A 2bcgskew Fused by a RHSP Veerle Desmet Hans Vandierendonck Koen De Bosschere Ghent University Member HiPEAC.

Slides:



Advertisements
Similar presentations
Group - E Group Members Burdukov, Ilya Igorevich Lewczyk, Tomasz Sithambaram, Sasitharan Penmatcha, Bharath Steszewski, Andrew Joseph MAE412.
Advertisements

1 Inducements–Call Blocking. Aware of the Service?
Copyright © 2003 Pearson Education, Inc. Slide 1 Computer Systems Organization & Architecture Chapters 8-12 John D. Carpinelli.
Properties Use, share, or modify this drill on mathematic properties. There is too much material for a single class, so you’ll have to select for your.
ALGEBRA Number Walls
Doc.: IEEE Submission January 2010 Rick Roberts (Intel)Slide 1 Project: IEEE P Working Group for Wireless Personal Area Networks.
Geometric Sequences and Series
0 - 0.
Jeopardy ExponentsOrder of Operations Mult/Div Fractions Converting Fractions Q $100 Q $200 Q $300 Q $400 Q $500 Q $100 Q $200 Q $300 Q $400 Q $500 Final.
Year 6 mental test 5 second questions
9/09. A = 4 points B = 3 points C = 2 points D = 1 point F = 0 points Example: Hist B17a grade of A will receive 4 points for each unit of A. Hist B17a.
Decoding Circuits Made by Adham Barghouti
Richmond House, Liverpool (1) 26 th January 2004.
Reducing Order Enforcement Cost in Complex Query Plans Ravindra Guravannavar and S. Sudarshan (To appear in ICDE 2007)
EU Market Situation for Eggs and Poultry Management Committee 21 June 2012.
Ch 6, Principle of Biostatistics
Testing “Multiple Conditions” with Decision Table Technique
Idealized Piecewise Linear Branch Prediction Daniel A. Jiménez Department of Computer Science Rutgers University and Departament d'Arquitectura de Computadors.
André Seznec Caps Team IRISA/INRIA 1 Looking for limits in branch prediction with the GTL predictor André Seznec IRISA/INRIA/HIPEAC.
IP Multicast Information management 2 Groep T Leuven – Information department 2/14 Agenda •Why IP Multicast ? •Multicast fundamentals •Intradomain.
BEEF & VEAL MARKET SITUATION "Single CMO" Management Committee 18 April 2013.
SOLID GEOMETRY. (d) Projections of Right & Regular Solids like; (Prisms, Pyramids, Cylinder and Cone) SOLID GEOMETRY Following topics will be covered.
Based on Mike Feeley’s original slides; Modified by George Tsiknis Unit 11 Local Variables, Parameters and the Stack Relevant Information CPSC 213 Companion.
Factor P 16 8(8-5ab) 4(d² + 4) 3rs(2r – s) 15cd(1 + 2cd) 8(4a² + 3b²)
Welcome To The Four Umpire System. © 2007 Softball Canada All Rights Reserved Softball Canada Four Umpire System FP & SP 2 Tips for the Four Umpire System.
Welcome To The Three Umpire System. © 2007 Softball Canada All Rights Reserved Softball Canada Three Umpire System FP 2 Tips for the Three Umpire System.
©Evergreen Public Schools /11/2011 Arithmetic Sequences Explicit Rules Teacher Notes Notes : We will continue work students have done with arithmetic.
Addition 1’s to 20.
12 October, 2014 St Mungo's Academy 1 ADVANCED HIGHER MATHS REVISION AND FORMULAE UNIT 1.
2 x0 0 12/13/2014 Know Your Facts!. 2 x1 2 12/13/2014 Know Your Facts!
H-Pattern: A Hybrid Pattern Based Dynamic Branch Predictor with Performance Based Adaptation Samir Otiv Second Year Undergraduate Kaushik Garikipati Second.
©Brooks/Cole, 2001 Chapter 12 Derived Types-- Enumerated, Structure and Union.
2 x /18/2014 Know Your Facts!. 11 x /18/2014 Know Your Facts!
Intracellular Compartments and Transport
PSSA Preparation.
Essential Cell Biology
2 x /10/2015 Know Your Facts!. 8 x /10/2015 Know Your Facts!
Overheads for Computers as Components 2nd ed.
Introduction to Computer Systems
Chapter 3 โพรเซสเซอร์และการทำงาน The Processing Unit
CPSC 330 Fall 1999 HW #1 Assigned September 1, 1999 Due September 8, 1999 Submit in class Use a word processor (although you may hand-draw answers to Problems.
The Project Please read the project’s description first. Each router will have a unique ID, with your router’s ID of 0 Any two connected routers will have.
5 x4. 10 x2 9 x3 10 x9 10 x4 10 x8 9 x2 9 x4.
1 General Structural Equation (LISREL) Models Week 3 #2 A.Multiple Group Models with > 2 groups B.Relationship to ANOVA, ANCOVA models C.Introduction to.
0 x x2 0 0 x1 0 0 x3 0 1 x7 7 2 x0 0 9 x0 0.
MULTIPLICATION OF INTEGERS
Revisiting Local History to Improve the Fused Two-Level Branch Predictor Yasuo Ishii (The University of Tokyo, NEC) Keisuke Kuroyanagi (The University.
André Seznec Caps Team IRISA/INRIA 1 The O-GEHL branch predictor Optimized GEometric History Length André Seznec IRISA/INRIA/HIPEAC.
TAGE-SC-L Branch Predictors
Clustered Indexing for Conditional Branch Predictors Veerle Desmet Ghent University Belgium.
EECS 470 Branch Prediction Lecture 6 Coverage: Chapter 3.
VLSI Project Neural Networks based Branch Prediction Alexander ZlotnikMarcel Apfelbaum Supervised by: Michael Behar, Spring 2005.
Perceptrons Branch Prediction and its’ recent developments
Optimized Hybrid Scaled Neural Analog Predictor Daniel A. Jiménez Department of Computer Science The University of Texas at San Antonio.
1 Storage Free Confidence Estimator for the TAGE predictor André Seznec IRISA/INRIA.
1 A 64 Kbytes ITTAGE indirect branch predictor André Seznec INRIA/IRISA.
Analysis of Branch Predictors
André Seznec Caps Team IRISA/INRIA 1 Analysis of the O-GEHL branch predictor Optimized GEometric History Length André Seznec IRISA/INRIA/HIPEAC.
1 A New Case for the TAGE Predictor André Seznec INRIA/IRISA.
1 Revisiting the perceptron predictor André Seznec IRISA/ INRIA.
The Journal of Instruction Level Parallelism Championship Branch Prediction website: Dan Connors, Univ. of Colorado Tom Conte,
Not- Taken? Taken? The Frankenpredictor Gabriel H. Loh Georgia Tech College of Computing MICRO Dec 5, 2004.
André Seznec Caps Team IRISA/INRIA 1 A 256 Kbits L-TAGE branch predictor André Seznec IRISA/INRIA/HIPEAC.
André Seznec Caps Team IRISA/INRIA 1 Analysis of the O-GEHL branch predictor Optimized GEometric History Length André Seznec IRISA/INRIA/HIPEAC.
Fast Path-Based Neural Branch Prediction Daniel A. Jimenez Presented by: Ioana Burcea.
Samira Khan University of Virginia April 12, 2016
Looking for limits in branch prediction with the GTL predictor
TAGE-SC-L Again MTAGE-SC
5th JILP Workshop on Computer Architecture Competitions
The O-GEHL branch predictor
Presentation transcript:

