SlideShare uma empresa Scribd logo

Pre-Si Verification for Post-Si Validation

•
•
D
DVClub
1 de 29
Baixar para ler offline
Pre-Si Verification for Post-Si Validation Jason Stinson Principal Engineer Enterprise Microprocessor Division Intel Corporation Dec 5th, 2006
2 Intel Corporation Overview
3 Intel Corporation Post-Silicon Validation • Increasingly larger percentage of time-to-money – Design advances result in flattening/shrinking schedules • Better design tools/methodologies • More focus on integration vs. creation – Post-Si Validation schedules continue to grow • Low investment in tools/methodologies • Complexity driven by design size not design “source” • Multiple SKUs effectively create multiple validation cycles
4 Intel Corporation ASIC Time-to-Money CAD: > 50 companies IP: > 100 companies Headcount ~ 50 Stanford Courses ~ 20 Headcount ~ 5 Test: ~ 10 companies Debug: ~ 3 companies Stanford Courses ~ 3 Pre-Silicon Post-Silicon Arch RTL & Synthesis Verification P & R Fab Debug Fab Ramp 2.5 years
5 Intel Corporation Post-Silicon Validation (cont’d) • Industry starting to make larger post-Si investments – Growing EDA industry (DFx insertion, content generation, etc) – Greater design investment in post-Si feature sets – Increase in academic interest/research (slow) • Two largest VLSI challenges: Verification and Validation – Verification = Pre-silicon verification of design thru simulation – Validation = Post-silicon validation of design thru silicon testing – Design is no longer the most difficult issue
6 Intel Corporation How does Verification affect Validation? • Pre-Si verification = Post-Si validation – Same goal: find the bugs, debug ‘em, fix ‘em • Verification needs to: – Move beyond functional verification • Although still complex, not the biggest challenge – Own mixed-signal verification • Interaction of complex electrical effects is huge challenge – Enable faster post-Si validation • Validate functional & mixed-signal power on and reset • Verify design-for-silicon feature set functionality – Provide better Silicon functional content • Electrical validation requires functional content
7 Intel Corporation Focus Areas • Mixed Signal Verification • Power on/Reset Verification • DFx Verification
8 Intel Corporation Mixed Signal Verification • Distinction between Analog and Digital is becoming blurred – Multiple voltage domains – Complex clocking systems – High speed, asynchronous I/O – High level integration (graphics, wireless, etc.) • Major gap in current pre-silicon verification – Functional verification and switch-level simulation insufficient • Lack asynchronous support • Lack levels simulation • Lack complex timing simulation – Manual SPICE simulation cannot keep pace w/complexity • Significant rise in mixed signal bugs during post-silicon
9 Intel Corporation Mixed Signal Bug: Example Fuse Block State Machine fuse_out state_in Vtt Vcache Vtt Vfuse fuse_out state_in Vcache Incorrect State Correct State
10 Intel Corporation Mixed Signal Bug: Example Fuse Block State Machine fuse_out state_in Vtt Vcache Vtt Vfuse fuse_out state_in Vcache Creates “runt” pulses -> invalid state Non-monotonic switching
11 Intel Corporation Mixed Signal Bug Conclusions • Functional verification and switch-level simulation insufficient – No supply ramp transition (0 -> 1) – No analog understanding of ramp (metastable transition) • Manual simulation or Electrical Rules checking could catch – Requires fullchip visibility and power on understanding • Mixed signal verification – Formal extraction of critical power on circuitry & known analog interactions – Basic SPICE engine combined with functional vector stimulus
12 Intel Corporation Power On/Reset Verification • Complexity of power on and reset increasing – More state to reset = more interactions – Assumptions are often invalid (mutex, DC, valid state, etc) – Analog components factor significantly • Emphasizing need for greater mixed signal verification – Unknown IP block interactions • Traditionally treated as lower verification priority – Old mantra: “Gate with PWRGOOD and route RESET to all state” • Eliminated need for significant (any) pre-silicon verification – New mantra: “Formally verify power on and reset” • Costly to get power on/reset signals to all blocks • Complexity of system no longer ensure valid bring up
13 Intel Corporation Power On Bug: Example Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk
14 Intel Corporation Power On Bug: Example (clock assumption) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk Sysclk begins toggling
15 Intel Corporation Power On Bug: Example (clock assumption) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk Three fuse blocks read out • Provide cfg info to PLL • Tell PLL1 & PLL2 to start to lock
16 Intel Corporation Power On Bug: Example (clock assumption) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk PLL1 and PLL2 acquire lock • PLL3 requires other PLLs to be locked • PLL3 starts to lock once PLL1/2 acquire
17 Intel Corporation Power On Bug: Example (clock assumption) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk PLL3 acquires lock
18 Intel Corporation Power On Bug: Example (fuse assumption) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk Fuse reads have to be staged • Supply issue w/simultaneous read out
19 Intel Corporation Power On Bug: Example (reality) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk Fuse block1 readout • PLL1 starts to acquire lock
20 Intel Corporation Power On Bug: Example (reality) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk Fuse block2 readout • Readout occurs after fixed internal delay (µs) • PLL2 starts to acquire lock PLL1 PLL1 acquires lock
21 Intel Corporation Power On Bug: Example (reality) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk Fuse block3 readout • Readout occurs after fixed internal delay (µs) • PLL3 waits for PLL2 to acquire lock PLL1 PLL2 acquires lock PLL2 • PLL3 starts to acquire lock
22 Intel Corporation Power On Bug: Example (reality) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk PLL3 acquires lock PLL1 No problem, right? PLL2
23 Intel Corporation Power On Bug: Example (reality) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 lock1 lock2 enable lock3 sysclk PLL timeout “feature” • If PWRGOOD and sysclk, then XX time to lock • Shutdown if no lock after XX time PLL1 PLL3 Issue: XX time < Fuse block1 + block2 + PLL2 • Clock assumption ≠ fuse assumption
24 Intel Corporation Power on/Reset Conclusions • Bug verified with both functional verification and switch-level – Original verification effort “zero’d out” fuse delay elements • Fuses readout simultaneously in original verification – Low priority low effort poor assumptions • Mixed signal verification also would have caught issue – Requires long simulation times • Fuse delay elements & PLL lock times are very long
25 Intel Corporation DFx Verification • Design for {Manufacturability, Test, Debug, etc} – “Must have” features to enable low cost and rapid product ramp – “Must have” features to enable validation of design & platform • DFx must-have feature need is increasing – Higher state-to-pin ratio (same pin count, more internal state) • More difficult to get at internal state • Takes longer to get at internal state – Design size continues to increase • Must test blocks in parallel to reduce test time & validation effort • Must enable high coverage, automated test generation – ATE I/O capabilities falling behind • Highly configurable ATE I/O cannot keep up with product I/O • DFx verification is typically lowest or zero priority – Results in longer post-silicon schedules and/or steppings
26 Intel Corporation DFx Verification: Case Study • Swap Core DFT Feature – Used on multicore microprocessor designs – Enables single core diagnostic vector to run on any core • Can be run on single core or all cores in parallel – Huge savings in effort and cost • Reduces vector generation requirements • Reduces core validation effort • Reduces test time (cores run in parallel) • Two microprocessor designs – Both multi-core + large cache – Processor #1: Significant effort in SwapCore verification – Processor #2: Low effort in SwapCore verification
27 Intel Corporation DFx Verification: Case Study (cont’d) • Processor #1 Experience – Swap core worked seamlessly with first silicon – Able to run single-core or cores-in-parallel – 80% of validation effort focused on single-core • Manufacturing ensured all cores matched single core validation • Remaining validation verified multi-core operation – Manufacturing test cost met all goals • Processor #2 Experience – Swap core didn’t work completely with first silicon • Required effort to debug issues • Included fixes in next stepping – Swap core still broken on next stepping (onion peeling issue) – Impacted post-Silicon validation effort/schedule • Had to ensure same coverage/functionality – Impacted manufacturing test vector memory • Threatened to become test time issue
28 Intel Corporation Last Thoughts
29 Intel Corporation Last Thoughts • Verification and Validation are the most critical VLSI challenge – Design is no longer the most difficult challenge • Key post-silicon challenges – Schedule/effort limiting bugs • Power on/reset • Electrical bugs – Drives need for mixed signal verification • High speed I/O – Bug escapes • Comprehensive test content coverage Pre-Si verification has huge impact on post-Si schedules

