Looking for limits in branch prediction with the GTL predictor

Slides:



Advertisements
Similar presentations
Bimode Cascading: Adaptive Rehashing for ITTAGE Indirect Branch Predictor Y.Ishii, K.Kuroyanagi, T.Sawada, M.Inaba, and K.Hiraki.
Advertisements

Dead Block Replacement and Bypass with a Sampling Predictor Daniel A. Jiménez Department of Computer Science The University of Texas at San Antonio.
A 2bcgskew Fused by a RHSP Veerle Desmet Hans Vandierendonck Koen De Bosschere Ghent University Member HiPEAC.
André Seznec Caps Team IRISA/INRIA 1 Looking for limits in branch prediction with the GTL predictor André Seznec IRISA/INRIA/HIPEAC.
H-Pattern: A Hybrid Pattern Based Dynamic Branch Predictor with Performance Based Adaptation Samir Otiv Second Year Undergraduate Kaushik Garikipati Second.
Adding the Jump Instruction
Hardware-based Devirtualization (VPC Prediction) Hyesoon Kim, Jose A. Joao, Onur Mutlu ++, Chang Joo Lee, Yale N. Patt, Robert Cohn* ++ *
1 A Hybrid Adaptive Feedback Based Prefetcher Santhosh Verma, David Koppelman and Lu Peng Louisiana State University.
Exploring Correlation for Indirect Branch Prediction 1 Nikunj Bhansali, Chintan Panirwala, Huiyang Zhou Department of Electrical and Computer Engineering.
André Seznec Caps Team IRISA/INRIA 1 The O-GEHL branch predictor Optimized GEometric History Length André Seznec IRISA/INRIA/HIPEAC.
Dynamic Branch Prediction
Exploiting Spatial Locality in Data Caches using Spatial Footprints Sanjeev Kumar, Princeton University Christopher Wilkerson, MRL, Intel.
André Seznec Caps Team IRISA/INRIA Design tradeoffs for the Alpha EV8 Conditional Branch Predictor André Seznec, IRISA/INRIA Stephen Felix, Intel Venkata.
Yue Hu David M. Koppelman Lu Peng A Penalty-Sensitive Branch Predictor Department of Electrical and Computer Engineering Louisiana State University.
A PPM-like, tag-based predictor Pierre Michaud. 2 Main characteristics global history based 5 tables –one 4k-entry bimodal (indexed with PC) –four 1k-entry.
TAGE-SC-L Branch Predictors
CPE 731 Advanced Computer Architecture ILP: Part II – Branch Prediction Dr. Gheith Abandah Adapted from the slides of Prof. David Patterson, University.
EECC722 - Shaaban #1 Lec # 10 Fall Conventional & Block-based Trace Caches In high performance superscalar processors the instruction fetch.
Branch Target Buffers BPB: Tag + Prediction
Computer Architecture Instruction Level Parallelism Dr. Esam Al-Qaralleh.
1 Lecture 8: Branch Prediction, Dynamic ILP Topics: static speculation and branch prediction (Sections )
Dynamic Branch Prediction
Prophet/Critic Hybrid Branch Prediction Falcon, Stark, Ramirez, Lai, Valero Presenter: Christian Wanamaker.
CIS 429/529 Winter 2007 Branch Prediction.1 Branch Prediction, Multiple Issue.
Perceptrons Branch Prediction and its’ recent developments
EECC722 - Shaaban #1 Lec # 9 Fall Conventional & Block-based Trace Caches In high performance superscalar processors the instruction fetch.
Evaluation of Dynamic Branch Prediction Schemes in a MIPS Pipeline Debajit Bhattacharya Ali JavadiAbhari ELE 475 Final Project 9 th May, 2012.
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.
Evaluation of the Gini-index for Studying Branch Prediction Features Veerle Desmet Lieven Eeckhout Koen De Bosschere.
1 A 64 Kbytes ITTAGE indirect branch predictor André Seznec INRIA/IRISA.
Analysis of Branch Predictors
1 Two research studies related to branch prediction and instruction sequencing André Seznec INRIA/IRISA.
André Seznec Caps Team IRISA/INRIA 1 Analysis of the O-GEHL branch predictor Optimized GEometric History Length André Seznec IRISA/INRIA/HIPEAC.
Microprocessor Microarchitecture Instruction Fetch Lynn Choi Dept. Of Computer and Electronics Engineering.
1 A New Case for the TAGE Predictor André Seznec INRIA/IRISA.
1 Revisiting the perceptron predictor André Seznec IRISA/ INRIA.
Not- Taken? Taken? The Frankenpredictor Gabriel H. Loh Georgia Tech College of Computing MICRO Dec 5, 2004.
Predicting Conditional Branches With Fusion-Based Hybrid Predictors Gabriel H. Loh Yale University Dept. of Computer Science Dana S. Henry Yale University.
T-BAG: Bootstrap Aggregating the TAGE Predictor Ibrahim Burak Karsli, Resit Sendag University of Rhode Island.
André Seznec Caps Team IRISA/INRIA 1 A 256 Kbits L-TAGE branch predictor André Seznec IRISA/INRIA/HIPEAC.
CS 6290 Branch Prediction. Control Dependencies Branches are very frequent –Approx. 20% of all instructions Can not wait until we know where it goes –Long.
Idealized Piecewise Linear Branch Prediction Daniel A. Jiménez Department of Computer Science Rutgers University.
1 The Inner Most Loop Iteration counter a new dimension in branch history André Seznec, Joshua San Miguel, Jorge Albericio.
Prophet/Critic Hybrid Branch Prediction B B B
Branch Prediction Perspectives Using Machine Learning Veerle Desmet Ghent University.
André Seznec Caps Team IRISA/INRIA 1 Analysis of the O-GEHL branch predictor Optimized GEometric History Length André Seznec IRISA/INRIA/HIPEAC.
FAT predictor Sabareesh Ganapathy, Prasanna Venkatesh Srinivasan, Maribel Monica.
Fast Path-Based Neural Branch Prediction Daniel A. Jimenez Presented by: Ioana Burcea.
Multiperspective Perceptron Predictor Daniel A. Jiménez Department of Computer Science & Engineering Texas A&M University.
Dynamic Branch Prediction
Multilayer Perceptron based Branch Predictor
CS203 – Advanced Computer Architecture
Distributive Property of Multiplication 306 x 2
Multiperspective Perceptron Predictor with TAGE
COSC3330 Computer Architecture Lecture 15. Branch Prediction
Dynamically Sizing the TAGE Branch Predictor
CS 704 Advanced Computer Architecture
Partial Products Algorithm for Multiplication
FA-TAGE Frequency Aware TAgged GEometric History Length Branch Predictor Boyu Zhang, Christopher Bodden, Dillon Skeehan ECE/CS 752 Advanced Computer Architecture.
Exploring Value Prediction with the EVES predictor
Dynamic Branch Prediction
Lecture 10: Branch Prediction and Instruction Delivery
TAGE-SC-L Again MTAGE-SC
Serene Banerjee, Lizy K. John, Brian L. Evans
5th JILP Workshop on Computer Architecture Competitions
Adapted from the slides of Prof
Dynamic Hardware Prediction
The O-GEHL branch predictor
Gang Luo, Hongfei Guo {gangluo,
Presentation transcript:

