1 / 10

Computer Architecture Principles Dr. Mike Frank

Computer Architecture Principles Dr. Mike Frank. CDA 5155 Summer 2003 Module #15 Multi-Cycle Operations. Multi-Cycle Operations. Depends on: Module #12 (Hazards) & Module #14 (Exceptions). Multi-cycle Operations for FP. Latency & Initiation Interval. Latency:

vila
Télécharger la présentation

Computer Architecture Principles Dr. Mike Frank

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Computer Architecture PrinciplesDr. Mike Frank CDA 5155Summer 2003 Module #15Multi-Cycle Operations

  2. Multi-Cycle Operations Depends on: Module #12 (Hazards) & Module #14 (Exceptions)

  3. Multi-cycle Operations for FP

  4. Latency & Initiation Interval • Latency: • Extra delay cycles before result is available. • Initiation interval: • Minimum number of cycles before a new input can be given to that functional unit.

  5. Pipelined Multiple-Issue FPU

  6. FPU Pipelining Issues in DLX • Notice instructions may complete out-of-order: • MULTD IF ID M1 M2 M3 M4 M5 M6 M7 ME WB • ADDD IF ID A1 A2 A3 A4 ME WB • LD IF ID EXME WB • SD IF ID EXME WB • Raises the possibility of WAW hazards, and structural hazards in MEM & WB stages. • Structural hazards may occur especially often with non-pipelined DIV unit. • Out-of-order completion impacts exception handling.

  7. Typical FP Code Seq. w. Stalls • MUL.D stalls • in ID 1 cycle waiting for new value of F4 from MEM stage of L.D • ADD.D stalls • 1 cycle in IF waiting for MUL.D to leave ID, • then 6 cycles in ID waiting for new F0 to be returned by MUL.D stage M7. • S.D stalls • 6 cycles in IF waiting for ADD.D to leave ID, • then 2 cycles in EX waiting for new F2 to be returned by ADD.D stage A4, then 1 more cycle in EX waiting for ADD.D to clear MEM stage.

  8. FP stall statistics 3rd ed.,fig. A.35,p. A-57 Per FP operation

  9. More FP stall stats 2nd ed., fig. 3.49,p. 199 (1.21 total) (0.65 total) Per inst. Overall avg. 82%of stalls

  10. ISA Design Impacts Pipelining • Variable instruction lengths & run times: • Introduces delays due to pipeline inequities. • Complicates hazard-detection & precise exceptions. • Sophisticated addressing modes: • Post-autoincrement complicates hazard detection, restarting, introduces WAR & WAW hazards. • Multiple-indirect modes complicate pipeline control & timing. • Self-modifying code: • What if you overwrite an instruction in the pipe? • Implicit condition codes: WAR hazards, restarts

More Related