Recomendados

2019 2 testing and verification of vlsi design_verification
2019 2 testing and verification of vlsi design_verification2019 2 testing and verification of vlsi design_verification
2019 2 testing and verification of vlsi design_verification
Usha Mehta
 
SOC Verification using SystemVerilog
SOC Verification using SystemVerilog SOC Verification using SystemVerilog
SOC Verification using SystemVerilog
Ramdas Mozhikunnath
 
Implementing Useful Clock Skew Using Skew Groups
Implementing Useful Clock Skew Using Skew GroupsImplementing Useful Clock Skew Using Skew Groups
Implementing Useful Clock Skew Using Skew Groups
M Mei
 
Spyglass dft
Spyglass dftSpyglass dft
Spyglass dft
kumar gavanurmath
 
UPF-Based Static Low-Power Verification in Complex Power Structure SoC Design...
UPF-Based Static Low-Power Verification in Complex Power Structure SoC Design...UPF-Based Static Low-Power Verification in Complex Power Structure SoC Design...
UPF-Based Static Low-Power Verification in Complex Power Structure SoC Design...
shaotao liu
 
Vlsi Synthesis
Vlsi SynthesisVlsi Synthesis
Vlsi Synthesis
SIVA NAGENDRA REDDY
 
2019 5 testing and verification of vlsi design_fault_modeling
2019 5 testing and verification of vlsi design_fault_modeling2019 5 testing and verification of vlsi design_fault_modeling
2019 5 testing and verification of vlsi design_fault_modeling
Usha Mehta
 
Introduction of testing and verification of vlsi design
Introduction of testing and verification of vlsi designIntroduction of testing and verification of vlsi design
Introduction of testing and verification of vlsi design
Usha Mehta
 

Recomendados

2019 2 testing and verification of vlsi design_verification
2019 2 testing and verification of vlsi design_verification2019 2 testing and verification of vlsi design_verification
2019 2 testing and verification of vlsi design_verification
Usha Mehta
 
SOC Verification using SystemVerilog
SOC Verification using SystemVerilog SOC Verification using SystemVerilog
SOC Verification using SystemVerilog
Ramdas Mozhikunnath
 
Implementing Useful Clock Skew Using Skew Groups
Implementing Useful Clock Skew Using Skew GroupsImplementing Useful Clock Skew Using Skew Groups
Implementing Useful Clock Skew Using Skew Groups
M Mei
 
Spyglass dft
Spyglass dftSpyglass dft
Spyglass dft
kumar gavanurmath
 
UPF-Based Static Low-Power Verification in Complex Power Structure SoC Design...
UPF-Based Static Low-Power Verification in Complex Power Structure SoC Design...UPF-Based Static Low-Power Verification in Complex Power Structure SoC Design...
UPF-Based Static Low-Power Verification in Complex Power Structure SoC Design...
shaotao liu
 
Vlsi Synthesis
Vlsi SynthesisVlsi Synthesis
Vlsi Synthesis
SIVA NAGENDRA REDDY
 
2019 5 testing and verification of vlsi design_fault_modeling
2019 5 testing and verification of vlsi design_fault_modeling2019 5 testing and verification of vlsi design_fault_modeling
2019 5 testing and verification of vlsi design_fault_modeling
Usha Mehta
 
