SlideShare une entreprise Scribd logo
1  sur  8
Télécharger pour lire hors ligne
Cost Evaluation for Adopting
                        Formal Property Checking

                              Richard Boulton
                                 Icera Inc.

 Related paper to appear in the proceedings of DVCon 2009, 24–26 February 2009


Copyright 2009 Icera Inc        Bristol DVClub, 21 Jan 2009                      1
Motivation

• Formal verification offers improvements in quality
       – But at what cost?
• Design teams / companies may not have extra
  resource to apply formal verification
       – Can FV replace some simulation without extra effort?
       – Can FV be mixed effectively with simulation?
• Icera conducted an evaluation of formal property
  checking (the type of FV most used in industry)
       – Not a tool evaluation!


Copyright 2009 Icera Inc   Cost Evaluation for Adopting FPC, Bristol DVClub, 21 Jan 2009   2
A Few Details

• Small SoC peripheral block (data pump)
       – Significant complexity in the design including a high
         degree of configurability
       – Previously verified at the block level using coverage-
         driven constrained-random simulation
       – Good functional coverage achieved
• Around 100 English properties extracted from the
  functional specification document
       – Some of these were formalized in PSL or SVA
       – Attempted to prove formalized properties using FPC

Copyright 2009 Icera Inc   Cost Evaluation for Adopting FPC, Bristol DVClub, 21 Jan 2009   3
Time Spent on Formal Property Checking

Activity                                                                                   Hours Weeks
Process the RTL                                                                              2.0    0.1
Write natural-language properties for the functional spec.                                  80.5    2.7
Select props for FPC and sim. and write coverage specs                                      61.0    2.0
Deploy property IP                                                                          18.5    0.6
Implement properties in formal language and try to prove                                   109.0    3.6
Total                                                                                      271.0    9.0

• Time sheets kept in half-hour increments
• Conversion to weeks on basis of six hours project work a day


Copyright 2009 Icera Inc   Cost Evaluation for Adopting FPC, Bristol DVClub, 21 Jan 2009                  4
Number of Properties Specified and Formalized

Type of Property Specified (English) Formalized Proportion Done
Definition                                               19                       2        10.5%
Assumption                                                 8                      3        37.5%
Assertion                                                87                     26         28.7%
Total                                                  114                      31         27.2%

•     Intermediate notions were defined for conciseness
•     At least 80% of block behaviour specified
•     31 English properties gave rise to 60 PSL/SVA properties
•     Some additional PSL/SVA assumptions required


Copyright 2009 Icera Inc   Cost Evaluation for Adopting FPC, Bristol DVClub, 21 Jan 2009           5
Estimated Time to Fully Formally Verify the Block

Activity                                                                                   Hours Weeks
Process the RTL                                                                              2.0    0.1
Write natural-language properties for the functional spec.                                 100.6    3.4
Select props for FPC and sim. and write coverage specs                                      76.2    2.5
Deploy property IP                                                                          18.5    0.6
Implement properties in formal language and try to prove                                   500.9   16.7
Total                                                                                      698.2   23.3

• About 23 weeks for FPC versus about 20 weeks for block-
  level simulation (estimated)
• Assumes testbench infrastructure not required elsewhere

Copyright 2009 Icera Inc   Cost Evaluation for Adopting FPC, Bristol DVClub, 21 Jan 2009                  6
Results

•    Two minor bugs found that had eluded simulation
•    Two further omissions from the specification found
•    Plus over 20 significant corrections to the spec.
•    Most assertions proved; a few failures remained
•    Only one assertion hit tool capacity limits
       – But very few “end-to-end” properties were used
• A run with all properties took about an hour



Copyright 2009 Icera Inc   Cost Evaluation for Adopting FPC, Bristol DVClub, 21 Jan 2009   7
Conclusions

• Similar effort for FPC as for simulation
       – If simulation can be avoided completely
• Found some issues not found by simulation
• Challenges:
       – Complexity of clocking
       – Analysis of counter-examples
       – Configurability of design
• Value in writing natural-language properties
• FPC makes assumptions in use of block explicit

Copyright 2009 Icera Inc   Cost Evaluation for Adopting FPC, Bristol DVClub, 21 Jan 2009   8

Contenu connexe

En vedette (8)

Jonathan bromley doulos
Jonathan bromley doulosJonathan bromley doulos
Jonathan bromley doulos
 
Ludden power7 verification
Ludden power7 verificationLudden power7 verification
Ludden power7 verification
 
3 d to _hpc
3 d to _hpc3 d to _hpc
3 d to _hpc
 
D2 audio dv_club_verification_flow
D2 audio dv_club_verification_flowD2 audio dv_club_verification_flow
D2 audio dv_club_verification_flow
 
Gareth edwards xilinx
Gareth edwards xilinxGareth edwards xilinx
Gareth edwards xilinx
 
