1 / 17

Additional Slides for Friday Discussion Session

Additional Slides for Friday Discussion Session. The 6-stage TYPICAL pipeline:. Scalar Pipelined Processors. Interface to Memory Subsystem. Pipeline Interface to Register File:. Pipeline Interface to Register File:. add R1 <= R2 + R3. x0246. 1. 2. 3. x0123. x0123.

abena
Download Presentation

Additional Slides for Friday Discussion Session

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. Additional Slides for Friday Discussion Session

  2. The 6-stage TYPICAL pipeline: Scalar Pipelined Processors

  3. Interface to Memory Subsystem

  4. Pipeline Interface to Register File:

  5. Pipeline Interface to Register File: add R1 <= R2 + R3 x0246 1 2 3 x0123 x0123

  6. 6-stage TYP Pipeline

  7. 6-stage TYP Pipeline load R3 <= M[R4 + R5] x99 x84 3 4 5 x80 x04 +

  8. IF ID RD R1 R1 i+1: i+2: i+3: R1 b a c ALU i+2: R1 R1 i: R1 i+1: ALU i+1: i: R1 MEM FORWARDING PATHS i: R1 WB (i i+1) (i i+2) (i i+3) Forwarding Forwarding i writes R1 before i+3 via Path b via Path a reads R1 ALU Interlock and Penalty

  9. ALU Forwarding Paths

  10. IF ID RD i+1: R1 R1 i+2: i+1: R1 e d ALU i+1: R1 LOAD i:R1 MEM[] FORWARDING PATH(s) MEM i:R1 MEM[] WB i:R1 MEM[] (i i+2) (i i+1) (i i+1) i writes R1 Stall i+1 Forwarding before i+2 via Path d reads R1 Load Interlock and Penalty

  11. • Register File • • • • • Comp Comp Comp Comp 1 0 1 0 1 0 1 0 • 1 0 1 0 ALU ALU • • Load Forwarding Path • A d d r D-Cache D a • t a • LOADWB • • LoadMEM Stall IF,ID,RD,ALU

  12. Branch Instruction Interlock:

  13. IF ID RD ALU BRANCH PENALTY ALU PENALTY MEM LOAD PENALTY WB 3 Major Penalty Loops of Pipelining Performance Objective: Reduce CPI to 1.

  14. IPC x GHz Processor Performance = ----------------- inst. count “Iron Law” of Processor Performance Time 1/ProcessorPerformance = --------------- Program Cycles Time Instructions = ------------------ X ---------------- X ------------ Instruction Program Cycle (inst. count) (CPI) (cycle time)

  15. Landscape of Microprocessor Families(SPECint92)

  16. Landscape of Microprocessor Families(SPECint95)

  17. Landscape of Microprocessor Families(SPECint2000)

More Related