Introduction of testing and verification of vlsi design
Introduction of testing and verification of vlsi designIntroduction of testing and verification of vlsi design
Introduction of testing and verification of vlsi design
Usha Mehta
 
Verification Challenges and Methodologies
Verification Challenges and MethodologiesVerification Challenges and Methodologies
Verification Challenges and Methodologies
Dr. Shivananda Koteshwar
 
01 Transition Fault Detection methods by Swetha
01 Transition Fault Detection methods by Swetha01 Transition Fault Detection methods by Swetha
01 Transition Fault Detection methods by Swetha
swethamg18
 
ASIC design verification
ASIC design verificationASIC design verification
ASIC design verification
Gireesh Kallihal
 
UVM Methodology Tutorial
UVM Methodology TutorialUVM Methodology Tutorial
UVM Methodology Tutorial
Arrow Devices
 
Static_Timing_Analysis_in_detail.pdf
Static_Timing_Analysis_in_detail.pdfStatic_Timing_Analysis_in_detail.pdf
Static_Timing_Analysis_in_detail.pdf
Usha Mehta
 
2019 3 testing and verification of vlsi design_sta
2019 3 testing and verification of vlsi design_sta2019 3 testing and verification of vlsi design_sta
2019 3 testing and verification of vlsi design_sta
Usha Mehta
 
Clock Tree Synthesis.pdf
Clock Tree Synthesis.pdfClock Tree Synthesis.pdf
Clock Tree Synthesis.pdf
Ahmed Abdelazeem
 
CPU Verification
CPU VerificationCPU Verification
CPU Verification
Ramdas Mozhikunnath
 
Session 6 sv_randomization
Session 6 sv_randomizationSession 6 sv_randomization
Session 6 sv_randomization
Nirav Desai
 
SOC design
SOC design SOC design
SOC design
Vinchipsytm Vlsitraining
 
Vlsi interview questions compilation
Vlsi interview questions compilationVlsi interview questions compilation
Vlsi interview questions compilation
Rajesh M
 
Functional verification techniques EW16 session
Functional verification techniques EW16 sessionFunctional verification techniques EW16 session
Functional verification techniques EW16 session
Sameh El-Ashry
 
5. DFT.pptx
5. DFT.pptx5. DFT.pptx
5. DFT.pptx
Ahmed Abdelazeem
 
Challenges in Using UVM at SoC Level
Challenges in Using UVM at SoC LevelChallenges in Using UVM at SoC Level
Challenges in Using UVM at SoC Level
DVClub
 
System verilog control flow
System verilog control flowSystem verilog control flow
System verilog control flow
Pushpa Yakkala
 
Physical design-complete
Physical design-completePhysical design-complete
Physical design-complete
Murali Rai
 
Verification challenges and methodologies - SoC and ASICs
Verification challenges and methodologies - SoC and ASICsVerification challenges and methodologies - SoC and ASICs
Verification challenges and methodologies - SoC and ASICs
Dr. Shivananda Koteshwar
 
Low power in vlsi with upf basics part 1
Low power in vlsi with upf basics part 1Low power in vlsi with upf basics part 1
Low power in vlsi with upf basics part 1
SUNODH GARLAPATI
 
Low-Power Design and Verification
Low-Power Design and VerificationLow-Power Design and Verification
Low-Power Design and Verification
DVClub
 
Timing and Design Closure in Physical Design Flows
Timing and Design Closure in Physical Design Flows Timing and Design Closure in Physical Design Flows
Timing and Design Closure in Physical Design Flows
Olivier Coudert
 
Validating Next Generation CPUs
Validating Next Generation CPUsValidating Next Generation CPUs
Validating Next Generation CPUs
DVClub
 
Intel Atom Processor Pre-Silicon Verification Experience
Intel Atom Processor Pre-Silicon Verification ExperienceIntel Atom Processor Pre-Silicon Verification Experience
Intel Atom Processor Pre-Silicon Verification Experience
DVClub
 

Mais conteúdo relacionado

Mais procurados

Clock Tree Synthesis.pdf
Clock Tree Synthesis.pdfClock Tree Synthesis.pdf
Clock Tree Synthesis.pdf
Ahmed Abdelazeem
 
Session 6 sv_randomization
Session 6 sv_randomizationSession 6 sv_randomization
Session 6 sv_randomization
Nirav Desai
 
5. DFT.pptx
5. DFT.pptx5. DFT.pptx
5. DFT.pptx
Ahmed Abdelazeem
 
Challenges in Using UVM at SoC Level
Challenges in Using UVM at SoC LevelChallenges in Using UVM at SoC Level
Challenges in Using UVM at SoC Level
DVClub
 
Low-Power Design and Verification
Low-Power Design and VerificationLow-Power Design and Verification
Low-Power Design and Verification
DVClub
 

Mais procurados (20)

Verification Challenges and Methodologies
Verification Challenges and MethodologiesVerification Challenges and Methodologies
Verification Challenges and Methodologies
 
01 Transition Fault Detection methods by Swetha
01 Transition Fault Detection methods by Swetha01 Transition Fault Detection methods by Swetha
01 Transition Fault Detection methods by Swetha
 
ASIC design verification
ASIC design verificationASIC design verification
ASIC design verification
 
UVM Methodology Tutorial
UVM Methodology TutorialUVM Methodology Tutorial
UVM Methodology Tutorial
 
Static_Timing_Analysis_in_detail.pdf
Static_Timing_Analysis_in_detail.pdfStatic_Timing_Analysis_in_detail.pdf
Static_Timing_Analysis_in_detail.pdf
 
2019 3 testing and verification of vlsi design_sta
2019 3 testing and verification of vlsi design_sta2019 3 testing and verification of vlsi design_sta
2019 3 testing and verification of vlsi design_sta
 
