EE 660: Computer Architecture Out-of-Order Processors
|
|
|
- Christiana Jessica Jackson
- 5 years ago
- Views:
Transcription
1 EE 660: Computer Architecture Out-of-Order Processors Yao Zheng Department of Electrical Engineering University of Hawaiʻi at Mānoa Based on the slides of Prof. David entzlaff
2 Agenda I4 Processors I2O2 Processors I2OI Processors IO3 Processors IO2I Processors 2
3 Agenda I4 Processors I2O2 Processors I2OI Processors IO3 Processors IO2I Processors 3
4 Out-Of-Order (OOO) Introduction Name Frontend Issue riteback Commit I4 IO IO IO IO Fixed Length Pipelines Scoreboard I2O2 IO IO OOO OOO Scoreboard I2OI IO IO OOO IO Scoreboard, Reorder Buffer, and Store Buffer I03 IO OOO OOO OOO Scoreboard and Issue Queue IO2I IO OOO OOO IO Scoreboard, Issue Queue, Reorder Buffer, and Store Buffer 4
5 OOO Motivating Code Sequence Two independent sequences of instructions enable flexibility in terms of how instructions are scheduled in total order e can schedule statically in software or dynamically in hardware 5
6 I4: In-Order Front-End, Issue, riteback, Commit F D X M 6
7 I4: In-Order Front-End, Issue, riteback, Commit X0 X1 F D M0 M1 7
8 I4: In-Order Front-End, Issue, riteback, Commit (4-stage MUL) X0 X1 X2 X3 F D X2 X3 M0 M1 Y0 Y1 Y2 Y3 To avoid increasing CPI, needs full bypassing which can be expensive. To help cycle time, add Issue stage where register file read and instruction issued to Functional Un 18 it
9 I4: In-Order Front-End, Issue, riteback, Commit (4-stage MUL) SB X0 X1 X2 X3 ARF F D I X2 X3 M0 M1 Y0 Y1 Y2 Y3 ARF R SB R/ 19
10 Basic Scoreboard R1 R2 R3 R31 Data Avail. P F P: Pending, rite to Destination in flight F: hich functional unit is writing register Data Avail.: here is the write data in the functional unit pipeline A One in Data Avail. In column I means that result data is in stage I of functional unit F Can use F and Data Avail. fields to determine when to bypass and where to bypass from A one in column zero means that cycle functional unit is in the riteback stage Bits in Data Avail. field shift right every cycle. 10
11 Basic Scoreboard Data Avail. P F R1 1 R2 R3 R31 P: Pending, rite to Destination in flight F: hich functional unit is writing register Data Avail.: here is the write data in the functional unit pipeline A One in Data Avail. In column I means that result data is in stage I of functional unit F Can use F and Data Avail. fields to determine when to bypass and where to bypass from A one in column zero means that cycle functional unit is in the riteback stage Bits in Data Avail. field shift right every cycle. 11
12 0 MUL R1, R2, R3 F 1 ADDIU R11,R10,1 2 MUL R5, R1, R4 3 MUL R7, R5, R6 4 ADDIU R12,R11,1 5 ADDIU R13,R12,1 6 ADDIU R14,R12,2 D I Y0 Y1 Y2 Y3 F D I X0 X1 X2 X3 F D I I I Y0 Y1 Y2 Y3 F D D D I I I I Y0 Y1 Y2 Y3 F F F D D D D I X0 X1 X2 X3 F F F F D I X0 X1 X2 X3 F D I X0 X1 X2 X3 Cyc D I Dest Regs 1 R1 1 1 R R R7 1 1 R R R RED Indicates if we look at F Field, we can bypass on this cycle 22
13 Agenda I4 Processors I2O2 Processors I2OI Processors IO3 Processors IO2I Processors 1 3
14 I2O2: In-order Frontend/Issue, Out-oforder riteback/commit SB X0 ARF F D I M0 M1 Y0 Y1 Y2 Y3 ARF R SB R R/ 14
15 I2O2 Scoreboard Similar to I4, but we can now use it to track structural hazards on riteback port Set bit in Data Avail. according to length of pipeline Architecture conservatively stalls to avoid A hazards by stalling in Decode therefore current scoreboard sufficient. More complicated scoreboard needed for processing A Hazards 15
16 0 MUL R1, R2, R3 F 1 ADDIU R11,R10,1 2 MUL R5, R1, R4 3 MUL R7, R5, R6 4 ADDIU R12,R11,1 5 ADDIU R13,R12,1 6 ADDIU R14,R12,2 D I Y0 Y1 Y2 Y3 F D I X0 F D I I I Y0 Y1 Y2 Y3 F D D D I I I I Y0 Y1 Y2 Y3 F F F D D D D I X0 F F F F D I X0 F D I I X0 Cyc D I Dest Regs 1 R1 1 1 R R R R R R RED Indicates if we look at F Field, we can bypass on this cycle rites with two cycle latency. Structural Hazard 25
17 Early Commit Point? 0 MUL R1, R2, R3 F D I Y0 Y1 Y2 Y3 / 1 ADDIU R11,R10,1 F D I X0 / 2 MUL R5, R1, R4 F D I I I / 3 MUL R7, R5, R6 F D D D / 4 ADDIU R12,R11,1 F F F / 5 ADDIU R13,R12,1 / 6 ADDIU R14,R12,2 Limits certain types of exceptions. 26
