The Era of Many-Module SoC : Revisiting the NoC Mapping Problem

Module

Module Module Module

Module

RR R R R

Module

Technion – Israel Institute of Technology

The Era of Many-Module SoC:Revisiting the NoC Mapping Problem

Isask’har (Zigi) Walter, Israel Cidon, Avinoam Kolodny, Daniel Sigalov

December, 2009

SoC Revolution

Bus-based system NoC-based system

SoC Evolution

R R R R R

Processor Evolution

Single Core

Dual Core

Cache CPU1Cache

Quad Core

CPU3Cache

CPU2Cache

CPU4Cache

How would such chips be like?

Most likely Power still important Highly parallel IP reuse

Ease of design and verification

The Era of Many-Module SoC

High Certainty

Totally unknown

Large number of modules

NoC Interconne

Applications

Special purpose cores replace general purpose processors Power considerations

Future SoCs - Observation#1

GeneralPurpose

Pre.Proc.

DSP CPU

Memory

Processing pipes are getting longer

Future SoCs - Observation#2

Large diversity All modules are

unique

Highly regular Classes of Replicated

cores standard modules (DSP,

HW accelerators, Cache banks, etc.)

Increased use of specialized cores Pipes are getting longer

Replication of processing elements

How is the design flow affected? This work – mapping of the NoC

The Era of Many-Module SoCObservation#

Observation#2

The Era of Many Module SoC Revisiting the Mapping Problem Cross-Entropy Optimization Evaluation

Outline

Given Traffic pattern(s)

a set (or sets) of pair-wise bandwidth requirements and timing constraints

Routing Topology

Goal Find efficient mapping of cores to tiles

NoC Mapping

PE1 PE2

PE7 PE8 PE9 PE4

PE2PE5

PE6PE7

PE4 PE1

PE3 PE6

PE7PE8

PE1PE2

PE8 PE9

An important design step Mapping affects power and performance!

A difficult problem! Often heuristic algorithms are used

Common optimization goals Minimize (dynamic) power Minimize power + maximize performance Minimize power subject to performance

constraints

Mapping Optimization

Typical modeling Power and latency proportional to

distance Cost function:

Modeling

( ) ( , )l i jl L i j N

Cost P BW b Dist i j

Calculating Mapping Cost

( ) ( , )l i jl L i j N

Cost P BW b Dist i j

2( ) 30 2 100 2 260Cost

PE1 PE2

PE4 PE5

1 2 6 4 3( ) 30 ( ) 100 ( )Cost Dist PE PE Dist PE PE

1 2 6 2 6 4 3 4 3( ) ( ) ( ) ( ) ( )Cost bw PE PE Dist PE PE bw PE PE bw PE PE

1( ) 30 2 100 3 360Cost

Mapping π1

Mapping π2

Motivation - Example #1

Optimal mapping (π1):

PE1 MEM2

PE2 PE3

1 1 1 1

( ) ( , ) 9i ji j N

Cost b Dist i j

Optimal mapping (π2):

Let the mapping algorithm assign the flows!

Motivation - Example #1 (cont.)

PE3 PE4

MEM2 PE1 MEM1

PE2 PE3

Cost(π1)=9Cost(π2)=7

Motivation - Example #1 (cont.)

PE3 PE4

Cost(π2)=7

The mapping algorithm should be aware of replicated modules!

Pair-wise point-to-point requirements For example, in a 4-module system:

Classic Performance Constraints

1 1PE4

PE4 PE3 PE2 PE1

PE1 PE2 PE3

Motivation - Example #2

Timing Requirement

PEs Stream ID

4 PE1PE2PE3PE4 Stream 1

1 PE2PE4 Stream 2

Stream 1

Stream 2

Example #2 – Pair-wise req.

No feasible mapping!

PE1 PE2 PE3

PE3 PE2. PE4.

PE1PE3

PE2.PE4

PE1PE3

1 PE3.

1 1 PE4

PE3 PE2 PE1

Req=2 Req=1

Req=1Req=1

PE1 PE2 PE3

Application-Level Requirements

Requirement PEs Stream ID

4 PE1PE2PE3PE4 Stream 1

1 PE2PE4 Stream 2

Stream 1

Stream 2

A feasible mapping does exist!