Vishakantaiah validating
Vishakantaiah validatingVishakantaiah validating
Vishakantaiah validating
 
Darren galpin q4_2008_bristol
Darren galpin q4_2008_bristolDarren galpin q4_2008_bristol
Darren galpin q4_2008_bristol
 
Zhang rtp q307
Zhang rtp q307Zhang rtp q307
Zhang rtp q307
 

Similaire à Bristol 2009 q1_boulton_richard

Bank Data Frank Peterson DB2 10-Early_Experiences_pdf
Bank Data   Frank Peterson DB2 10-Early_Experiences_pdfBank Data   Frank Peterson DB2 10-Early_Experiences_pdf
Bank Data Frank Peterson DB2 10-Early_Experiences_pdfSurekha Parekh
 
The BC Spatial Project – Integrating BC’s Cadastre with FME Server
The BC Spatial Project – Integrating BC’s Cadastre with FME ServerThe BC Spatial Project – Integrating BC’s Cadastre with FME Server
The BC Spatial Project – Integrating BC’s Cadastre with FME ServerSafe Software
 
Richard Dawson Sustainable Ict
Richard Dawson   Sustainable IctRichard Dawson   Sustainable Ict
Richard Dawson Sustainable IctSocitm
 
Hlb private cloud rules of engagement idc
Hlb private cloud rules of engagement   idcHlb private cloud rules of engagement   idc
Hlb private cloud rules of engagement idcYew Jin Kang
 
Elevating Application Performance with the latest IBM COBOL offerings
Elevating Application Performance with the latest IBM COBOL offeringsElevating Application Performance with the latest IBM COBOL offerings
Elevating Application Performance with the latest IBM COBOL offeringsDevOps for Enterprise Systems
 
Track g test strategy - delta
Track g   test strategy - deltaTrack g   test strategy - delta
Track g test strategy - deltachiportal
 
Software System Scalability: Concepts and Techniques (keynote talk at ISEC 2009)
Software System Scalability: Concepts and Techniques (keynote talk at ISEC 2009)Software System Scalability: Concepts and Techniques (keynote talk at ISEC 2009)
Software System Scalability: Concepts and Techniques (keynote talk at ISEC 2009)David Rosenblum
 
Coverage Solutions on Emulators
Coverage Solutions on EmulatorsCoverage Solutions on Emulators
Coverage Solutions on EmulatorsDVClub
 
IBM Z for the Digital Enterprise 2018 - Leverage best language for Transforma...
IBM Z for the Digital Enterprise 2018 - Leverage best language for Transforma...IBM Z for the Digital Enterprise 2018 - Leverage best language for Transforma...
IBM Z for the Digital Enterprise 2018 - Leverage best language for Transforma...DevOps for Enterprise Systems
 
CICS TS for z/OS, From Waterfall to Agile using Rational Jazz Technology - no...
CICS TS for z/OS, From Waterfall to Agile using Rational Jazz Technology - no...CICS TS for z/OS, From Waterfall to Agile using Rational Jazz Technology - no...
CICS TS for z/OS, From Waterfall to Agile using Rational Jazz Technology - no...IBM Danmark
 
21st Century SOA
21st Century SOA21st Century SOA
21st Century SOABob Rhubart
 
Run Your Oracle BI QA Cycles More Effectively
Run Your Oracle BI QA Cycles More EffectivelyRun Your Oracle BI QA Cycles More Effectively
Run Your Oracle BI QA Cycles More EffectivelyKPI Partners
 
Next Generation Network Management for Service Providers
Next Generation Network Management for Service ProvidersNext Generation Network Management for Service Providers
Next Generation Network Management for Service ProvidersCisco Service Provider
 
Agile development in practical world
Agile development in practical worldAgile development in practical world
Agile development in practical worldPerfecto Mobile
 
Experiences with High-Definition Video Conferencing
Experiences with High-Definition Video ConferencingExperiences with High-Definition Video Conferencing
Experiences with High-Definition Video ConferencingVideoguy
 
Intel: Industry Enablement of IO Technologies
Intel: Industry Enablement of IO TechnologiesIntel: Industry Enablement of IO Technologies
Intel: Industry Enablement of IO TechnologiesMemory Fabric Forum
 
What You Missed: OpenStack Summit Austin
What You Missed: OpenStack Summit AustinWhat You Missed: OpenStack Summit Austin
What You Missed: OpenStack Summit AustinNetApp
 

Similaire à Bristol 2009 q1_boulton_richard (20)

Bank Data Frank Peterson DB2 10-Early_Experiences_pdf
Bank Data   Frank Peterson DB2 10-Early_Experiences_pdfBank Data   Frank Peterson DB2 10-Early_Experiences_pdf
Bank Data Frank Peterson DB2 10-Early_Experiences_pdf
 