18 Agenda I4 Processors I2O2 Processors I2OI Processors IO3 Processors IO2I Processors 1 8
19 F I2OI: In-order Frontend/Issue, Out-oforder riteback, In-order Commit D SB X0 I L0 L1 S0 Y0 Y1 Y2 Y3 PRF ARF ROB C FSB ARF SB PRF ROB FSB R/ R R/ PRF=Physical Register File(Future File), ROB=Reorder Buffer, FSB=Finished Store Buffer (1 entry) R/ R/ 27
20 State S ST V Preg -- P 1 F 1 P 1 P F P P Reorder Buffer (ROB) State: {Free, Pending, Finished} S: Speculative ST: Store bit V: Physical Register File Specifier Valid Preg: Physical Register File Specifier 20
21 State S ST V Preg -- P 1 F 1 P 1 P F P P State: {Free, Pending, Finished} S: Speculative ST: Store bit Reorder Buffer (ROB) Next instruction allocates here in D Tail of ROB Speculative because branch is in flight Instruction wrote ROB out of order Head of ROB Commit stage is waiting for Head of ROB to be finished V: Physical Register File Specifier Valid Preg: Physical Register File Specifier 21
22 Finished Store Buffer (FSB) V Op Addr Data -- Only need one entry if we only support one memory instruction inflight at a time. Single Entry FSB makes allocation trivial. If support more than one memory instruction, we need to worry about Load/Store address aliasing. 30
23 0 MUL R1, R2, R3 F D I Y0 Y1 Y2 Y3 C 1 ADDIU R11,R10,1 F D I X0 r C 2 MUL R5, R1, R4 F D I I I Y0 Y1 Y2 Y3 C 3 MUL R7, R5, R6 F D D D I I I I Y0 Y1 Y2 Y3 C 4 ADDIU R12,R11,1 F F F D D D D I X0 r C 5 ADDIU R13,R12,1 F F F F D I X0 r C 6 ADDIU R14,R12,2 F D I I X0 r C Cyc D I ROB R1 2 1 R R11 R1 R12 6 R12 R13 R13 R5 R5 R14 R14 R7 R7 Empty = free entry in ROB State of ROB at beginning of cycle Pending entry inrob Circle=Finished (Cycle after ) Last cycle before entry is freed from ROB (Cycle in C stage) Entry becomes free and is freed on next cycle 31
24 hat if First Instruction Causes an Exception? 0 MUL R1, R2, R3 F D I Y0 Y1 Y2 Y3 / 1 ADDIU R11,R10,1 F D I X0 r -- / 2 MUL R5, R1, R4 F D I I I Y0 / 3 MUL R7, R5, R6 F D D D I / 4 ADDIU R12,R11,1 F F F D / F D I... 24
25 hat About Branches? Option 2 0 BEQZ R1, target F D I X0 C 1 ADDIU R11,R10,1 F D I X0 / 2 ADDIU R5, R1, R4 F D I / 3 ADDIU R7, R5, R6 F D / T ADDIU R12,R11,1 F D I... Squash instructions in ROB when Branch commits Option 1 0 BEQZ R1, target F D I X0 C 1 ADDIU R11,R10,1 F D I - 2 ADDIU R5, R1, R4 F D - 3 ADDIU R7, R5, R6 F - T ADDIU R12,R11,1 F D I... Squash instructions earlier. Has more complexity. ROB needs many ports. Option 3 0 BEQZ R1, target F 1 ADDIU R11,R10,1 2ADDIU R5, R1, R4 3ADDIU R7, R5, R6 T ADDIU R12,R11,1 D I X0 C F D I X0 / F D F I X0 / D I X0 / F D I X0 C ait for speculative instructions to reach the Commit stage and squash in Commit stage 25
26 hat About Branches? Three possible designs with decreasing complexity based on when to squash speculative instructions and de-allocate ROB entry: 1. As soon as branch resolves 2. hen branch commits 3. hen speculative instructions reach commit Base design only allows one branch at a time. Second branch stalls in decode. Can add more bits to track multiple in-flight branches. 26
27 Avoiding Stalling Commit on Store Miss PRF ARF ROB C CSB R FSB 0 OpA F D I X0 C CSB=Committed Store Buffer 1 S F D I S0 C C C C 2 OpB F D I X0 C 3 OpC F D I X X X X C 4 OpD F D I I I I X C ith Retire Stage 0 OpA F D I X0 C 1 S F D I S0 C R R R 2 OpB F D I X0 C 3 OpC F D I X C 4 OpD F D I X C 35
28 Agenda I4 Processors I2O2 Processors I2OI Processors IO3 Processors IO2I Processors 2 8
29 IO3: In-order Frontend, Out-of-order Issue/riteback/Commit SB X0 ARF F D I Q I M0 M1 Y0 Y1 Y2 Y3 ARF SB I Q R R R/ R/ 36
30 Issue Queue (IQ) Op Imm S V Dest V P Src0 V P Src1 Op: Opcode Imm.: Immediate S: Speculative Bit V: Valid (Instruction has corresponding Src/Dest) P: Pending (aiting on operands to be produced) Instruction Ready = (!Vsrc0!Psrc0) && (!Vsrc1!Psrc1) && no structural hazards For high performance, factor in bypassing 30
31 Centralized vs. Distributed Issue Queue X0 I Q A I X0 F D I Q I M0 F D M0 Y0 I Q B I Y0 Centralized Distributed 31
32 Advanced Scoreboard R1 R2 R3 R31 Data Avail. P P: Pending, rite to Destination in flight Data Avail.: here is the write data in the pipeline and which functional unit Data Avail. now contains functional unit identfier A non-empty value in column zero means that cycle functional unit is in the riteback stage Bits in Data Avail. field shift right every cycle. 32
