Design as You See FIT: System-Level Soft Error Analysis of Sequential Circuits

Similar documents
1. Introduction. 2. Fault modeling in logic

Project UPSET: Understanding and Protecting Against Single Event Transients

This work is supported in part by grants from GSRC and NSF (Career No )

Multiple Transient Faults in Combinational and Sequential Circuits: A Systematic Approach

An Efficient Static Algorithm for Computing the Soft Error Rates of Combinational Circuits

Symbolic Simulation of the Propagation and Filtering of Transient Faulty Pulses

An Accurate Single Event Effect Digital Design Flow for Reliable System Level Design

Design of Soft Error Tolerant Memory and Logic Circuits

Auto-tuning Fault Tolerance Technique for DSP-Based Circuits in Transportation Systems

The Effect of Threshold Voltages on the Soft Error Rate. - V Degalahal, N Rajaram, N Vijaykrishnan, Y Xie, MJ Irwin

On the Role of Timing Masking in Reliable Logic Circuit Design

EECS150 - Digital Design Lecture 28 Course Wrap Up. Recap 1

Soft Error Rate Analysis for Combinational Logic Using An Accurate Electrical Masking Model

Course Outcome of M.Tech (VLSI Design)

Cost-Effective Radiation Hardening Technique for Combinational Logic

Device and Architecture Concurrent Optimization for FPGA Transient Soft Error Rate

Design of Robust CMOS Circuits for Soft Error Tolerance

The challenges of low power design Karen Yorav

SOFT errors are radiation-induced transient errors caused by

On-chip Networks in Multi-core era

Low Power Radiation Tolerant CMOS Design using Commercial Fabrication Processes

Single Event Transient Effects on Microsemi ProASIC Flash-based FPGAs: analysis and possible solutions

A Novel Low-Power Scan Design Technique Using Supply Gating

Partial Error Masking to Reduce Soft Error Failure Rate in Logic Circuits

Low Power System-On-Chip-Design Chapter 12: Physical Libraries

Separate Dual-Transistor Registers - A Circuit Solution for On-line Testing of Transient Error in UDSM-IC

Modeling the Effect of Technology Trends on the Soft Error Rate of Combinational Logic

THE increasing threat of soft errors in nanometer technologies

CHAPTER III THE FPGA IMPLEMENTATION OF PULSE WIDTH MODULATION

Method for Qcrit Measurement in Bulk CMOS Using a Switched Capacitor Circuit

Low Power Design of Successive Approximation Registers

Modeling the Impact of Device and Pipeline Scaling on the Soft Error Rate of Processor Elements

Overview ECE 553: TESTING AND TESTABLE DESIGN OF DIGITAL SYSTES. Motivation. Modeling Levels. Hierarchical Model: A Full-Adder 9/6/2002

Soft Error Susceptibility in SRAM-Based FPGAs. With the increasing emphasis on minimizing mass and volume along with

CHAPTER 7 A BICS DESIGN TO DETECT SOFT ERROR IN CMOS SRAM

EC 1354-Principles of VLSI Design

Modeling the Effect of Technology Trends on Soft Error Rate of Combinational Logic

Early SEU Fault Injection in Digital, Analog and Mixed Signal Circuits: a Global Flow

EE 434 ASIC & Digital Systems

Logic SER Reduction through Flipflop Redesign

A Novel Radiation Tolerant SRAM Design Based on Synergetic Functional Component Separation for Nanoscale CMOS.

An Overview of the NASA Goddard Methodology for FPGA Radiation Testing and Soft Error Rate (SER) Prediction

STM RH-ASIC capability

INF3430 Clock and Synchronization

DIGITAL INTEGRATED CIRCUITS A DESIGN PERSPECTIVE 2 N D E D I T I O N

DIRECT IONIZATION INDUCED TRANSIENT FAULT ANALYSIS FOR COMBINATIONAL LOGIC AND SEQUENTIAL CAPTURE IN DIGITAL INTEGRATED

Debugging a Boundary-Scan I 2 C Script Test with the BusPro - I and I2C Exerciser Software: A Case Study