Clock Tree Synthesis.pdf
Clock Tree Synthesis.pdfClock Tree Synthesis.pdf
Clock Tree Synthesis.pdf
 
CPU Verification
CPU VerificationCPU Verification
CPU Verification
 
Session 6 sv_randomization
Session 6 sv_randomizationSession 6 sv_randomization
Session 6 sv_randomization
 
SOC design
SOC design SOC design
SOC design
 
Vlsi interview questions compilation
Vlsi interview questions compilationVlsi interview questions compilation
Vlsi interview questions compilation
 
Functional verification techniques EW16 session
Functional verification techniques EW16 sessionFunctional verification techniques EW16 session
Functional verification techniques EW16 session
 
5. DFT.pptx
5. DFT.pptx5. DFT.pptx
5. DFT.pptx
 
Challenges in Using UVM at SoC Level
Challenges in Using UVM at SoC LevelChallenges in Using UVM at SoC Level
Challenges in Using UVM at SoC Level
 
System verilog control flow
System verilog control flowSystem verilog control flow
System verilog control flow
 
Physical design-complete
Physical design-completePhysical design-complete
Physical design-complete
 
Verification challenges and methodologies - SoC and ASICs
Verification challenges and methodologies - SoC and ASICsVerification challenges and methodologies - SoC and ASICs
Verification challenges and methodologies - SoC and ASICs
 
Low power in vlsi with upf basics part 1
Low power in vlsi with upf basics part 1Low power in vlsi with upf basics part 1
Low power in vlsi with upf basics part 1
 
Low-Power Design and Verification
Low-Power Design and VerificationLow-Power Design and Verification
Low-Power Design and Verification
 
Timing and Design Closure in Physical Design Flows
Timing and Design Closure in Physical Design Flows Timing and Design Closure in Physical Design Flows
Timing and Design Closure in Physical Design Flows
 

Destaque

Validating Next Generation CPUs
Validating Next Generation CPUsValidating Next Generation CPUs
Validating Next Generation CPUs
DVClub
 
Intel Atom Processor Pre-Silicon Verification Experience
Intel Atom Processor Pre-Silicon Verification ExperienceIntel Atom Processor Pre-Silicon Verification Experience
Intel Atom Processor Pre-Silicon Verification Experience
DVClub
 
Intel Xeon Pre-Silicon Validation: Introduction and Challenges
Intel Xeon Pre-Silicon Validation: Introduction and ChallengesIntel Xeon Pre-Silicon Validation: Introduction and Challenges
Intel Xeon Pre-Silicon Validation: Introduction and Challenges
DVClub
 
Stop Writing Assertions! Efficient Verification Methodology
Stop Writing Assertions! Efficient Verification MethodologyStop Writing Assertions! Efficient Verification Methodology
Stop Writing Assertions! Efficient Verification Methodology
DVClub
 
Verification of Graphics ASICs (Part I)
Verification of Graphics ASICs (Part I)Verification of Graphics ASICs (Part I)
Verification of Graphics ASICs (Part I)
DVClub
 
Verification of Graphics ASICs (Part II)
Verification of Graphics ASICs (Part II)Verification of Graphics ASICs (Part II)
Verification of Graphics ASICs (Part II)
DVClub
 
Intel_Low Power Intelligent Solutions with Intel Atom Processor
Intel_Low Power Intelligent Solutions with Intel Atom ProcessorIntel_Low Power Intelligent Solutions with Intel Atom Processor
Intel_Low Power Intelligent Solutions with Intel Atom Processor
Işınsu Akçetin
 
Efficiency Through Methodology
Efficiency Through MethodologyEfficiency Through Methodology
Efficiency Through Methodology
DVClub
 

Destaque (16)

Validating Next Generation CPUs
Validating Next Generation CPUsValidating Next Generation CPUs
Validating Next Generation CPUs
 
Intel Atom Processor Pre-Silicon Verification Experience
Intel Atom Processor Pre-Silicon Verification ExperienceIntel Atom Processor Pre-Silicon Verification Experience
Intel Atom Processor Pre-Silicon Verification Experience
 
Intel Xeon Pre-Silicon Validation: Introduction and Challenges
Intel Xeon Pre-Silicon Validation: Introduction and ChallengesIntel Xeon Pre-Silicon Validation: Introduction and Challenges
Intel Xeon Pre-Silicon Validation: Introduction and Challenges
 
Stop Writing Assertions! Efficient Verification Methodology
Stop Writing Assertions! Efficient Verification MethodologyStop Writing Assertions! Efficient Verification Methodology
Stop Writing Assertions! Efficient Verification Methodology
 
Intel Atom X Mobile Processors Announcement Slides MWC 2015
Intel Atom X Mobile Processors Announcement Slides MWC 2015Intel Atom X Mobile Processors Announcement Slides MWC 2015
Intel Atom X Mobile Processors Announcement Slides MWC 2015
 
Verification of Graphics ASICs (Part I)
Verification of Graphics ASICs (Part I)Verification of Graphics ASICs (Part I)
Verification of Graphics ASICs (Part I)
 
Verification of Graphics ASICs (Part II)
Verification of Graphics ASICs (Part II)Verification of Graphics ASICs (Part II)
Verification of Graphics ASICs (Part II)
 
AMD_11th_Intl_SoC_Conf_UCI_Irvine
AMD_11th_Intl_SoC_Conf_UCI_IrvineAMD_11th_Intl_SoC_Conf_UCI_Irvine
AMD_11th_Intl_SoC_Conf_UCI_Irvine
 
Intel_Low Power Intelligent Solutions with Intel Atom Processor
Intel_Low Power Intelligent Solutions with Intel Atom ProcessorIntel_Low Power Intelligent Solutions with Intel Atom Processor
Intel_Low Power Intelligent Solutions with Intel Atom Processor
 