33 MUL R1, R2, R3 F D I Y0 Y1 Y2 Y3 1 ADDIU R11,R10,1 F D I X0 2 MUL R5, R1, R4 F D i I Y0 Y1 Y2 Y3 3 MUL R7, R5, R6 F D i I Y0 Y1 Y2 Y3 4 ADDIU R12,R11,1 F D i I X0 5 ADDIU R13,R12,1 F D i I X0 6 ADDIU R14,R12,2 F D i I X0 Cyc D I IQ R1/R2/R3 R11/R10 R5/R1/R4 R7/R5/R6 R12/R11 R13/R12 5 R14/R R14/R12 Dest/Src0/Src1, Circle denotes value present in ARF Value bypassed so no circle, present bit Value set present by Instruction 1 in cycle 5, Stage 40
34 Assume All Instruction in Issue Queue MUL R1, R2, R3 F D i I Y0 Y1 Y2 Y3 1 ADDIU R11,R10,1 F D i I X0 2 MUL R5, R1, R4 F D i I Y0 Y1 Y2 Y3 3 MUL R7, R5, R6 F D i I Y0 Y1 Y2 Y3 4 ADDIU R12,R11,1 F D i I X0 5 ADDIU R13,R12,1 F D i I X0 6 ADDIU R14,R12,2 F D i I X0 Better performance than previous? 41
35 Agenda I4 Processors I2O2 Processors I2OI Processors IO3 Processors IO2I Processors 3 5
36 F IO2I: In-order Frontend, Out-of-order Issue/riteback, In-order Commit D I Q SB X0 I L0 L1 S0 Y0 Y1 Y2 Y3 PRF ROB FSB ARF C ARF SB PRF ROB FSB IQ R/ R/ R R/ R/ R/ 42
37 MUL R1, R2, R3 F D I Y0 Y1 Y2 Y3 C 1 ADDIU R11,R10,1 F D I X0 r C 2 MUL R5, R1, R4 F D i I Y0 Y1 Y2 Y3 C 3 MUL R7, R5, R6 F D i I Y0 Y1 Y2 Y3 C 4 ADDIU R12,R11,1 F D i I X0 r C 5 ADDIU R13,R12,1 F D i I X0 r C 6 ADDIU R14,R12,2 F D i I X0 r C 0 MUL R1, R2, R3 F D I Y0 Y1 Y2 Y3 C 1 ADDIU R11,R10,1 F D I X0 r C 2 MUL R5, R1, R4 F D i I Y0 Y1 Y2 Y3 C 3 MUL R7, R5, R6 F D i I Y0 Y1 Y2 Y3 C 4 ADDIU R12,R11,1 F D i I X0 r C 5 ADDIU R13,R12,1 F D i I X0 r C 6 ADDIU R14,R12,2 F D i I X0 r C 37
38 Out-of-order 2-ide Superscalar with 1 ALU MUL R1, R2, R3 F D I Y0 Y1 Y2 Y3 C 1 ADDIU R11,R10,1 F D I X0 r C 2 MUL R5, R1, R4 F D i I Y0 Y1 Y2 Y3 C 3 MUL R7, R5, R6 F D i I Y0 Y1 Y2 Y3 C 4 ADDIU R12,R11,1 F D I X0 r C 5 ADDIU R13,R12,1 F D i I X0 r C 6 ADDIU R14,R12,2 F D i I X0 r C 38
39 Acknowledgements These slides contain material developed and copyright by: Arvind (MIT) Krste Asanovic (MIT/UCB) Joel Emer (Intel/MIT) James Hoe (CMU) John Kubiatowicz (UCB) David Patterson (UCB) Christopher Batten (Cornell) MIT material derived from course UCB material derived from course CS252 & CS152 Cornell material derived from course ECE
40 Copyright 2013 David entzlaff 40
CSCI-564 Advanced Computer Architecture
CSCI-564 Advanced Computer Architecture Lecture 8: Handling Exceptions and Interrupts / Superscalar Bo Wu Colorado School of Mines Branch Delay Slots (expose control hazard to software) Change the ISA
Computer Architecture ELEC2401 & ELEC3441
Last Time Pipeline Hazard Computer Architecture ELEC2401 & ELEC3441 Lecture 8 Pipelining (3) Dr. Hayden Kwok-Hay So Department of Electrical and Electronic Engineering Structural Hazard Hazard Control
Issue = Select + Wakeup. Out-of-order Pipeline. Issue. Issue = Select + Wakeup. OOO execution (2-wide) OOO execution (2-wide)
Out-of-order Pipeline Buffer of instructions Issue = Select + Wakeup Select N oldest, read instructions N=, xor N=, xor and sub Note: ma have execution resource constraints: i.e., load/store/fp Fetch Decode
CS 52 Computer rchitecture and Engineering Lecture 4 - Pipelining Krste sanovic Electrical Engineering and Computer Sciences University of California at Berkeley http://www.eecs.berkeley.edu/~krste! http://inst.eecs.berkeley.edu/~cs52!
This Unit: Scheduling (Static + Dynamic) CIS 501 Computer Architecture. Readings. Review Example
This Unit: Scheduling (Static + Dnamic) CIS 50 Computer Architecture Unit 8: Static and Dnamic Scheduling Application OS Compiler Firmware CPU I/O Memor Digital Circuits Gates & Transistors! Previousl:!
CPSC 3300 Spring 2017 Exam 2
CPSC 3300 Spring 2017 Exam 2 Name: 1. Matching. Write the correct term from the list into each blank. (2 pts. each) structural hazard EPIC forwarding precise exception hardwired load-use data hazard VLIW
Simple Instruction-Pipelining. Pipelined Harvard Datapath
6.823, L8--1 Simple ruction-pipelining Laboratory for Computer Science M.I.T. http://www.csg.lcs.mit.edu/6.823 Pipelined Harvard path 6.823, L8--2. I fetch decode & eg-fetch execute memory Clock period
4. (3) What do we mean when we say something is an N-operand machine?