EECS 427 Lecture 21: Design for Test (DFT) Reminders

VLSI Design Verification and Test Delay Faults II CMPE 646

Low-Power Soft Error Hardened Latch

Evaluation of the Masked Logic Style MDPL on a Prototype Chip

Testing of Complex Digital Chips. Juri Schmidt Advanced Seminar

CS250 VLSI Systems Design. Lecture 3: Physical Realities: Beneath the Digital Abstraction, Part 1: Timing

A New Asymmetric SRAM Cell to Reduce Soft Errors and Leakage Power in FPGA

UNIT-III POWER ESTIMATION AND ANALYSIS

Announcements. Advanced Digital Integrated Circuits. Project proposals due today. Homework 1. Lecture 8: Gate delays,

28nm and below: New Frontiers and Innovations in Design for Manufacturing. Vito Dai, Ph.D. Sr. Member of Technical Staff, DFM

Automated FSM Error Correction for Single Event Upsets

PROCESS-VOLTAGE-TEMPERATURE (PVT) VARIATIONS AND STATIC TIMING ANALYSIS

EE241 - Spring 2013 Advanced Digital Integrated Circuits. Announcements. Lecture 13: Timing revisited

Computer Aided Design of Electronics

VLSI Physical Design Prof. Indranil Sengupta Department of Computer Science and Engineering Indian Institute of Technology, Kharagpur

ELEC Digital Logic Circuits Fall 2015 Delay and Power

Research Article Single-Event-Upset Sensitivity Analysis on Low-Swing Drivers

! Review: MOS IV Curves and Switch Model. ! MOS Device Layout. ! Inverter Layout. ! Gate Layout and Stick Diagrams. ! Design Rules. !

ASICs Concept to Product

A Review of Clock Gating Techniques in Low Power Applications

In the previous chapters, efficient and new methods and. algorithms have been presented in analog fault diagnosis. Also a

Dual-K K Versus Dual-T T Technique for Gate Leakage Reduction : A Comparative Perspective

Low Power VLSI Circuit Synthesis: Introduction and Course Outline

A Highly-Efficient Technique for Reducing Soft Errors in Static CMOS Circuits

Extending Modular Redundancy to NTV: Costs and Limits of Resiliency at Reduced Supply Voltage

Cascaded Propagation and Reduction Techniques for Fault Binary Decision Diagram in Single-event Transient Analysis

Optimizing Design of Fault-tolerant Computing Systems

Jack Keil Wolf Lecture. ESE 570: Digital Integrated Circuits and VLSI Fundamentals. Lecture Outline. MOSFET N-Type, P-Type.

Energy Efficiency of Power-Gating in Low-Power Clocked Storage Elements

! Review: MOS IV Curves and Switch Model. ! MOS Device Layout. ! Inverter Layout. ! Gate Layout and Stick Diagrams. ! Design Rules. !

ESE 570: Digital Integrated Circuits and VLSI Fundamentals

SOFT ERROR TOLERANT HIGHLY RELIABLE MULTIPORT MEMORY CELL DESIGN

WHEN high-energy neutrons (present in terrestrial cosmic

University of Minnesota, Minneapolis, MN 2. Intel Corporation, Hillsboro, OR 3. Los Alamos National Laboratory, Los Alamos, NM

ESE 570: Digital Integrated Circuits and VLSI Fundamentals

Highly Reliable Arithmetic Multipliers for Future Technologies

Policy-Based RTL Design

A BICS Design to Detect Soft Error in CMOS SRAM

Energy Efficient Memory Design using Low Voltage Complementary Metal Oxide Semiconductor on 28nm FPGA

Low-Power Soft Error Hardened Latch

Embedded Error Compensation for Energy Efficient DSP Systems

Tunable transient filters for soft error rate reduction in combinational circuits

(VE2: Verilog HDL) Software Development & Education Center

A Power-Efficient Design Approach to Radiation Hardened Digital Circuitry using Dynamically Selectable Triple Modulo Redundancy

Lecture 1. Tinoosh Mohsenin

CS302 - Digital Logic Design Glossary By

Synthesis of Non-Intrusive Concurrent Error Detection Using an Even Error Detecting Function

A design of 16-bit adiabatic Microprocessor core

Interconnect-Power Dissipation in a Microprocessor

Inspector Data Sheet. EM-FI Transient Probe. High speed pulsed EM fault injection probe for localized glitches. Riscure EM-FI Transient Probe 1/8