Efficiency Through Methodology
Efficiency Through MethodologyEfficiency Through Methodology
Efficiency Through Methodology
 
UVM: Basic Sequences
UVM: Basic SequencesUVM: Basic Sequences
UVM: Basic Sequences
 
Coverage and Introduction to UVM
Coverage and Introduction to UVMCoverage and Introduction to UVM
Coverage and Introduction to UVM
 
Performance testing using hp load runner
Performance testing using hp load runnerPerformance testing using hp load runner
Performance testing using hp load runner
 
ASIC DESIGN FLOW
ASIC DESIGN FLOWASIC DESIGN FLOW
ASIC DESIGN FLOW
 
Intel Processors
Intel ProcessorsIntel Processors
Intel Processors
 
Project Metrics & Measures
Project Metrics & MeasuresProject Metrics & Measures
Project Metrics & Measures
 

Semelhante a Pre-Si Verification for Post-Si Validation

Stinson post si and verification
Stinson post si and verificationStinson post si and verification
Stinson post si and verification
Obsidian Software
 
Challenges in Assessing Single Event Upset Impact on Processor Systems
Challenges in Assessing Single Event Upset Impact on Processor SystemsChallenges in Assessing Single Event Upset Impact on Processor Systems
Challenges in Assessing Single Event Upset Impact on Processor Systems
Wojciech Koszek
 
Validation and-design-in-a-small-team-environment
Validation and-design-in-a-small-team-environmentValidation and-design-in-a-small-team-environment
Validation and-design-in-a-small-team-environment
Obsidian Software
 
Validation and Design in a Small Team Environment
Validation and Design in a Small Team EnvironmentValidation and Design in a Small Team Environment
Validation and Design in a Small Team Environment
DVClub
 
Real-time Inverted Search in the Cloud Using Lucene and Storm
Real-time Inverted Search in the Cloud Using Lucene and StormReal-time Inverted Search in the Cloud Using Lucene and Storm
Real-time Inverted Search in the Cloud Using Lucene and Storm
lucenerevolution
 
Real-Time Inverted Search NYC ASLUG Oct 2014
Real-Time Inverted Search NYC ASLUG Oct 2014Real-Time Inverted Search NYC ASLUG Oct 2014
Real-Time Inverted Search NYC ASLUG Oct 2014
Bryan Bende
 

Semelhante a Pre-Si Verification for Post-Si Validation (20)

Stinson post si and verification
Stinson post si and verificationStinson post si and verification
Stinson post si and verification
 
mod7.pdf
mod7.pdfmod7.pdf
mod7.pdf
 
ICS Performance Lab
ICS Performance LabICS Performance Lab
ICS Performance Lab
 
Module5 Testing and Verification.pdf
Module5 Testing and Verification.pdfModule5 Testing and Verification.pdf
Module5 Testing and Verification.pdf
 
Challenges in Assessing Single Event Upset Impact on Processor Systems
Challenges in Assessing Single Event Upset Impact on Processor SystemsChallenges in Assessing Single Event Upset Impact on Processor Systems
Challenges in Assessing Single Event Upset Impact on Processor Systems
 
Swimming upstream: OPNFV Doctor project case study
Swimming upstream: OPNFV Doctor project case studySwimming upstream: OPNFV Doctor project case study
Swimming upstream: OPNFV Doctor project case study
 
Mike Bartley - Innovations for Testing Parallel Software - EuroSTAR 2012
Mike Bartley - Innovations for Testing Parallel Software - EuroSTAR 2012Mike Bartley - Innovations for Testing Parallel Software - EuroSTAR 2012
Mike Bartley - Innovations for Testing Parallel Software - EuroSTAR 2012
 
Validation and-design-in-a-small-team-environment
Validation and-design-in-a-small-team-environmentValidation and-design-in-a-small-team-environment
Validation and-design-in-a-small-team-environment
 
Validation and Design in a Small Team Environment
Validation and Design in a Small Team EnvironmentValidation and Design in a Small Team Environment
Validation and Design in a Small Team Environment
 
Systems on chip (so c)
Systems on chip (so c)Systems on chip (so c)
Systems on chip (so c)
 
Kubernetes at NU.nl (Kubernetes meetup 2019-09-05)
Kubernetes at NU.nl (Kubernetes meetup 2019-09-05)Kubernetes at NU.nl (Kubernetes meetup 2019-09-05)
Kubernetes at NU.nl (Kubernetes meetup 2019-09-05)
 
Improving Batch-Process Testing Techniques with a Domain-Specific Language
Improving Batch-Process Testing Techniques with a Domain-Specific LanguageImproving Batch-Process Testing Techniques with a Domain-Specific Language
Improving Batch-Process Testing Techniques with a Domain-Specific Language
 
Real-time Inverted Search in the Cloud Using Lucene and Storm
Real-time Inverted Search in the Cloud Using Lucene and StormReal-time Inverted Search in the Cloud Using Lucene and Storm
Real-time Inverted Search in the Cloud Using Lucene and Storm
 
Kernel Recipes 2015: Solving the Linux storage scalability bottlenecks
Kernel Recipes 2015: Solving the Linux storage scalability bottlenecksKernel Recipes 2015: Solving the Linux storage scalability bottlenecks
Kernel Recipes 2015: Solving the Linux storage scalability bottlenecks
 
MODULE 3 process synchronizationnnn.pptx
MODULE 3 process synchronizationnnn.pptxMODULE 3 process synchronizationnnn.pptx
MODULE 3 process synchronizationnnn.pptx
 
Docker–Grid (A On demand and Scalable dockerized selenium grid architecture)
Docker–Grid (A On demand and Scalable dockerized selenium grid architecture)Docker–Grid (A On demand and Scalable dockerized selenium grid architecture)
Docker–Grid (A On demand and Scalable dockerized selenium grid architecture)
 