1. (2) What are the two main ways to define performance? 2. (2) When dealing with control hazards, a prediction is not enough - what else is necessary in order to eliminate stalls? 3. (3) What is an "unbalanced"
1. (2 )Clock rates have grown by a factor of 1000 while power consumed has only grown by a factor of 30. How was this accomplished?
1. (2 )Clock rates have grown by a factor of 1000 while power consumed has only grown by a factor of 30. How was this accomplished? 2. (2 )What are the two main ways to define performance? 3. (2 )What
3. (2) What is the difference between fixed and hybrid instructions?
1. (2 pts) What is a "balanced" pipeline? 2. (2 pts) What are the two main ways to define performance? 3. (2) What is the difference between fixed and hybrid instructions? 4. (2 pts) Clock rates have grown
Simple Instruction-Pipelining. Pipelined Harvard Datapath
6.823, L8--1 Simple ruction-pipelining Updated March 6, 2000 Laboratory for Computer Science M.I.T. http://www.csg.lcs.mit.edu/6.823 Pipelined Harvard path 6.823, L8--2. fetch decode & eg-fetch execute
CMP N 301 Computer Architecture. Appendix C
CMP N 301 Computer Architecture Appendix C Outline Introduction Pipelining Hazards Pipelining Implementation Exception Handling Advanced Issues (Dynamic Scheduling, Out of order Issue, Superscalar, etc)
Department of Electrical and Computer Engineering University of Wisconsin - Madison. ECE/CS 752 Advanced Computer Architecture I.
Last (family) name: Solution First (given) name: Student I.D. #: Department of Electrical and Computer Engineering University of Wisconsin - Madison ECE/CS 752 Advanced Computer Architecture I Midterm
[2] Predicting the direction of a branch is not enough. What else is necessary?
![[2] Predicting the direction of a branch is not enough. What else is necessary?](/thumbs/74/69859984.jpg "[2] Predicting the direction of a branch is not enough. What else is necessary?")
[2] When we talk about the number of operands in an instruction (a 1-operand or a 2-operand instruction, for example), what do we mean? [2] What are the two main ways to define performance? [2] Predicting
[2] Predicting the direction of a branch is not enough. What else is necessary?
![[2] Predicting the direction of a branch is not enough. What else is necessary?](/thumbs/74/69860001.jpg "[2] Predicting the direction of a branch is not enough. What else is necessary?")
[2] What are the two main ways to define performance? [2] Predicting the direction of a branch is not enough. What else is necessary? [2] The power consumed by a chip has increased over time, but the clock
Computer Architecture. ESE 345 Computer Architecture. Design Process. CA: Design process
Computer Architecture ESE 345 Computer Architecture Design Process 1 The Design Process "To Design Is To Represent" Design activity yields description/representation of an object -- Traditional craftsman
ECE 3401 Lecture 23. Pipeline Design. State Table for 2-Cycle Instructions. Control Unit. ISA: Instruction Specifications (for reference)
ECE 3401 Lecture 23 Pipeline Design Control State Register Combinational Control Logic New/ Modified Control Word ISA: Instruction Specifications (for reference) P C P C + 1 I N F I R M [ P C ] E X 0 PC
ICS 233 Computer Architecture & Assembly Language
ICS 233 Computer Architecture & Assembly Language Assignment 6 Solution 1. Identify all of the RAW data dependencies in the following code. Which dependencies are data hazards that will be resolved by
/ : Computer Architecture and Design
16.482 / 16.561: Computer Architecture and Design Summer 2015 Homework #5 Solution 1. Dynamic scheduling (30 points) Given the loop below: DADDI R3, R0, #4 outer: DADDI R2, R1, #32 inner: L.D F0, 0(R1)
Fall 2011 Prof. Hyesoon Kim
Fall 2011 Prof. Hyesoon Kim Add: 2 cycles FE_stage add r1, r2, r3 FE L ID L EX L MEM L WB L add add sub r4, r1, r3 sub sub add add mul r5, r2, r3 mul sub sub add add mul sub sub add add mul sub sub add
Project Two RISC Processor Implementation ECE 485
Project Two RISC Processor Implementation ECE 485 Chenqi Bao Peter Chinetti November 6, 2013 Instructor: Professor Borkar 1 Statement of Problem This project requires the design and test of a RISC processor
Portland State University ECE 587/687. Branch Prediction
Portland State University ECE 587/687 Branch Prediction Copyright by Alaa Alameldeen and Haitham Akkary 2015 Branch Penalty Example: Comparing perfect branch prediction to 90%, 95%, 99% prediction accuracy,
CS 152 Computer Architecture and Engineering. Lecture 17: Synchronization and Sequential Consistency
CS 152 Computer Architecture and Engineering Lecture 17: Synchronization and Sequential Consistency Dr. George Michelogiannakis EECS, University of California at Berkeley CRD, Lawrence Berkeley National
Implementing the Controller. Harvard-Style Datapath for DLX
6.823, L6--1 Implementing the Controller Laboratory for Computer Science M.I.T. http://www.csg.lcs.mit.edu/6.823 6.823, L6--2 Harvard-Style Datapath for DLX Src1 ( j / ~j ) Src2 ( R / RInd) RegWrite MemWrite
Unit 6: Branch Prediction
CIS 501: Computer Architecture Unit 6: Branch Prediction Slides developed by Joe Devie/, Milo Mar4n & Amir Roth at Upenn with sources that included University of Wisconsin slides by Mark Hill, Guri Sohi,
Pipelining. Traditional Execution. CS 365 Lecture 12 Prof. Yih Huang. add ld beq CS CS 365 2
Pipelining CS 365 Lecture 12 Prof. Yih Huang CS 365 1 Traditional Execution 1 2 3 4 1 2 3 4 5 1 2 3 add ld beq CS 365 2 1 Pipelined Execution 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
Simple Instruction-Pipelining (cont.) Pipelining Jumps