Advanced Digital Design

A Current-based Method for Short Circuit Power Calculation under Noisy Input Waveforms *

Selective Hardening in Early Design Steps

Transcription:

Design as You See FIT: System-Level Soft Error Analysis of Sequential Circuits Dan Holcomb Wenchao Li Sanjit A. Seshia Department of EECS University of California, Berkeley Design Automation and Test in Europe, 2009 Holcomb, Li, Seshia Design as You See FIT 1

Soft errors in VLSI circuits Spurious radiation-induced flip of one or more stored bits Does not permanently damage devices Measured in units of FIT; 1 FIT is 1 failure in 10 9 hrs Strikes to logic or directly to memory Holcomb, Li, Seshia Design as You See FIT 2

Soft errors in VLSI circuits Spurious radiation-induced flip of one or more stored bits Does not permanently damage devices Measured in units of FIT; 1 FIT is 1 failure in 10 9 hrs Strikes to logic or directly to memory Logic FIT may increase 9 orders of magnitude from 1992-2011 [Shivakumar 02] Logic FIT approaching memory FIT around 100nm [Shivakumar 02, Baumann 05] Circuit-level hardening techniques exist, but have costs Holcomb, Li, Seshia Design as You See FIT 2

Our contribution the verification guided error resilience methodology Use formal specifications to capture system-level correctness Holcomb, Li, Seshia Design as You See FIT 3

Our contribution the verification guided error resilience methodology Use formal specifications to capture system-level correctness Use verification to analyze system-level impact of circuit-level upsets Holcomb, Li, Seshia Design as You See FIT 3

Our contribution the verification guided error resilience methodology Use formal specifications to capture system-level correctness Use verification to analyze system-level impact of circuit-level upsets Guide efficient circuit hardening techniques Holcomb, Li, Seshia Design as You See FIT 3

Our contribution the verification guided error resilience methodology Use formal specifications to capture system-level correctness Use verification to analyze system-level impact of circuit-level upsets Guide efficient circuit hardening techniques Particularly suited for communication protocols and on-chip networks Holcomb, Li, Seshia Design as You See FIT 3

Outline Introduction Single Event Upset Circuit-level masking System-level masking Introduction Single Event Upset Circuit-level masking System-level masking BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Analysis Efficient Hardening Holcomb, Li, Seshia Design as You See FIT 4

Single Event Upset Circuit-level masking System-level masking Basic mechanisms of single event upset (SEU) in logic Strike near sensitive diffusion Sensitive diffusion is a function of gate input If sufficient charge, glitch results at gate output Glitch propagates downstream toward sequential element(s) Holcomb, Li, Seshia Design as You See FIT 5

Circuit-level masking not all glitches are equally likely to flip bits Single Event Upset Circuit-level masking System-level masking Logical Masking Is there a sensitized path from strike to latch(es)? Timing Masking Does the glitch arrive at latch(es) while open? Electrical Masking Is the strike magnitude sufficient to cause upset? SEU can lead to single (SBU) or multi-bit upset (MBU) Holcomb, Li, Seshia Design as You See FIT 6

Circuit-level masking Introduction Single Event Upset Circuit-level masking System-level masking Related work Holcomb, Li, Seshia Design as You See FIT 7

Circuit-level masking Introduction Single Event Upset Circuit-level masking System-level masking Related work Static analysis of circuit structure [Miskov-Zivanov 06, B. Zhang 06] Holcomb, Li, Seshia Design as You See FIT 7

Circuit-level masking Introduction Single Event Upset Circuit-level masking System-level masking Related work Static analysis of circuit structure [Miskov-Zivanov 06, B. Zhang 06] Electrical masking using input vectors [M. Zhang 06, Rao 07] Path-based methods Use only a subset of paths Lose accuracy in timing masking Holcomb, Li, Seshia Design as You See FIT 7

Circuit-level masking Introduction Single Event Upset Circuit-level masking System-level masking Related work Static analysis of circuit structure [Miskov-Zivanov 06, B. Zhang 06] Electrical masking using input vectors [M. Zhang 06, Rao 07] Path-based methods Use only a subset of paths Lose accuracy in timing masking Timing masking, with no electrical masking [Krishnaswamy 08] Holcomb, Li, Seshia Design as You See FIT 7