A 2bcgskew Fused by a RHSP Veerle Desmet Hans Vandierendonck Koen De Bosschere Ghent University Member HiPEAC

1st JILP Championship Branch Prediction (CBP-1) --- December 5, A 2bcgskew Fused by a RHSP

1st JILP Championship Branch Prediction (CBP-1) --- December 5, A 2bcgskew Fused by a RHSP 2bcgskew RHSP

1st JILP Championship Branch Prediction (CBP-1) --- December 5, A 2bcgskew... 2bcgskew 4 component predictors: Bimodal Gshare (G0) Gshare (G1) Meta META selects between Majority vote BIM/G0/G1 BIM prediction [André Seznec, 2003]

1st JILP Championship Branch Prediction (CBP-1) --- December 5, A 2bcgskew... 2bcgskew Shared prediction tables: 2 N-1 bias G0/G1 2 N-2 bias BIM/META 2 N-2 hysteresis bits BIM/G0/G1/META Indexing: Bimodal pure addr G0 addr + 12 global hist G1 addr + 24 global hist META addr + 3 global hist

1st JILP Championship Branch Prediction (CBP-1) --- December 5, Fusion principle Hybrid selects component based on META- predictor follows outcome of one component Fusion learns patterns from component outcomes improved accuracy if component outcomes differ [Gabriel Loh, 2002] Component predictors fusion Component predictors meta

1st JILP Championship Branch Prediction (CBP-1) --- December 5, Fused by... Extracts multiple bits from 2bcgskew 9 outcome bits: 2bcgskew BIM bias + hysteresis G0 bias + hysteresis G1 bias + hysteresis META bias + hysteresis

1st JILP Championship Branch Prediction (CBP-1) --- December 5, a RHSP predictor RHSP [André Seznec, 2003]

1st JILP Championship Branch Prediction (CBP-1) --- December 5, Information#bits Global History33 Local History6 Pseudo-tag8 Outcomes9 Bias1 Perceptron Predictor Local history table 512 entries indexed by branch address Pseudo-tag bits higher order address bits not used to select weights [André Seznec, 2003]

1st JILP Championship Branch Prediction (CBP-1) --- December 5, InformationRedundancy#bits Global Historypure33 1 Local Historypure Pseudo-tagpure Outcomespure9 Biaspure1 Redundant History Perceptron Predictor 116 n h5h4h3h2h1h0 n-th order redundancy of h: h5h4h3h2h1h0 x4x3x2x1x0

1st JILP Championship Branch Prediction (CBP-1) --- December 5, Redundant History Skewed Perceptron Predictor InformationRedundancy#bitsBank 1Bank 2Bank 3Bank 4 Global Historypure Local Historypure Pseudo-tagpure Outcomespure99000 Biaspure

1st JILP Championship Branch Prediction (CBP-1) --- December 5, Redundant History Skewed Perceptron Predictor Bank 1Bank 2Bank 3Bank 4 4 banks 32 entries = 5 bits index indexed by global history and rotation hashing function of branch address 8-bit weights MAC: (A+B, A-B) coding together 116 weights/entry a5a4 a0a1a2a3 r4r3r2r1r0 a8a7a6

1st JILP Championship Branch Prediction (CBP-1) --- December 5, Budget 2bcgskew 32K bits Fusion RHSP predictor 29K bits Local history 3K bits

Performance???

The end

Feedback: Privacy Policy Feedback
Do Not Sell My Personal Information
About project: SlidePlayer Terms of Service

© 2024 SlidePlayer.com Inc. All rights reserved.