6.823, L9--1 Simple ruction-pipelining (cont.) + Interrupts Updated March 6, 2000 Laboratory for Computer Science M.I.T. http://www.csg.lcs.mit.edu/6.823 Src1 ( j / ~j ) Src2 ( / Ind) Pipelining Jumps
Exploring the Potential of Instruction-Level Parallelism of Exposed Datapath Architectures with Buffered Processing Units
Exploring the Potential of Instruction-Level Parallelism of Exposed Datapath Architectures with Buffered Processing Units Anoop Bhagyanath and Klaus Schneider Embedded Systems Chair University of Kaiserslautern
CMU Introduction to Computer Architecture, Spring 2015 HW 2: ISA Tradeoffs, Microprogramming and Pipelining
CMU 18-447 Introduction to Computer Architecture, Spring 2015 HW 2: ISA Tradeoffs, Microprogramming and Pipelining Instructor: Prof Onur Mutlu TAs: Rachata Ausavarungnirun, Kevin Chang, Albert Cho, Jeremie
Computer Engineering Department. CC 311- Computer Architecture. Chapter 4. The Processor: Datapath and Control. Single Cycle
Computer Engineering Department CC 311- Computer Architecture Chapter 4 The Processor: Datapath and Control Single Cycle Introduction The 5 classic components of a computer Processor Input Control Memory
CS 152 Computer Architecture and Engineering. Lecture 17: Synchroniza<on and Sequen<al Consistency. Last Time, Lecture 16: GPUs. NOW Handout Page 1
CS 152 Computer Architecture and Engineering Lecture 17: Synchroniza
Microprocessor Power Analysis by Labeled Simulation
Microprocessor Power Analysis by Labeled Simulation Cheng-Ta Hsieh, Kevin Chen and Massoud Pedram University of Southern California Dept. of EE-Systems Los Angeles CA 989 Outline! Introduction! Problem
Lecture: Pipelining Basics
Lecture: Pipelining Basics Topics: Performance equations wrap-up, Basic pipelining implementation Video 1: What is pipelining? Video 2: Clocks and latches Video 3: An example 5-stage pipeline Video 4:
Performance Prediction Models
Performance Prediction Models Shruti Padmanabha, Andrew Lukefahr, Reetuparna Das, Scott Mahlke {shrupad,lukefahr,reetudas,mahlke}@umich.edu Gem5 workshop Micro 2012 December 2, 2012 Composite Cores Pushing
Compiling Techniques
Lecture 11: Introduction to 13 November 2015 Table of contents 1 Introduction Overview The Backend The Big Picture 2 Code Shape Overview Introduction Overview The Backend The Big Picture Source code FrontEnd
Scalable Store-Load Forwarding via Store Queue Index Prediction
Scalable Store-Load Forwarding via Store Queue Index Prediction Tingting Sha, Milo M.K. Martin, Amir Roth University of Pennsylvania {shatingt, milom, amir}@cis.upenn.edu addr addr addr (CAM) predictor
Performance, Power & Energy. ELEC8106/ELEC6102 Spring 2010 Hayden Kwok-Hay So
Performance, Power & Energy ELEC8106/ELEC6102 Spring 2010 Hayden Kwok-Hay So Recall: Goal of this class Performance Reconfiguration Power/ Energy H. So, Sp10 Lecture 3 - ELEC8106/6102 2 PERFORMANCE EVALUATION
NCU EE -- DSP VLSI Design. Tsung-Han Tsai 1
NCU EE -- DSP VLSI Design. Tsung-Han Tsai 1 Multi-processor vs. Multi-computer architecture µp vs. DSP RISC vs. DSP RISC Reduced-instruction-set Register-to-register operation Higher throughput by using
Lecture 12: Pipelined Implementations: Control Hazards and Resolutions
18-447 Lectre 12: Pipelined Implementations: Control Hazards and Resoltions S 09 L12-1 James C. Hoe Dept of ECE, CU arch 2, 2009 Annoncements: Spring break net week!! Project 2 de the week after spring
INF2270 Spring Philipp Häfliger. Lecture 8: Superscalar CPUs, Course Summary/Repetition (1/2)
INF2270 Spring 2010 Philipp Häfliger Summary/Repetition (1/2) content From Scalar to Superscalar Lecture Summary and Brief Repetition Binary numbers Boolean Algebra Combinational Logic Circuits Encoder/Decoder
Worst-Case Execution Time Analysis. LS 12, TU Dortmund
Worst-Case Execution Time Analysis Prof. Dr. Jian-Jia Chen LS 12, TU Dortmund 02, 03 May 2016 Prof. Dr. Jian-Jia Chen (LS 12, TU Dortmund) 1 / 53 Most Essential Assumptions for Real-Time Systems Upper
ECE 172 Digital Systems. Chapter 12 Instruction Pipelining. Herbert G. Mayer, PSU Status 7/20/2018
ECE 172 Digital Systems Chapter 12 Instruction Pipelining Herbert G. Mayer, PSU Status 7/20/2018 1 Syllabus l Scheduling on Pipelined Architecture l Idealized Pipeline l Goal of Scheduling l Causes for
UNIVERSITY OF WISCONSIN MADISON
CS/ECE 252: INTRODUCTION TO COMPUTER ENGINEERING UNIVERSITY OF WISCONSIN MADISON Prof. Gurindar Sohi TAs: Minsub Shin, Lisa Ossian, Sujith Surendran Midterm Examination 2 In Class (50 minutes) Friday,
Department of Electrical and Computer Engineering The University of Texas at Austin
Department of Electrical and Computer Engineering The University of Texas at Austin EE 360N, Fall 2004 Yale Patt, Instructor Aater Suleman, Huzefa Sanjeliwala, Dam Sunwoo, TAs Exam 1, October 6, 2004 Name:
Branch Prediction using Advanced Neural Methods
Branch Prediction using Advanced Neural Methods Sunghoon Kim Department of Mechanical Engineering University of California, Berkeley shkim@newton.berkeley.edu Abstract Among the hardware techniques, two-level
Lecture 3, Performance
Lecture 3, Performance Repeating some definitions: CPI Clocks Per Instruction MHz megahertz, millions of cycles per second MIPS Millions of Instructions Per Second = MHz / CPI MOPS Millions of Operations
Lecture 3, Performance
Repeating some definitions: Lecture 3, Performance CPI MHz MIPS MOPS Clocks Per Instruction megahertz, millions of cycles per second Millions of Instructions Per Second = MHz / CPI Millions of Operations
Che-Wei Chang Department of Computer Science and Information Engineering, Chang Gung University
Che-Wei Chang chewei@mail.cgu.edu.tw Department of Computer Science and Information Engineering, Chang Gung University } 2017/11/15 Midterm } 2017/11/22 Final Project Announcement 2 1. Introduction 2.