Circuit-level masking Introduction Single Event Upset Circuit-level masking System-level masking Related work Static analysis of circuit structure [Miskov-Zivanov 06, B. Zhang 06] Electrical masking using input vectors [M. Zhang 06, Rao 07] Path-based methods Use only a subset of paths Lose accuracy in timing masking Timing masking, with no electrical masking [Krishnaswamy 08] No analysis of MBU Holcomb, Li, Seshia Design as You See FIT 7

Circuit-level masking Introduction Single Event Upset Circuit-level masking System-level masking Related work Static analysis of circuit structure [Miskov-Zivanov 06, B. Zhang 06] Electrical masking using input vectors [M. Zhang 06, Rao 07] Path-based methods Use only a subset of paths Lose accuracy in timing masking Timing masking, with no electrical masking [Krishnaswamy 08] No analysis of MBU Our approach An efficient method for estimating electrical/timing masking Analysis of circuits up to 20k gates Able to handle multiple sensitized paths Analysis of both SBU and MBU Holcomb, Li, Seshia Design as You See FIT 7

Single Event Upset Circuit-level masking System-level masking System level masking not all bit flips are equally likely to cause system failure Related work Holcomb, Li, Seshia Design as You See FIT 8

Single Event Upset Circuit-level masking System-level masking System level masking not all bit flips are equally likely to cause system failure Related work Architectural Vulnerability Factor [Mukherjee 03] Find probability of a bit flip leading to incorrect future execution Requires detailed architecture model Holcomb, Li, Seshia Design as You See FIT 8

Single Event Upset Circuit-level masking System-level masking System level masking not all bit flips are equally likely to cause system failure Related work Architectural Vulnerability Factor [Mukherjee 03] Find probability of a bit flip leading to incorrect future execution Requires detailed architecture model Output equivalence Find probability of a bit flip leading to incorrect outputs Model state using Markov Chain theory [Miskov-Zivanov 08] Holcomb, Li, Seshia Design as You See FIT 8

Single Event Upset Circuit-level masking System-level masking System level masking not all bit flips are equally likely to cause system failure Related work Architectural Vulnerability Factor [Mukherjee 03] Find probability of a bit flip leading to incorrect future execution Requires detailed architecture model Output equivalence Find probability of a bit flip leading to incorrect outputs Model state using Markov Chain theory [Miskov-Zivanov 08] Verification Guided [Seshia 07] Find probability possibility of a bit flip leading to bad behavior Bad behavior formalized using specifications Model checking identifies non-critical latches High confidence but binary Can apply to individual functional blocks Holcomb, Li, Seshia Design as You See FIT 8

Single Event Upset Circuit-level masking System-level masking System level masking not all bit flips are equally likely to cause system failure Related work Architectural Vulnerability Factor [Mukherjee 03] Find probability of a bit flip leading to incorrect future execution Requires detailed architecture model Output equivalence Find probability of a bit flip leading to incorrect outputs Model state using Markov Chain theory [Miskov-Zivanov 08] Verification Guided [Seshia 07] Find probability possibility of a bit flip leading to bad behavior Bad behavior formalized using specifications Model checking identifies non-critical latches High confidence but binary Can apply to individual functional blocks Our approach Verification guided, but produce a refined ranking Holcomb, Li, Seshia Design as You See FIT 8

Outline Introduction BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Introduction Single Event Upset Circuit-level masking System-level masking BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Analysis Efficient Hardening Holcomb, Li, Seshia Design as You See FIT 9

VGER toolkit Introduction BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Holcomb, Li, Seshia Design as You See FIT 10

VGER toolkit Introduction BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Holcomb, Li, Seshia Design as You See FIT 10

VGER toolkit Introduction BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Holcomb, Li, Seshia Design as You See FIT 10

VGER toolkit Introduction BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Holcomb, Li, Seshia Design as You See FIT 10

VGER toolkit Introduction BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Holcomb, Li, Seshia Design as You See FIT 10

VGER toolkit Introduction BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Holcomb, Li, Seshia Design as You See FIT 10

VGER toolkit Introduction BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Holcomb, Li, Seshia Design as You See FIT 10