Slide used at ACM-SAC 2014 by Suzaki
Slide used at ACM-SAC 2014 by SuzakiSlide used at ACM-SAC 2014 by Suzaki
Slide used at ACM-SAC 2014 by Suzaki
 
Model Based System Design PPT MODULE IV.pptx
Model Based System Design PPT MODULE IV.pptxModel Based System Design PPT MODULE IV.pptx
Model Based System Design PPT MODULE IV.pptx
 
What is HIL (HardWare In The Loop)
What is HIL (HardWare In The Loop)What is HIL (HardWare In The Loop)
What is HIL (HardWare In The Loop)
 
Real-Time Inverted Search NYC ASLUG Oct 2014
Real-Time Inverted Search NYC ASLUG Oct 2014Real-Time Inverted Search NYC ASLUG Oct 2014
Real-Time Inverted Search NYC ASLUG Oct 2014
 

Mais de DVClub

IP Reuse Impact on Design Verification Management Across the Enterprise
IP Reuse Impact on Design Verification Management Across the EnterpriseIP Reuse Impact on Design Verification Management Across the Enterprise
IP Reuse Impact on Design Verification Management Across the Enterprise
DVClub
 
Cisco Base Environment Overview
Cisco Base Environment OverviewCisco Base Environment Overview
Cisco Base Environment Overview
DVClub
 
Verification Automation Using IPXACT
Verification Automation Using IPXACTVerification Automation Using IPXACT
Verification Automation Using IPXACT
DVClub
 
Trends in Mixed Signal Validation
Trends in Mixed Signal ValidationTrends in Mixed Signal Validation
Trends in Mixed Signal Validation
DVClub
 
Verification In A Global Design Community
Verification In A Global Design CommunityVerification In A Global Design Community
Verification In A Global Design Community
DVClub
 
Design Verification Using SystemC
Design Verification Using SystemCDesign Verification Using SystemC
Design Verification Using SystemC
DVClub
 
Verification Strategy for PCI-Express
Verification Strategy for PCI-ExpressVerification Strategy for PCI-Express
Verification Strategy for PCI-Express
DVClub
 
SystemVerilog Assertions (SVA) in the Design/Verification Process
SystemVerilog Assertions (SVA) in the Design/Verification ProcessSystemVerilog Assertions (SVA) in the Design/Verification Process
SystemVerilog Assertions (SVA) in the Design/Verification Process
DVClub
 
OpenSPARC T1 Processor
OpenSPARC T1 ProcessorOpenSPARC T1 Processor
OpenSPARC T1 Processor
DVClub
 
Using Assertions in AMS Verification
Using Assertions in AMS VerificationUsing Assertions in AMS Verification
Using Assertions in AMS Verification
DVClub
 
UVM Update: Register Package
UVM Update: Register PackageUVM Update: Register Package
UVM Update: Register Package
DVClub
 
Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...
Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...
Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...
DVClub
 
Verification of Wireless SoCs: No Longer in the Dark Ages
Verification of Wireless SoCs: No Longer in the Dark AgesVerification of Wireless SoCs: No Longer in the Dark Ages
Verification of Wireless SoCs: No Longer in the Dark Ages
DVClub
 
I Never Thought I Would Grow Up to be This Formal
I Never Thought I Would Grow Up to be This FormalI Never Thought I Would Grow Up to be This Formal
I Never Thought I Would Grow Up to be This Formal
DVClub
 
OMAP Verification
OMAP VerificationOMAP Verification
OMAP Verification
DVClub
 
Roy aeroVerifying Power Domains in AeroFONE
Roy aeroVerifying Power Domains in AeroFONERoy aeroVerifying Power Domains in AeroFONE
Roy aeroVerifying Power Domains in AeroFONE
DVClub
 
Advanced Methodologies Used for Top-Level Verification of Mixed Signal Products
Advanced Methodologies Used for Top-Level Verification of Mixed Signal ProductsAdvanced Methodologies Used for Top-Level Verification of Mixed Signal Products
Advanced Methodologies Used for Top-Level Verification of Mixed Signal Products
DVClub
 
Verification Planning and Metrics to Ensure Efficient Program Execution
Verification Planning and Metrics to Ensure Efficient Program ExecutionVerification Planning and Metrics to Ensure Efficient Program Execution
Verification Planning and Metrics to Ensure Efficient Program Execution
DVClub
 
Need of AMS Simulation in Mix IP Verification: Benefits and Challenges
Need of AMS Simulation in Mix IP Verification: Benefits and ChallengesNeed of AMS Simulation in Mix IP Verification: Benefits and Challenges
Need of AMS Simulation in Mix IP Verification: Benefits and Challenges
DVClub
 
Wireless Low Power and Verification Challenges
Wireless Low Power and Verification ChallengesWireless Low Power and Verification Challenges
Wireless Low Power and Verification Challenges
DVClub
 

Mais de DVClub (20)

IP Reuse Impact on Design Verification Management Across the Enterprise
IP Reuse Impact on Design Verification Management Across the EnterpriseIP Reuse Impact on Design Verification Management Across the Enterprise
IP Reuse Impact on Design Verification Management Across the Enterprise
 
Cisco Base Environment Overview
Cisco Base Environment OverviewCisco Base Environment Overview
Cisco Base Environment Overview
 
Verification Automation Using IPXACT
Verification Automation Using IPXACTVerification Automation Using IPXACT
Verification Automation Using IPXACT
 
Trends in Mixed Signal Validation
Trends in Mixed Signal ValidationTrends in Mixed Signal Validation
Trends in Mixed Signal Validation
 
Verification In A Global Design Community
Verification In A Global Design CommunityVerification In A Global Design Community
Verification In A Global Design Community
 
Design Verification Using SystemC
Design Verification Using SystemCDesign Verification Using SystemC
Design Verification Using SystemC
 