Req=1Req=2

Req=1Req=1

It’s better to work with the application level requirements

Find efficient mappings by extending the formulation of the mapping problem Adding degrees of freedom

Degree of freedom #1 Leverage existence of replicated modules

Degree of freedom #2 Replace p2p constraints with end-to-end,

application-level requirements

This Work

Modifying the Formulation (1)

Time Req. BW Flow

3 100 PE1DSP3

12 200 PE2DSP4

15 100 PE2SRAM1

5 100 PE3SRAM2

… … …

Time Req. BW Flow

3 100 PE1<ANY DSP>

12 200 PE2<ANY DSP>

15 100 PE2<ANY SRAM>

5 100 PE3<ANY SRAM>

… … …

Leverage existence of replicated modules Allow the mapping algorithm to allocate

flows to the best replicated module

∞ 3 1 3

∞ 4 7 7 3

∞ 3 ∞ 2 4 ∞

4 ∞ 7 2 3 ∞ 5

∞ ∞ 2 ∞ 1 5 6 ∞

7 7 3 ∞ ∞ 5 3 ∞ 1

∞ 3 ∞ 3 2 1 2 ∞ 3 1

Modifying the Formulation (2)

E2E Req.

Stream’s PEs Stream ID

23 PE1PE3PE9PE4PE10

12 PE5PE2PE3PE8PE7PE6PE10

15 PE5PE3PE9 3

20 PE7PE8PE2PE3 4

2 PE1PE2 5

… … …

In this paper, for synthetic task graphs Did so for a real application too

P2P timing req. E2E timing req.

Replace p2p constraints with end-to-end, application-level requirements

Outline

Modern optimization heuristic Good at combinatorial optimization problems

Akin to evolutionary algorithms Generation of new solutions is based on

sampling and estimation Inherently a global search method

Reduced risk of getting trapped in a local minimum

Cross Entropy Optimization

Given an initial parameter vector v=v0, sample a random population of K solutions x1,x2,…,xk from the distribution given by f(x;v).

Evaluate the costs S(xi),i=1,…,K. Using the ρK (0<ρ<1) elite (lowest cost) samples, obtain a new density function

f(x;v) by calculating a new vector v via Maximum Likelihood (ML) estimation. Repeat steps 1-3 with the new vector v unless maximum number of iterations is

reached or no improvement is obtained for a predefined number of iterations.

Cross Entropy Optimization

1. Generate 10 random mappings: π1, π2, …, π10

2. Find 3 lowest cost mappings: π2, π5, π7

3. Examine those 3 best mappings:A. For each tile, calculate the probability

core PEi is mapped to that tile

B. Update probabilities accordingly

For example:

Prob(TileAPE1)= Prob(TileAPE2)= Prob(TileAPE3)=Prob(TileAPE4)=0.25

Prob(TileBPE1)= Prob(TileBPE2)= Prob(TileBPE3)=Prob(TileBPE4)=0.25

Prob(TileCPE1)= Prob(TileCPE2)= Prob(TileCPE3)=Prob(TileCPE4)=0.25

Prob(TileDPE1)= Prob(TileDPE2)= Prob(TileDPE3)=Prob(TileDPE4)=0.25

CE Example

PE1 PE2

PE1PE2

PE1 PE2

PE3 PE1

PE2PE4

PE1 PE2

Prob(TileAPE1)=1

Updating Probabilities

PE1 PE2

PE4PE3PE3

PE4 PE1 PE2

PE4 PE3

Prob(TileBPE2)=2/3 Prob(TileBPE4)=1/3

Prob(TileDPE2)=1/3 Prob(TileDPE3)=1/3 Prob(TileDPE4)=1/3

Prob(TileCPE3)=2/3 Prob(TileCPE4)=1/3

π2 π5 π3

Following iteration uses these updates probabilities Gradually, probabilities converge to 0/1

Outline

Scenario 6x6 mesh NoC Synthetic, randomized SoC

Task graphs (and task-to-core mapping) Varying number of replicated modules Varying timing constraints (Real application in DATE10 paper)

Compare with best cost of classic mapping Averaging multiple runs

Evaluation

“Class”: a group of identical PEs Total number of replicated cores=

{Number of classes}*{class size}

Accounting for Replication

Replicated Modules [%]

One ClassTwo ClassesFour Classes

Reduct

Total number of Replicated Modules]%[ 30 40 50

SoCs with a pipeline data path and background P2P traffic Varying pipeline slack Different amounts of background

constraints

Application-Level Requirements

Tight Medium Loose Very Loose

Allowed Pipeline Slack

extra constraints extra constraints extra constraints extra constraints

Reduct

Pipeline Slack

We are going into the era of “Many module SoC”

Extend the mapping to account for Classes of replicated modules Application-level requirements

Meaningful power savings

But mapping is an example Routing? Task assignment? Link design?