BFIT circuit level analysis tool BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Open source C++ simulation tool for combinational logic circuits Based on Nangate 45nm open cell library Inputs Comb. circuit and sampled states Outputs FIT of all events: FIT g E Struck gate g Set of upset sequential elements E some E are SBU, others are MBU Holcomb, Li, Seshia Design as You See FIT 11

What determines the FIT of an event? BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Every possible strike is represented by a collected charge and time (q, t) FIT g E R g (q, t) N g E (q, t) dt dq R g(q, t) R N g E (q, t) {0, 1} probability of observing strike q, t at gate g conditional probability of upset in set E of latches, given a strike q, t at gate g Encompasses logical, electrical, timing masking Holcomb, Li, Seshia Design as You See FIT 12

BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors BFIT approach to Masking N(q, t) of single path can be characterized using path delay and gate input Inputs determine drive strength Input determines shape of N(q, t) 00 01 10 11 5.0E-6 7.5E-6 16.1E-6 2.9E-6 Table: FIT vs gate input Holcomb, Li, Seshia Design as You See FIT 13

BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors BFIT approach to Masking N(q, t) of single path can be characterized using path delay and gate input Inputs determine drive strength Input determines shape of N(q, t) Path delay is time-shift to N(q, t) 00 01 10 11 5.0E-6 7.5E-6 16.1E-6 2.9E-6 Table: FIT vs gate input 15 5 3 1 6.2E-6 6.4E-6 6.4E-6 6.7E-6 Table: FIT vs path length Holcomb, Li, Seshia Design as You See FIT 13

Demonstration of BFIT algorithm BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Holcomb, Li, Seshia Design as You See FIT 14

Demonstration of BFIT algorithm BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors 1. Forward propagate input vector in levelized DAG Holcomb, Li, Seshia Design as You See FIT 14

Demonstration of BFIT algorithm BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors 1. Forward propagate input vector in levelized DAG 2. Dynamic programming back trace in reverse levelized order If an input can flip current gate, back propagate delays Holcomb, Li, Seshia Design as You See FIT 14

Demonstration of BFIT algorithm BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors 1. Forward propagate input vector in levelized DAG 2. Dynamic programming back trace in reverse levelized order If an input can flip current gate, back propagate delays 3. For each gate g, find all possible N g E (q, t) using: List of path delays and terminating latches Gate input state and load capacitance Cell precharacterization Holcomb, Li, Seshia Design as You See FIT 14

Demonstration of BFIT algorithm BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors 1. Forward propagate input vector in levelized DAG 2. Dynamic programming back trace in reverse levelized order If an input can flip current gate, back propagate delays 3. For each gate g, find all possible N g E (q, t) using: List of path delays and terminating latches Gate input state and load capacitance Cell precharacterization Holcomb, Li, Seshia Design as You See FIT 14

Demonstration of BFIT algorithm BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors 1. Forward propagate input vector in levelized DAG 2. Dynamic programming back trace in reverse levelized order If an input can flip current gate, back propagate delays 3. For each gate g, find all possible N g E (q, t) using: List of path delays and terminating latches Gate input state and load capacitance Cell precharacterization Holcomb, Li, Seshia Design as You See FIT 14

BFIT results Introduction BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Circuit INPUTS LATCHES GATES RUNTIME (s/1k vectors) FIT s5378 214 179 3232 35 3.27e-3 s9234 247 228 7230 68 1.06e-2 s13207 700 669 10277 136 1.77e-2 s15850 611 597 12712 207 2.24e-2 s38417 1664 1636 28223 451 5.62e-2 s38584 1464 1452 28854 1311 4.85e-2 s35932 1763 1728 23012 204 3.99e-2 Holcomb, Li, Seshia Design as You See FIT 15

BFIT results multiple bit upsets Introduction BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Do MBU occur? Holcomb, Li, Seshia Design as You See FIT 16

BFIT results multiple bit upsets Introduction BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Do MBU occur? BFIT HSPICE SBU 8.6e-7 7.4e-7 MBU 9.0e-7 12.6e-7 total FIT 1.7e-6 2.0e-6 Holcomb, Li, Seshia Design as You See FIT 16

BFIT results multiple bit upsets Introduction BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Do MBU occur? BFIT HSPICE SBU 8.6e-7 7.4e-7 MBU 9.0e-7 12.6e-7 total FIT 1.7e-6 2.0e-6 Is MBU a concern? Holcomb, Li, Seshia Design as You See FIT 16