Verification Strategy for PCI-Express
Verification Strategy for PCI-ExpressVerification Strategy for PCI-Express
Verification Strategy for PCI-Express
 
SystemVerilog Assertions (SVA) in the Design/Verification Process
SystemVerilog Assertions (SVA) in the Design/Verification ProcessSystemVerilog Assertions (SVA) in the Design/Verification Process
SystemVerilog Assertions (SVA) in the Design/Verification Process
 
OpenSPARC T1 Processor
OpenSPARC T1 ProcessorOpenSPARC T1 Processor
OpenSPARC T1 Processor
 
Using Assertions in AMS Verification
Using Assertions in AMS VerificationUsing Assertions in AMS Verification
Using Assertions in AMS Verification
 
UVM Update: Register Package
UVM Update: Register PackageUVM Update: Register Package
UVM Update: Register Package
 
Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...
Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...
Verification of the QorIQ Communication Platform Containing CoreNet Fabric wi...
 
Verification of Wireless SoCs: No Longer in the Dark Ages
Verification of Wireless SoCs: No Longer in the Dark AgesVerification of Wireless SoCs: No Longer in the Dark Ages
Verification of Wireless SoCs: No Longer in the Dark Ages
 
I Never Thought I Would Grow Up to be This Formal
I Never Thought I Would Grow Up to be This FormalI Never Thought I Would Grow Up to be This Formal
I Never Thought I Would Grow Up to be This Formal
 
OMAP Verification
OMAP VerificationOMAP Verification
OMAP Verification
 
Roy aeroVerifying Power Domains in AeroFONE
Roy aeroVerifying Power Domains in AeroFONERoy aeroVerifying Power Domains in AeroFONE
Roy aeroVerifying Power Domains in AeroFONE
 
Advanced Methodologies Used for Top-Level Verification of Mixed Signal Products
Advanced Methodologies Used for Top-Level Verification of Mixed Signal ProductsAdvanced Methodologies Used for Top-Level Verification of Mixed Signal Products
Advanced Methodologies Used for Top-Level Verification of Mixed Signal Products
 
Verification Planning and Metrics to Ensure Efficient Program Execution
Verification Planning and Metrics to Ensure Efficient Program ExecutionVerification Planning and Metrics to Ensure Efficient Program Execution
Verification Planning and Metrics to Ensure Efficient Program Execution
 
Need of AMS Simulation in Mix IP Verification: Benefits and Challenges
Need of AMS Simulation in Mix IP Verification: Benefits and ChallengesNeed of AMS Simulation in Mix IP Verification: Benefits and Challenges
Need of AMS Simulation in Mix IP Verification: Benefits and Challenges
 
Wireless Low Power and Verification Challenges
Wireless Low Power and Verification ChallengesWireless Low Power and Verification Challenges
Wireless Low Power and Verification Challenges
 