Topology selection?

Conclusions and Future Work

Thank you!

Questions?

zigi@tx.technion.ac.il

The Era of Many-Module SoC

M odule

M odule M odule

M odule

QNoCResearch

GroupGroup

ResearchQNoC

The Era of Many-Module SoC : Revisiting the NoC Mapping Problem

Documents

Transcript of The Era of Many-Module SoC : Revisiting the NoC Mapping Problem

Adaptive NoC for reconfigurable SoC

Remote Infrastructure Management · • Deployment Architecture for NOC/SOC • Operation Scope & Deployment Methodology • Build Specification for Integration Transition & Implementation

[Tutorial] NoC the Next Generation of Multi-Processor SoC

Leveraging Application-Level Requirements in the Design of a NoC for a 4G SoC – a Case Study

Next Generation NOC/SOC Solutions

Next Generation NOC/SOC Solutions - mycom-osi.com OSI Solutions brochure - Q117.pdfNOC/SOC video walls with context-sensitive linking to other dashboards, reports and interactive analysis

SOLUTION DESCRIPTION · PDF fileSOLUTION DESCRIPTION ... Alarming, Automation, Extraction Subscriber Analytics Service Analytics ... Solution NOC/SOC, Engineering, CsOs Architects

Preventing & Mitigating Cyber Attacks · Digital Forensics Escalation Management SOC/NOC Monitoring Asset Management Continuous Monitoring/ Situational Awareness Security Operations

Adaptive NoC for recon gurable SoC - core.ac.uk · PDF fileAdaptive NoC for recon gurable SoC Istas Pratomo To cite this version: Istas Pratomo. Adaptive NoC for recon gurable SoC.

Securing your IT infrastructure with SOC-NOC collaboration TWP

NOC-centric security of reconﬁgurable SoC - Lab …gogniat/papers/diguetnoc2007.pdfNOC-centric security of reconﬁgurable SoC Jean-Philippe Diguet, Samuel Evain, Romain Vaslin,

SoC Design Lecture 13: NoC - Sharifce.sharif.edu/courses/88-89/1/ce757-1/resources/root/Slides/lec13.pdf · Outline SoC Interconnect NoC – Introduction NoC layers Typical NoC Router

Optimized Design of 2D Mesh NOC Router using Custom …aircconline.com/vlsics/V2N4/2411vlsics15.pdf · Optimized Design of 2D Mesh NOC Router using ... design of SOC using traditional

Resilient System-on-Chip Designs With NoC Fabricssandip.ece.ufl.edu/publications/tifs20.pdfSystem-on-Chip (SoC) designs by integration of pre-designed hardware blocks, — referred

Enhanced Security Mitigation via AutomationVerizon 2018 Data Breach Investigations Report ... •Reduce Time to Detection •Mitigate the Threat •NOC-SOC Integration Automated Security

Revised 1/14/2008 - PA - eMarketplace Web viewSupplier’s perimeter security and network operations alerting and NOC/SOC services interfaced to enterprise NOC/SOC services that are

TRANSACTION LEVEL MODELING FOR NOC AND SOC …eece.cu.edu.eg/~hfahmy/thesis/2015_04_tlm_noc_soc.pdf · TRANSACTION LEVEL MODELING FOR NOC AND SOC By Amr Ahmed Hany Mohamed A Thesis

(NOC) (SOC) I Assistant Director (NO C)

Reduce Threat Containment Time to Seconds Fortinet ...€¦ · Gap between SOC & NOC Corporate Network 1 Command & Control Server Internet Switch Compromised endpoint attempts to

SMART: A Single- Cycle Reconfigurable NoC for SoC Applications -Jyoti Wadhwani Chia-Hsin Owen Chen, Sunghyun Park, Tushar Krishna, Suvinay Subramaniam,