BFIT results multiple bit upsets Introduction BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Do MBU occur? BFIT HSPICE SBU 8.6e-7 7.4e-7 MBU 9.0e-7 12.6e-7 total FIT 1.7e-6 2.0e-6 Is MBU a concern? Circuit %SBU %MBU s5378 0.8388 0.1612 s9234 0.8625 0.1375 s13207 0.9379 0.0621 s15850 0.9148 0.0852 s38417 0.8812 0.1188 s38584 0.9538 0.0462 s35932 0.9960 0.0030 Holcomb, Li, Seshia Design as You See FIT 16

Verification Guided Error Resilience using sequential simulation with monitors BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Correctness captured in specifications Hardware monitors synthesized from specifications Estimate failure probabilities using random fault injections Accurately incorporate workload Produces refined ranking Inputs List of formal specifications List of circuit errors E output FP E - the probability that upset E will lead to a violated specification Holcomb, Li, Seshia Design as You See FIT 17

Outline Introduction Analysis Efficient Hardening Introduction Single Event Upset Circuit-level masking System-level masking BFIT: Circuit-level soft error analysis Sequential Simulation with Monitors Analysis Efficient Hardening Holcomb, Li, Seshia Design as You See FIT 18

Chip Multiprocessor (CMP) Router Analysis Efficient Hardening Simplified 2 port version of 5 port design[peh 01] 174 latches, 1300 gates Specification Every incoming flit must be routed correctly within 11 cycles Target Select gates or latches to harden to reduce combinational FIT to FIT TARGET Holcomb, Li, Seshia Design as You See FIT 19

BFIT analysis of CMP router Introduction Analysis Efficient Hardening Output from BFIT tool: Holcomb, Li, Seshia Design as You See FIT 20

BFIT analysis of CMP router Introduction Analysis Efficient Hardening Output from BFIT tool: Thousands of E observed, but over 94% are SBU Holcomb, Li, Seshia Design as You See FIT 20

System level masking in CMP Analysis Efficient Hardening 94% of SEU are SBU 94% of SEU are SBU Find failure probability (FP E ) for all SBU Holcomb, Li, Seshia Design as You See FIT 21

CMP - all masking factors Introduction Analysis Efficient Hardening Figure: System-level SBU FIT Holcomb, Li, Seshia Design as You See FIT 22

CMP Achieving FIT target by hardening gates or latches Analysis Efficient Hardening Hardening gates Holcomb, Li, Seshia Design as You See FIT 23

CMP Achieving FIT target by hardening gates or latches Analysis Efficient Hardening Hardening gates Assume a hardened gate contributes no FIT Holcomb, Li, Seshia Design as You See FIT 23

CMP Achieving FIT target by hardening gates or latches Analysis Efficient Hardening Hardening gates Assume a hardened gate contributes no FIT To achieve FIT FIT TARGET : Harden 81/1300 gates Holcomb, Li, Seshia Design as You See FIT 23

CMP Achieving FIT target by hardening gates or latches Analysis Efficient Hardening Hardening gates Assume a hardened gate contributes no FIT To achieve FIT FIT TARGET : Harden 81/1300 gates Pareto optimal coverage if all MBU causes system failure Holcomb, Li, Seshia Design as You See FIT 23

CMP Achieving FIT target by hardening gates or latches Analysis Efficient Hardening Hardening gates Hardening latches Assume a hardened gate contributes no FIT To achieve FIT FIT TARGET : Harden 81/1300 gates Pareto optimal coverage if all MBU causes system failure Holcomb, Li, Seshia Design as You See FIT 23

CMP Achieving FIT target by hardening gates or latches Analysis Efficient Hardening Hardening gates Assume a hardened gate contributes no FIT Hardening latches Assume a hardened latch captures no FIT To achieve FIT FIT TARGET : Harden 81/1300 gates Pareto optimal coverage if all MBU causes system failure Holcomb, Li, Seshia Design as You See FIT 23

CMP Achieving FIT target by hardening gates or latches Analysis Efficient Hardening Hardening gates Assume a hardened gate contributes no FIT To achieve FIT FIT TARGET : Hardening latches Assume a hardened latch captures no FIT To achieve FIT FIT TARGET : Harden 81/1300 gates Harden 39/174 latches Pareto optimal coverage if all MBU causes system failure Holcomb, Li, Seshia Design as You See FIT 23