Looking for limits in branch prediction with the GTL predictor André Seznec IRISA/INRIA/HIPEAC

Motivations Geometric history length predictors introduced in 2004-2006 OGEHL, CBP-1, dec. 2004 TAGE, JILP ’06, feb. 2006 Storage effective Exploits very long global histories Were defined with possible implementation in mind What are the limits of accuracy that can be captured with these schemes ? How do they compare with unconstrained prediction schemes ?

Geometric history length predictors: global history +multiple lengths ? L(0) T3 L(1) L(2) T4 L(3) L(4)

GEometric History Length predictor The set of history lengths forms a geometric series Capture correlation on very long histories {0, 2, 4, 8, 16, 32, 64, 128} most of the storage for short history !! What is important: L(i)-L(i-1) is drastically increasing

Combining multiple predictions Neural inspired predictors Use a (multiply)-add tree Partial matching Use tagged tables and the longest matching history O-GEHL, CBP-1 TAGE, JILP’ 06

Final computation through a sum 256Kbits: 12 components 3.670 misp/KI CBP-1 (2004): O-GEHL Final computation through a sum ∑ L(4) L(3) L(2) L(1) TO T1 T2 T3 T4 L(0) Prediction=Sign 256Kbits: 12 components 3.670 misp/KI

JILP ‘06: TAGE longest matching history =? =? =? 1 1 1 1 1 1 1 1 1 256Kbits: 3.358 misp/KI

What is global history conditional branch history: path confusion on short histories  path history: Direct hashing leads to path confusion  Represent all branches in branch history Use path AND direction history

Using a kernel history and a user history Traces mix user and kernel activities: Kernel activity after exception Global history pollution Solution: use two separate global histories User history is updated only in user mode Kernel history is updated in both modes

Accuracy limits for TAGE Varying the predictor size, the number of components, the tag width, the history length. Allowing multiple allocations The best accuracy on distributed traces: 3.054 misp/KI History length around 1,000 15-20 components No need for tags wider than 16 bits

Accuracy limits for GEHL Varying the predictor size, the number of components, the history length, counter width (slightly) improving the update policy and fitting in the two hours simulation rule  on the distributed traces: 2.842 misp/KI 97 components 8 bits counter 2,000 bits global history

GEHL vs TAGE Realistic implementation parameters (storage budget, number of components) TAGE is more accurate than (O-)GEHL Unlimited budget, huge number of components GEHL is more accurate than TAGE

Will it be sufficient to win The Championship ? GEHL history length: 2,000 97 components 2.842 misp/KI

A step further: hybrid GEHL-TAGE On a few benchmarks, TAGE is more accurate than GEHL, Let us try an hybrid GEHL-TAGE predictor

Branch/path history + PC Hybrid GEHL-TAGE Branch/path history + PC GEHL TAGE Meta = egskew mux Inherit from: Agree/bimode, YAGS, 2bcgskew,

GEHL+TAGE GEHL provides the main prediction: also used as the base predictor for TAGE (YAGS inspired) TAGE records when GEHL fails: {prediction, address, history} (agree/bimode, YAGS inspired) Meta selects between GEHL and TAGE (2bcgskew inspired)

Let us have fun !! GEHL history length: 400 TAGE history length: 100,000 2.774 misp/KI

Might still be unsufficient  GEHL history length: 400 TAGE history length: 100,000 2.774 misp/KI

Adding a loop predictor The loop predictor captures the number of iterations of a loop When successively encounters 8 times the same number of iterations, the loop predictor provides the prediction. Advantage: Very reliable

Branch/path history + PC GTL predictor Branch/path history + PC GEHL TAGE Meta = egskew mux Loop predictor confidence + static prediction on first occurrence

Hope this will be sufficient to win the Championship !! GTL GEHL, 97 comp., 400 hist. + TAGE, 19 comp., 100,000 hist + loop predictor 2.717 misp/KI

Geometric History Length predictors and limits on branch prediction Unlimited budget, huge number of components GEHL is more accurate than TAGE Very old correlation can be captured: On two benchmarks, using 10,000 history is really helping Does not seem to be a lot of potential extra benefit from local history We did not find any interesting extra scheme apart loop prediction Loop prediction, very marginal apart gzip

The End 