Pre-Si Verification for Post-Si Validation

  • 1. Pre-Si Verification for Post-Si Validation Jason Stinson Principal Engineer Enterprise Microprocessor Division Intel Corporation Dec 5th, 2006
  • 3. 3 Intel Corporation Post-Silicon Validation • Increasingly larger percentage of time-to-money – Design advances result in flattening/shrinking schedules • Better design tools/methodologies • More focus on integration vs. creation – Post-Si Validation schedules continue to grow • Low investment in tools/methodologies • Complexity driven by design size not design “source” • Multiple SKUs effectively create multiple validation cycles
  • 4. 4 Intel Corporation ASIC Time-to-Money CAD: > 50 companies IP: > 100 companies Headcount ~ 50 Stanford Courses ~ 20 Headcount ~ 5 Test: ~ 10 companies Debug: ~ 3 companies Stanford Courses ~ 3 Pre-Silicon Post-Silicon Arch RTL & Synthesis Verification P & R Fab Debug Fab Ramp 2.5 years
  • 5. 5 Intel Corporation Post-Silicon Validation (cont’d) • Industry starting to make larger post-Si investments – Growing EDA industry (DFx insertion, content generation, etc) – Greater design investment in post-Si feature sets – Increase in academic interest/research (slow) • Two largest VLSI challenges: Verification and Validation – Verification = Pre-silicon verification of design thru simulation – Validation = Post-silicon validation of design thru silicon testing – Design is no longer the most difficult issue
  • 6. 6 Intel Corporation How does Verification affect Validation? • Pre-Si verification = Post-Si validation – Same goal: find the bugs, debug ‘em, fix ‘em • Verification needs to: – Move beyond functional verification • Although still complex, not the biggest challenge – Own mixed-signal verification • Interaction of complex electrical effects is huge challenge – Enable faster post-Si validation • Validate functional & mixed-signal power on and reset • Verify design-for-silicon feature set functionality – Provide better Silicon functional content • Electrical validation requires functional content
  • 7. 7 Intel Corporation Focus Areas • Mixed Signal Verification • Power on/Reset Verification • DFx Verification
  • 8. 8 Intel Corporation Mixed Signal Verification • Distinction between Analog and Digital is becoming blurred – Multiple voltage domains – Complex clocking systems – High speed, asynchronous I/O – High level integration (graphics, wireless, etc.) • Major gap in current pre-silicon verification – Functional verification and switch-level simulation insufficient • Lack asynchronous support • Lack levels simulation • Lack complex timing simulation – Manual SPICE simulation cannot keep pace w/complexity • Significant rise in mixed signal bugs during post-silicon
  • 9. 9 Intel Corporation Mixed Signal Bug: Example Fuse Block State Machine fuse_out state_in Vtt Vcache Vtt Vfuse fuse_out state_in Vcache Incorrect State Correct State
  • 10. 10 Intel Corporation Mixed Signal Bug: Example Fuse Block State Machine fuse_out state_in Vtt Vcache Vtt Vfuse fuse_out state_in Vcache Creates “runt” pulses -> invalid state Non-monotonic switching
  • 11. 11 Intel Corporation Mixed Signal Bug Conclusions • Functional verification and switch-level simulation insufficient – No supply ramp transition (0 -> 1) – No analog understanding of ramp (metastable transition) • Manual simulation or Electrical Rules checking could catch – Requires fullchip visibility and power on understanding • Mixed signal verification – Formal extraction of critical power on circuitry & known analog interactions – Basic SPICE engine combined with functional vector stimulus
  • 12. 12 Intel Corporation Power On/Reset Verification • Complexity of power on and reset increasing – More state to reset = more interactions – Assumptions are often invalid (mutex, DC, valid state, etc) – Analog components factor significantly • Emphasizing need for greater mixed signal verification – Unknown IP block interactions • Traditionally treated as lower verification priority – Old mantra: “Gate with PWRGOOD and route RESET to all state” • Eliminated need for significant (any) pre-silicon verification – New mantra: “Formally verify power on and reset” • Costly to get power on/reset signals to all blocks • Complexity of system no longer ensure valid bring up
  • 13. 13 Intel Corporation Power On Bug: Example Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk
  • 14. 14 Intel Corporation Power On Bug: Example (clock assumption) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk Sysclk begins toggling
  • 15. 15 Intel Corporation Power On Bug: Example (clock assumption) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk Three fuse blocks read out • Provide cfg info to PLL • Tell PLL1 & PLL2 to start to lock
  • 16. 16 Intel Corporation Power On Bug: Example (clock assumption) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk PLL1 and PLL2 acquire lock • PLL3 requires other PLLs to be locked • PLL3 starts to lock once PLL1/2 acquire
  • 17. 17 Intel Corporation Power On Bug: Example (clock assumption) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk PLL3 acquires lock
  • 18. 18 Intel Corporation Power On Bug: Example (fuse assumption) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk Fuse reads have to be staged • Supply issue w/simultaneous read out
  • 19. 19 Intel Corporation Power On Bug: Example (reality) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk Fuse block1 readout • PLL1 starts to acquire lock
  • 20. 20 Intel Corporation Power On Bug: Example (reality) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk Fuse block2 readout • Readout occurs after fixed internal delay (µs) • PLL2 starts to acquire lock PLL1 PLL1 acquires lock
  • 21. 21 Intel Corporation Power On Bug: Example (reality) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk Fuse block3 readout • Readout occurs after fixed internal delay (µs) • PLL3 waits for PLL2 to acquire lock PLL1 PLL2 acquires lock PLL2 • PLL3 starts to acquire lock
  • 22. 22 Intel Corporation Power On Bug: Example (reality) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 PLL3 lock1 lock2 enable lock3 sysclk PLL3 acquires lock PLL1 No problem, right? PLL2
  • 23. 23 Intel Corporation Power On Bug: Example (reality) Fuse Block1 Fuse Block2 Fuse Block3 PLL1 PLL2 lock1 lock2 enable lock3 sysclk PLL timeout “feature” • If PWRGOOD and sysclk, then XX time to lock • Shutdown if no lock after XX time PLL1 PLL3 Issue: XX time < Fuse block1 + block2 + PLL2 • Clock assumption ≠ fuse assumption
  • 24. 24 Intel Corporation Power on/Reset Conclusions • Bug verified with both functional verification and switch-level – Original verification effort “zero’d out” fuse delay elements • Fuses readout simultaneously in original verification – Low priority low effort poor assumptions • Mixed signal verification also would have caught issue – Requires long simulation times • Fuse delay elements & PLL lock times are very long
  • 25. 25 Intel Corporation DFx Verification • Design for {Manufacturability, Test, Debug, etc} – “Must have” features to enable low cost and rapid product ramp – “Must have” features to enable validation of design & platform • DFx must-have feature need is increasing – Higher state-to-pin ratio (same pin count, more internal state) • More difficult to get at internal state • Takes longer to get at internal state – Design size continues to increase • Must test blocks in parallel to reduce test time & validation effort • Must enable high coverage, automated test generation – ATE I/O capabilities falling behind • Highly configurable ATE I/O cannot keep up with product I/O • DFx verification is typically lowest or zero priority – Results in longer post-silicon schedules and/or steppings
  • 26. 26 Intel Corporation DFx Verification: Case Study • Swap Core DFT Feature – Used on multicore microprocessor designs – Enables single core diagnostic vector to run on any core • Can be run on single core or all cores in parallel – Huge savings in effort and cost • Reduces vector generation requirements • Reduces core validation effort • Reduces test time (cores run in parallel) • Two microprocessor designs – Both multi-core + large cache – Processor #1: Significant effort in SwapCore verification – Processor #2: Low effort in SwapCore verification
  • 27. 27 Intel Corporation DFx Verification: Case Study (cont’d) • Processor #1 Experience – Swap core worked seamlessly with first silicon – Able to run single-core or cores-in-parallel – 80% of validation effort focused on single-core • Manufacturing ensured all cores matched single core validation • Remaining validation verified multi-core operation – Manufacturing test cost met all goals • Processor #2 Experience – Swap core didn’t work completely with first silicon • Required effort to debug issues • Included fixes in next stepping – Swap core still broken on next stepping (onion peeling issue) – Impacted post-Silicon validation effort/schedule • Had to ensure same coverage/functionality – Impacted manufacturing test vector memory • Threatened to become test time issue
  • 29. 29 Intel Corporation Last Thoughts • Verification and Validation are the most critical VLSI challenge – Design is no longer the most difficult challenge • Key post-silicon challenges – Schedule/effort limiting bugs • Power on/reset • Electrical bugs – Drives need for mixed signal verification • High speed I/O – Bug escapes • Comprehensive test content coverage Pre-Si verification has huge impact on post-Si schedules