a Computer processing system implementing out-of-order execution. The method includes the step of selecting and moving a next instruction from its current position in a sequence of instructions to an earlier position. It is determined whether the selected instruction may reference a memory location for read-access. It is determined whether the selected instruction was previously moved over a non-selected instruction which may ambiguously reference the memory location, when the selected instruction may reference the memory location for read-access. It is determined whether the selected instruction was previously moved over a branch instruction, when the selected instruction was previously moved over the non-selected instruction. A record of the selected instruction is stored for future reference, when the selected instruction was previously moved over the branch instruction. The record includes a path specifier for indicating a path from a current locus of execution to a basic block corresponding to a in-order position of the selected instruction.

Attorney, Agent or Firm: F. Chau & Associates, LLP ;

Primary / Asst. Examiners: Coleman, Eric;

Maintenance Status: E1 Expired  Check current status

INPADOC Legal Status: Show legal status actions

Family: None

First Claim:
Show all 28 claims
What is claimed is:     1. A method for scheduling instructions for execution along multiple paths in a computer processing system implementing out-of-order execution, the method comprising the steps of:
  • selecting and moving a next instruction from its current position in a sequence of instructions to an earlier position;
  • determining whether the selected instruction may reference a memory location for read-access;
  • determining whether the selected instruction was previously moved over a non-selected instruction which may ambiguously reference the memory location, when the selected instruction may reference the memory location for read-access;
  • determining whether the selected instruction was previously moved over a branch instruction, when the selected instruction was previously moved over the non-selected instruction; and
  • storing a record of the selected instruction for future reference, when the selected instruction was previously moved over the branch instruction, the record comprising a path specifier for indicating a path from a current locus of execution to a basic block corresponding to an in-order position of the selected instruction.


Background / Summary: Show background / summary

Drawing Descriptions: Show drawing descriptions

Description: Show description

Forward References: Show 2 U.S. patent(s) that reference this one

       
U.S. References: Go to Result Set: All U.S. references   |  Forward references (2)   |   Backward references (7)   |   Citation Link

Buy
PDF
Patent  Pub.Date  Inventor Assignee   Title
Buy PDF- 10pp US5119495  1992-06 King  Bull HN Information Systems Inc. Minimizing hardware pipeline breaks using software scheduling techniques during compilation
Buy PDF- 48pp US5642512  1997-06 Tanaka  Matsushita Electric Co. Compiler with improved live range interference investigation
Buy PDF- 10pp US5721893  1998-02 Holler  Hewlett-Packard Company Exploiting untagged branch prediction cache by relocating branches
  US5758051  1998-05 Moreno et al.  International Business Machines Corporation Method and apparatus for reordering memory operations in a processor
Buy PDF- 27pp US5930507  1999-07 Nakahira  Fujitsu Limited Compiling processing apparatus
Buy PDF- 23pp US6189088  2001-02 Gshwind  International Business Machines Corporation Forwarding stored dara fetched for out-of-order load/read operation to over-taken operation read-accessing same memory location
Buy PDF- 39pp US6260190  2001-07 Ju  Hewlett-Packard Company Unified compiler framework for control and data speculation with recovery code
       
Foreign References: None

Other References:
  • Klauser et al. "Selective Eager Execution on the PolyPath Architecture", 25th Annual International Symposium on Computer Architecture, pp. 250-259, (1998).
  • Ebcioglu et al. "An Eight-Issue Tree-VLIW Processor for Dynamic Binary Translation", International Conference on Computer Design (Oct. 1998).
  • T.F. Chen "Supporting Highly Speculative Execution via Adaptive Branch Trees", Fourth International Symposium on High-Performance Computer Architecture, pp. 185-194 (1998).
  • Moudgill et al. "Register Renaming and Dynamic Speculation: An Alternative Approach", Proceedings of the 26th Annual International Symposium on Microarchitecture, pp. 202-213 (Dec. 1993).
  • Uht et al. "Disjoint Eager Execution: An Optimal Form of Speculative Execution", Proceedings of the 28th International Symposium on Microarchitecture, pp. 1-13 (Nov./Dec. 1995).


  • Inquire Regarding Licensing

    Powered by Verity


    Plaques from Patent Awards      Gallery of Obscure PatentsNominate this for the Gallery...

    Thomson Reuters Copyright © 1997-2010 Thomson Reuters 
    Subscriptions  |  Web Seminars  |  Privacy  |  Terms & Conditions  |  Site Map  |  Contact Us  |  Help