Pipelined Datapath. Reading. Sections Practice Problems: 1, 3, 8, 12 (2) Lecture notes from MKP, H. H. Lee and S.
Pipelined Datapath Lectre notes from KP, H. H. Lee and S. Yalamanchili Sections 4.5 4. Practice Problems:, 3, 8, 2 Reading (2) Pipeline Performance Assme time for stages is v ps for register read or write
Computer Architecture
Lecture 2: Iakovos Mavroidis Computer Science Department University of Crete 1 Previous Lecture CPU Evolution What is? 2 Outline Measurements and metrics : Performance, Cost, Dependability, Power Guidelines
Performance, Power & Energy
Recall: Goal of this class Performance, Power & Energy ELE8106/ELE6102 Performance Reconfiguration Power/ Energy Spring 2010 Hayden Kwok-Hay So H. So, Sp10 Lecture 3 - ELE8106/6102 2 What is good performance?
Mothers of Pipelines
Electronic Notes in Theoretical Computer Science 174 (2007) 7 22 www.elsevier.com/locate/entcs Mothers of Pipelines Sava Krstić, Robert B. Jones, and John O Leary 1,2 Strategic CAD Labs, Intel Corporation,
ENEE350 Lecture Notes-Weeks 14 and 15
Pipelining & Amdahl s Law ENEE350 Lecture Notes-Weeks 14 and 15 Pipelining is a method of processing in which a problem is divided into a number of sub problems and solved and the solu8ons of the sub problems
System Data Bus (8-bit) Data Buffer. Internal Data Bus (8-bit) 8-bit register (R) 3-bit address 16-bit register pair (P) 2-bit address
Intel 8080 CPU block diagram 8 System Data Bus (8-bit) Data Buffer Registry Array B 8 C Internal Data Bus (8-bit) F D E H L ALU SP A PC Address Buffer 16 System Address Bus (16-bit) Internal register addressing:
ECE290 Fall 2012 Lecture 22. Dr. Zbigniew Kalbarczyk
ECE290 Fall 2012 Lecture 22 Dr. Zbigniew Kalbarczyk Today LC-3 Micro-sequencer (the control store) LC-3 Micro-programmed control memory LC-3 Micro-instruction format LC -3 Micro-sequencer (the circuitry)
SISD SIMD. Flynn s Classification 8/8/2016. CS528 Parallel Architecture Classification & Single Core Architecture C P M
8/8/26 S528 arallel Architecture lassification & Single ore Architecture arallel Architecture lassification A Sahu Dept of SE, IIT Guwahati A Sahu Flynn s lassification SISD Architecture ategories M SISD
Circuit Modeling for Practical Many-core Architecture Design Exploration
Circuit Modeling for Practical Many-core Architecture Design Exploration Redefining design abstractions Dean Truong Bevan Baas VLSI Computation Lab University of California, Davis Outline Motivation Circuit
Processor Design & ALU Design
3/8/2 Processor Design A. Sahu CSE, IIT Guwahati Please be updated with http://jatinga.iitg.ernet.in/~asahu/c22/ Outline Components of CPU Register, Multiplexor, Decoder, / Adder, substractor, Varity of
Arithmetic and Logic Unit First Part
Arithmetic and Logic Unit First Part Arquitectura de Computadoras Arturo Díaz D PérezP Centro de Investigación n y de Estudios Avanzados del IPN adiaz@cinvestav.mx Arquitectura de Computadoras ALU1-1 Typical
Building a Computer. Quiz #2 on 10/31, open book and notes. (This is the last lecture covered) I wonder where this goes? L16- Building a Computer 1
Building a Computer I wonder where this goes? B LU MIPS Kit Quiz # on /3, open book and notes (This is the last lecture covered) Comp 4 Fall 7 /4/7 L6- Building a Computer THIS IS IT! Motivating Force
Control. Control. the ALU. ALU control signals 11/4/14. Next: control. We built the instrument. Now we read music and play it...