The BC Spatial Project – Integrating BC’s Cadastre with FME Server
The BC Spatial Project – Integrating BC’s Cadastre with FME ServerThe BC Spatial Project – Integrating BC’s Cadastre with FME Server
The BC Spatial Project – Integrating BC’s Cadastre with FME Server
 
Richard Dawson Sustainable Ict
Richard Dawson   Sustainable IctRichard Dawson   Sustainable Ict
Richard Dawson Sustainable Ict
 
Gluecon 2018 JaC
Gluecon 2018 JaCGluecon 2018 JaC
Gluecon 2018 JaC
 
Strickland dvclub
Strickland dvclubStrickland dvclub
Strickland dvclub
 
Hlb private cloud rules of engagement idc
Hlb private cloud rules of engagement   idcHlb private cloud rules of engagement   idc
Hlb private cloud rules of engagement idc
 
Elevating Application Performance with the latest IBM COBOL offerings
Elevating Application Performance with the latest IBM COBOL offeringsElevating Application Performance with the latest IBM COBOL offerings
Elevating Application Performance with the latest IBM COBOL offerings
 
Track g test strategy - delta
Track g   test strategy - deltaTrack g   test strategy - delta
Track g test strategy - delta
 
Software System Scalability: Concepts and Techniques (keynote talk at ISEC 2009)
Software System Scalability: Concepts and Techniques (keynote talk at ISEC 2009)Software System Scalability: Concepts and Techniques (keynote talk at ISEC 2009)
Software System Scalability: Concepts and Techniques (keynote talk at ISEC 2009)
 
Coverage Solutions on Emulators
Coverage Solutions on EmulatorsCoverage Solutions on Emulators
Coverage Solutions on Emulators
 
IBM Z for the Digital Enterprise 2018 - Leverage best language for Transforma...
IBM Z for the Digital Enterprise 2018 - Leverage best language for Transforma...IBM Z for the Digital Enterprise 2018 - Leverage best language for Transforma...
IBM Z for the Digital Enterprise 2018 - Leverage best language for Transforma...
 
CICS TS for z/OS, From Waterfall to Agile using Rational Jazz Technology - no...
CICS TS for z/OS, From Waterfall to Agile using Rational Jazz Technology - no...CICS TS for z/OS, From Waterfall to Agile using Rational Jazz Technology - no...
CICS TS for z/OS, From Waterfall to Agile using Rational Jazz Technology - no...
 
21st Century SOA
21st Century SOA21st Century SOA
21st Century SOA
 
Run Your Oracle BI QA Cycles More Effectively
Run Your Oracle BI QA Cycles More EffectivelyRun Your Oracle BI QA Cycles More Effectively
Run Your Oracle BI QA Cycles More Effectively
 
Next Generation Network Management for Service Providers
Next Generation Network Management for Service ProvidersNext Generation Network Management for Service Providers
Next Generation Network Management for Service Providers
 
Agile development in practical world
Agile development in practical worldAgile development in practical world
Agile development in practical world
 
Experiences with High-Definition Video Conferencing
Experiences with High-Definition Video ConferencingExperiences with High-Definition Video Conferencing
Experiences with High-Definition Video Conferencing
 
Intel: Industry Enablement of IO Technologies
Intel: Industry Enablement of IO TechnologiesIntel: Industry Enablement of IO Technologies
Intel: Industry Enablement of IO Technologies
 
ISC HPCW talks
ISC HPCW talksISC HPCW talks
ISC HPCW talks
 
What You Missed: OpenStack Summit Austin
What You Missed: OpenStack Summit AustinWhat You Missed: OpenStack Summit Austin
What You Missed: OpenStack Summit Austin
 

Plus de Obsidian Software (20)

Zehr dv club_12052006
Zehr dv club_12052006Zehr dv club_12052006
Zehr dv club_12052006
 
Yang greenstein part_2
Yang greenstein part_2Yang greenstein part_2
Yang greenstein part_2
 
Yang greenstein part_1
Yang greenstein part_1Yang greenstein part_1
Yang greenstein part_1
 
Williamson arm validation metrics
Williamson arm validation metricsWilliamson arm validation metrics
Williamson arm validation metrics
 
Whipp q3 2008_sv
Whipp q3 2008_svWhipp q3 2008_sv
Whipp q3 2008_sv
 
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
 
Tobin verification isglobal
Tobin verification isglobalTobin verification isglobal
Tobin verification isglobal
 
Tierney bq207
Tierney bq207Tierney bq207
Tierney bq207
 
The validation attitude
The validation attitudeThe validation attitude
The validation attitude
 
Thaker q3 2008
Thaker q3 2008Thaker q3 2008
Thaker q3 2008
 
Thaker q3 2008
Thaker q3 2008Thaker q3 2008
Thaker q3 2008
 
Stinson post si and verification
Stinson post si and verificationStinson post si and verification
Stinson post si and verification
 