CMP Achieving FIT target by hardening gates or latches Analysis Efficient Hardening Hardening gates Assume a hardened gate contributes no FIT To achieve FIT FIT TARGET : Hardening latches Assume a hardened latch captures no FIT To achieve FIT FIT TARGET : Harden 81/1300 gates Harden 39/174 latches Pareto optimal coverage if all MBU causes system failure Coverage is Pareto optimal with respect to SBU Holcomb, Li, Seshia Design as You See FIT 23

CMP Achieving FIT target by hardening gates or latches Analysis Efficient Hardening Hardening gates Assume a hardened gate contributes no FIT To achieve FIT FIT TARGET : Hardening latches Assume a hardened latch captures no FIT To achieve FIT FIT TARGET : Harden 81/1300 gates Harden 39/174 latches Pareto optimal coverage if all MBU causes system failure Coverage is Pareto optimal with respect to SBU... design as you see FIT Holcomb, Li, Seshia Design as You See FIT 23

Summary Summary Holcomb, Li, Seshia Design as You See FIT 24

Summary Summary A new method for efficiently analyzing system level impact of circuit level upsets Holcomb, Li, Seshia Design as You See FIT 24

Summary Summary A new method for efficiently analyzing system level impact of circuit level upsets Can guide cost-effective hardening of circuit level features Holcomb, Li, Seshia Design as You See FIT 24

Summary Summary A new method for efficiently analyzing system level impact of circuit level upsets Can guide cost-effective hardening of circuit level features Provide designer flexibility to harden either gates or latches Holcomb, Li, Seshia Design as You See FIT 24

Summary Summary A new method for efficiently analyzing system level impact of circuit level upsets Can guide cost-effective hardening of circuit level features Provide designer flexibility to harden either gates or latches MBU poses threat to reliability Holcomb, Li, Seshia Design as You See FIT 24

Summary Thank You BFIT is available at www.eecs.berkeley.edu/~holcomb/bfit.htm Holcomb, Li, Seshia Design as You See FIT 25

N. Miskov-Zivanov and D. Marculescu. Modeling and Optimization for Soft-Error Reliability of Sequential Circuits. IEEE Trans. on CAD of Integ. Circ. and Sys., pages 803 816, May 2008. N. Miskov-Zivanov and D. Marculescu. Circuit Reliability Analysis Using Symbolic Techniques. IEEE Trans. on CAD of Integ. Circ. and Sys., pages 2638 2649, Dec. 2006. R. Baumann. Radiation-induced soft errors in advanced semiconductor technologies. IEEE Trans. Device and Materials Reliability, 5(3):305 316, Sept. 2005. P. Hazucha and C. Svensson. Impact of CMOS technology scaling on the atmospheric neutron soft error rate. IEEE Trans. Nuclear Science, 47(6):2586 2594, Dec 2000. S. Krishnaswamy, et al. On the role of timing masking in reliable logic circuit design. DAC 2008, pages 924 929. L.-S. Peh. Flow Control and Micro-Arch. Mechanisms for Extending the Performance of Interconnection Networks. PhD thesis, Stanford University, August 2001. R. R. Rao, et al. Computing the soft error rate of a combinational logic circuit using parameterized descriptors. IEEE Trans. on CAD of Integ. Circ. and Sys., 26(3):468 479, 2007. S. S. Mukherjee et al. A systematic methodology to compute the architectural vulnerability factors for a high-perf. microprocessor. In MICRO 2003, pages 29 40. S. A. Seshia, et al. Verification-guided soft error resilience. In DATE 2007, pages 1442 1447. P. Shivakumar, et al. Modeling the effect of technology trends on soft error rate of combinational logic, DSN 02, pp. 389-398. B. Zhang, et al. FASER: fast analysis of soft error susceptibility for cell-based designs. ISQED 2006, pages 755-760. M. Zhang and N. R. Shanbhag. Soft-error-rate-analysis (SERA) methodology. IEEE Trans. on CAD of Integrated Circuits and Systems, 25(10):2140 2155, 2006.

2024 © hobbydocbox.com Privacy Policy | Terms of Service | Feedback