// CS 2, Fall 2! CS 2, Fall 2! We built the instrument. Now we read music and play it... A simple implementa/on uc/on uct r r 2 Write r Src Src Extend 6 Mem Next: path 7-2 CS 2, Fall 2! signals CS 2, Fall
EXAMPLES 4/12/2018. The MIPS Pipeline. Hazard Summary. Show the pipeline diagram. Show the pipeline diagram. Pipeline Datapath and Control
The MIPS Pipeline CSCI206 - Computer Organization & Programming Pipeline Datapath and Control zybook: 11.6 Developed and maintained by the Bucknell University Computer Science Department - 2017 Hazard
ALU A functional unit
ALU A functional unit that performs arithmetic operations such as ADD, SUB, MPY logical operations such as AND, OR, XOR, NOT on given data types: 8-,16-,32-, or 64-bit values A n-1 A n-2... A 1 A 0 B n-1
CSE Computer Architecture I
Execution Sequence Summary CSE 30321 Computer Architecture I Lecture 17 - Multi Cycle Control Michael Niemier Department of Computer Science and Engineering Step name Instruction fetch Instruction decode/register
Developing an Architecture for a Single-Flux Quantum Based Reconfigurable Accelerator
Developing an Architecture for a Single-Flux Quantum Based Reconfigurable Accelerator F. Mehdipour, Hiroaki Honda*, * H. Kataoka, K. Inoue and K. Murakami Graduate School of Information Science and Electrical
CS 700: Quantitative Methods & Experimental Design in Computer Science
CS 700: Quantitative Methods & Experimental Design in Computer Science Sanjeev Setia Dept of Computer Science George Mason University Logistics Grade: 35% project, 25% Homework assignments 20% midterm,
Evaluating Overheads of Multi-bit Soft Error Protection Techniques at Hardware Level Sponsored by SRC and Freescale under SRC task number 2042
Evaluating Overheads of Multi-bit Soft Error Protection Techniques at Hardware Level Sponsored by SR and Freescale under SR task number 2042 Lukasz G. Szafaryn, Kevin Skadron Department of omputer Science
Pipeline no Prediction. Branch Delay Slots A. From before branch B. From branch target C. From fall through. Branch Prediction
Pipeline no Prediction Branching completes in 2 cycles We know the target address after the second stage? PC fetch Instruction Memory Decode Check the condition Calculate the branch target address PC+4
Administrivia. Course Objectives. Overview. Lecture Notes Week markem/cs333/ 2. Staff. 3. Prerequisites. 4. Grading. 1. Theory and application
Administrivia 1. markem/cs333/ 2. Staff 3. Prerequisites 4. Grading Course Objectives 1. Theory and application 2. Benefits 3. Labs TAs Overview 1. What is a computer system? CPU PC ALU System bus Memory
A Second Datapath Example YH16
A Second Datapath Example YH16 Lecture 09 Prof. Yih Huang S365 1 A 16-Bit Architecture: YH16 A word is 16 bit wide 32 general purpose registers, 16 bits each Like MIPS, 0 is hardwired zero. 16 bit P 16
Clock-driven scheduling
Clock-driven scheduling Also known as static or off-line scheduling Michal Sojka Czech Technical University in Prague, Faculty of Electrical Engineering, Department of Control Engineering November 8, 2017
CMP 338: Third Class
CMP 338: Third Class HW 2 solution Conversion between bases The TINY processor Abstraction and separation of concerns Circuit design big picture Moore s law and chip fabrication cost Performance What does
Verilog HDL:Digital Design and Modeling. Chapter 11. Additional Design Examples. Additional Figures
Chapter Additional Design Examples Verilog HDL:Digital Design and Modeling Chapter Additional Design Examples Additional Figures Chapter Additional Design Examples 2 Page 62 a b y y 2 y 3 c d e f Figure
CODE GENERATION REGISTER ALLOCATION. Goal. Interplay between. Translate intermediate code into target code
CODE GENERATION Goal Translate intermediate code into target code Interplay between Register Allocation Instruction Selection Instruction Scheduling 1 REGISTER ALLOCATION 1 REGISTER ALLOCATION Motivation
Professor Lee, Yong Surk. References. Topics Microprocessor & microcontroller. High Performance Microprocessor Architecture Overview
This lecture was mae by a generous contribution of C & S Technology corporation. (http://( http://www.cnstec.com) There is no copyright on this lecture. Processor Laboratory homepage, (http://( http://mpu.yonsei.ac.kr)
Implementing Absolute Addressing in a Motorola Processor (DRAFT)
Implementing Absolute Addressing in a Motorola 68000 Processor (DRAFT) Dylan Leigh (s3017239) 2009 This project involved the further development of a limited 68000 processor core, developed by Dylan Leigh
Introduction to CMOS VLSI Design Lecture 1: Introduction
Introduction to CMOS VLSI Design Lecture 1: Introduction David Harris, Harvey Mudd College Kartik Mohanram and Steven Levitan University of Pittsburgh Introduction Integrated circuits: many transistors
L07-L09 recap: Fundamental lesson(s)!
L7-L9 recap: Fundamental lesson(s)! Over the next 3 lectures (using the IPS ISA as context) I ll explain:! How functions are treated and processed in assembly! How system calls are enabled in assembly!