Shultz dallas q108
Shultz dallas q108Shultz dallas q108
Shultz dallas q108
 
Shreeve dv club_ams
Shreeve dv club_amsShreeve dv club_ams
Shreeve dv club_ams
 
Sharam salamian
Sharam salamianSharam salamian
Sharam salamian
 
Schulz sv q2_2009
Schulz sv q2_2009Schulz sv q2_2009
Schulz sv q2_2009
 
Schulz dallas q1_2008
Schulz dallas q1_2008Schulz dallas q1_2008
Schulz dallas q1_2008
 
Sakar jain
Sakar jainSakar jain
Sakar jain
 
Runner sv q307
Runner sv q307Runner sv q307
Runner sv q307
 
Roy omap validation_dvc_lub_092106
Roy omap validation_dvc_lub_092106Roy omap validation_dvc_lub_092106
Roy omap validation_dvc_lub_092106
 

Bristol 2009 q1_boulton_richard

  • 1. Cost Evaluation for Adopting Formal Property Checking Richard Boulton Icera Inc. Related paper to appear in the proceedings of DVCon 2009, 24–26 February 2009 Copyright 2009 Icera Inc Bristol DVClub, 21 Jan 2009 1
  • 2. Motivation • Formal verification offers improvements in quality – But at what cost? • Design teams / companies may not have extra resource to apply formal verification – Can FV replace some simulation without extra effort? – Can FV be mixed effectively with simulation? • Icera conducted an evaluation of formal property checking (the type of FV most used in industry) – Not a tool evaluation! Copyright 2009 Icera Inc Cost Evaluation for Adopting FPC, Bristol DVClub, 21 Jan 2009 2
  • 3. A Few Details • Small SoC peripheral block (data pump) – Significant complexity in the design including a high degree of configurability – Previously verified at the block level using coverage- driven constrained-random simulation – Good functional coverage achieved • Around 100 English properties extracted from the functional specification document – Some of these were formalized in PSL or SVA – Attempted to prove formalized properties using FPC Copyright 2009 Icera Inc Cost Evaluation for Adopting FPC, Bristol DVClub, 21 Jan 2009 3
  • 4. Time Spent on Formal Property Checking Activity Hours Weeks Process the RTL 2.0 0.1 Write natural-language properties for the functional spec. 80.5 2.7 Select props for FPC and sim. and write coverage specs 61.0 2.0 Deploy property IP 18.5 0.6 Implement properties in formal language and try to prove 109.0 3.6 Total 271.0 9.0 • Time sheets kept in half-hour increments • Conversion to weeks on basis of six hours project work a day Copyright 2009 Icera Inc Cost Evaluation for Adopting FPC, Bristol DVClub, 21 Jan 2009 4
  • 5. Number of Properties Specified and Formalized Type of Property Specified (English) Formalized Proportion Done Definition 19 2 10.5% Assumption 8 3 37.5% Assertion 87 26 28.7% Total 114 31 27.2% • Intermediate notions were defined for conciseness • At least 80% of block behaviour specified • 31 English properties gave rise to 60 PSL/SVA properties • Some additional PSL/SVA assumptions required Copyright 2009 Icera Inc Cost Evaluation for Adopting FPC, Bristol DVClub, 21 Jan 2009 5
  • 6. Estimated Time to Fully Formally Verify the Block Activity Hours Weeks Process the RTL 2.0 0.1 Write natural-language properties for the functional spec. 100.6 3.4 Select props for FPC and sim. and write coverage specs 76.2 2.5 Deploy property IP 18.5 0.6 Implement properties in formal language and try to prove 500.9 16.7 Total 698.2 23.3 • About 23 weeks for FPC versus about 20 weeks for block- level simulation (estimated) • Assumes testbench infrastructure not required elsewhere Copyright 2009 Icera Inc Cost Evaluation for Adopting FPC, Bristol DVClub, 21 Jan 2009 6
  • 7. Results • Two minor bugs found that had eluded simulation • Two further omissions from the specification found • Plus over 20 significant corrections to the spec. • Most assertions proved; a few failures remained • Only one assertion hit tool capacity limits – But very few “end-to-end” properties were used • A run with all properties took about an hour Copyright 2009 Icera Inc Cost Evaluation for Adopting FPC, Bristol DVClub, 21 Jan 2009 7
  • 8. Conclusions • Similar effort for FPC as for simulation – If simulation can be avoided completely • Found some issues not found by simulation • Challenges: – Complexity of clocking – Analysis of counter-examples – Configurability of design • Value in writing natural-language properties • FPC makes assumptions in use of block explicit Copyright 2009 Icera Inc Cost Evaluation for Adopting FPC, Bristol DVClub, 21 Jan 2009 8