Spiral 1 / Unit 3
-3. Spiral / Unit 3 Minterm and Maxterms Canonical Sums and Products 2- and 3-Variable Boolean Algebra Theorems DeMorgan's Theorem Function Synthesis use Canonical Sums/Products -3.2 Outcomes I know the
COVER SHEET: Problem#: Points
EEL 4712 Midterm 3 Spring 2017 VERSION 1 Name: UFID: Sign here to give permission for your test to be returned in class, where others might see your score: IMPORTANT: Please be neat and write (or draw)
CA Compiler Construction
CA4003 - Compiler Construction Code Generation to MIPS David Sinclair Code Generation The generation o machine code depends heavily on: the intermediate representation;and the target processor. We are
CIS 371 Computer Organization and Design
CIS 371 Computer Organization and Design Unit 7: Caches Based on slides by Prof. Amir Roth & Prof. Milo Martin CIS 371: Comp. Org. Prof. Milo Martin Caches 1 This Unit: Caches I$ Core L2 D$ Main Memory
Transposition Mechanism for Sparse Matrices on Vector Processors
Transposition Mechanism for Sparse Matrices on Vector Processors Pyrrhos Stathis Stamatis Vassiliadis Sorin Cotofana Electrical Engineering Department, Delft University of Technology, Delft, The Netherlands
Lecture 9: Control Hazard and Resolution. James C. Hoe Department of ECE Carnegie Mellon University
18 447 Lectre 9: Control Hazard and Resoltion James C. Hoe Department of ECE Carnegie ellon University 18 447 S18 L09 S1, James C. Hoe, CU/ECE/CALC, 2018 Yor goal today Hosekeeping simple control flow
CMPEN 411 VLSI Digital Circuits Spring Lecture 19: Adder Design
CMPEN 411 VLSI Digital Circuits Spring 2011 Lecture 19: Adder Design [Adapted from Rabaey s Digital Integrated Circuits, Second Edition, 2003 J. Rabaey, A. Chandrakasan, B. Nikolic] Sp11 CMPEN 411 L19
LRADNN: High-Throughput and Energy- Efficient Deep Neural Network Accelerator using Low Rank Approximation
LRADNN: High-Throughput and Energy- Efficient Deep Neural Network Accelerator using Low Rank Approximation Jingyang Zhu 1, Zhiliang Qian 2, and Chi-Ying Tsui 1 1 The Hong Kong University of Science and
Hardware Acceleration of DNNs
Lecture 12: Hardware Acceleration of DNNs Visual omputing Systems Stanford S348V, Winter 2018 Hardware acceleration for DNNs Huawei Kirin NPU Google TPU: Apple Neural Engine Intel Lake rest Deep Learning
Review. Combined Datapath
Review Topics:. A single cycle implementation 2. State Diagrams. A mltiple cycle implementation COSC 22: Compter Organization Instrctor: Dr. Amir Asif Department of Compter Science York University Handot
A Simple Architectural Enhancement for Fast and Flexible Elliptic Curve Cryptography over Binary Finite Fields GF(2 m )
A Simple Architectural Enhancement for Fast and Flexible Elliptic Curve Cryptography over Binary Finite Fields GF(2 m ) Stefan Tillich, Johann Großschädl Institute for Applied Information Processing and
Part I: Definitions and Properties
Turing Machines Part I: Definitions and Properties Finite State Automata Deterministic Automata (DFSA) M = {Q, Σ, δ, q 0, F} -- Σ = Symbols -- Q = States -- q 0 = Initial State -- F = Accepting States
Leveraging Transactional Memory for a Predictable Execution of Applications Composed of Hard Real-Time and Best-Effort Tasks
Leveraging Transactional Memory for a Predictable Execution of Applications Composed of Hard Real-Time and Best-Effort Tasks Stefan Metzlaff, Sebastian Weis, and Theo Ungerer Department of Computer Science,
Energy Delay Optimization
EE M216A.:. Fall 21 Lecture 8 Energy Delay Optimization Prof. Dejan Marković ee216a@gmail.com Some Common Questions Is sizing better than V DD for energy reduction? What are the optimal values of gate
Numbers and Arithmetic
Numbers and Arithmetic See: P&H Chapter 2.4 2.6, 3.2, C.5 C.6 Hakim Weatherspoon CS 3410, Spring 2013 Computer Science Cornell University Big Picture: Building a Processor memory inst register file alu
Numbers and Arithmetic
Numbers and Arithmetic See: P&H Chapter 2.4 2.6, 3.2, C.5 C.6 Hakim Weatherspoon CS 3410, Spring 2013 Computer Science Cornell University Big Picture: Building a Processor memory inst register file alu
Worst-Case Execution Time Analysis. LS 12, TU Dortmund
Worst-Case Execution Time Analysis Prof. Dr. Jian-Jia Chen LS 12, TU Dortmund 09/10, Jan., 2018 Prof. Dr. Jian-Jia Chen (LS 12, TU Dortmund) 1 / 43 Most Essential Assumptions for Real-Time Systems Upper
CSD Univ. of Crete Fall 2017 TUTORIAL ON EXTERNAL SORTING
TUTORIAL ON EXTERNAL SORTING 1 External Sorting External sorting has two main components: - Computation involved in sorting records in buffers in main memory - I/O necessary to move records between mass
EC 413 Computer Organization
EC 413 Computer Organization rithmetic Logic Unit (LU) and Register File Prof. Michel. Kinsy Computing: Computer Organization The DN of Modern Computing Computer CPU Memory System LU Register File Disks
CSE. 1. In following code. addi. r1, skip1 xor in r2. r3, skip2. counter r4, top. taken): PC1: PC2: PC3: TTTTTT TTTTTT
CSE 560 Practice Problem Set 4 Solution 1. In this question, you will examine several different schemes for branch prediction, using the following code sequence for a simple load store ISA with no branch
Designing Single-Cycle MIPS Processor
CSE 32: Introdction to Compter Architectre Designing Single-Cycle IPS Processor Presentation G Stdy:.-. Gojko Babić 2/9/28 Introdction We're now ready to look at an implementation of the system that incldes
GPU Acceleration of Cutoff Pair Potentials for Molecular Modeling Applications
GPU Acceleration of Cutoff Pair Potentials for Molecular Modeling Applications Christopher Rodrigues, David J. Hardy, John E. Stone, Klaus Schulten, Wen-Mei W. Hwu University of Illinois at Urbana-Champaign