The invention generally relates to the field of circuit design verification and particularly to integrated circuit design verification. More particularly the invention relates to a system, method and computer program product for functional verification of an integrated circuit.
Design verification, the process of validating the functionality of a design against its specification, is a crucial step in today's design flows. Still, today's large system-on-chip (SoC) integrated circuits (ICs) are only partially verified prior to tapeout, i.e., signing-off for preparation of masks and manufacturing, due to the complexities of the designs as well as the time and resources needed to conclude such a daunting task. This complexity forced designers to rely on coverage, design knowledge, and other metrics to decide when to conclude the verification effort and to declare an IC “verified”. Formal verification, as opposed to simulation, is an exhaustive verification technique often applied to fully verify specific, local functionality of a well-defined portion of an IC design, for example, a communication protocol. This verification is typically undertaken by verification experts who spend significant efforts to verify a specific aspect of the functionality of a design.
In recent years, major advances in the automatic application of formal verification expanded this technology to register-transfer level (RTL) designers who cannot afford the extensive time, effort and expertise required for verifying their design functionality. Automatic formal verification techniques now apply to various problems including clock-domain crossing (CDC), timing exception, power reduction and verification, and static analysis of RTL design, and can target the verification of thousands of properties in days. While formal verification of complex ICs is exhaustive, it usually requires significant time and memory resources to provide conclusive results. When resources are exhausted, current formal verification flows provide inconclusive results. For some formal engines, some information on the depth of analysis might be provided to the user. This information is usually insufficient for the user to make an educated decision on whether to increase system resources or to assume that the system is “reasonably” verified.
For example, formal CDC verification has produced mixed results. On one hand, it can detect functional failures and report them. On the other hand, it is time consuming, since a major effort is required to understand, debug and close any kind of failure reported. The difficulty stems from the fact that CDC verification requires structural as well as functional verification. Synchronization involves structure as well as functionality and both must be properly verified before a tapeout can take place. Incorrectly waiving a reported failure might result in missing a design bug that will manifest itself when the manufactured IC is tested or deployed within a system.
When current automatic formal verification flows exhaust allocated memory and time resources, they provide inconclusive results. This indicates that the analysis was not able to find a design defect, but at the same time it was not able to cover the full state space. As side information, some flows provide the depth to which the analysis was carried. While this information is useful, it is very hard for designers using automatic formal verification flows to correctly interpret this information and decide whether the property can be waived. This is due to the fact that this information is formal analysis-centric.
There is therefore a need in the art for a functional verification solution that can provide the user with design-centric metrics that can be used to catch real design problems. Furthermore, it would be advantageous if such a solution executes in a reasonable amount of time and resource. Such a solution should go beyond the results provided by prior art formal approaches and guide the designer of an IC while analyzing inconclusive properties.
A system and computerized method are provided for performing a reasonable functional verification of the design of an integrated circuit (IC) in which a selected portion of the design of the IC is brought to a target state that is close to a suspected point of failure based on an analysis of a setup for failure property and then functional verification is executed with a number of attempts to trigger the failure property and checking whether a failure is triggered. The functional verification results may be reported as any one or more of a failure to reach the target state, a failure to reach adequate coverage of the attempts to trigger the failure, a property failure being successfully triggered, or a reasonable pass of the IC property being analyzed (i.e., no property failures after adequate coverage of trigger the failure attempts). Adequate coverage is reported as a design-centric metric, such as a ratio of number of trigger the failure operations attempted to a design-specified minimum number of such operations. The functional verification may then proceed to other states of the IC that satisfy the setup for failure and also to other suspected points of failure of the IC. In addition to reporting information sufficient for a user to determine whether or not to sign off on a performed verification, the method can avoid long run times by limiting the verification to bounded analyses focused upon suspected points of failure.
A system and methods for reasonable formal verification provides a user with coverage information that is used for verification signoff. The coverage is calculated based on formal analysis techniques and is provided to the user in terms of design-centric metrics rather than formal-centric metrics. Design-centric metrics include the likes of a number of reads from or writes to memories and a number of bit changes for counters, among many others. Accordingly a setup for failure (SFF) function and a trigger the failure (TTF) function take place. During SFF, formal analysis is applied in an attempt to reach a set of states close enough to suspected failure states. During TTF, formal analysis is applied, starting from the SFF states, to search for a state violating a predetermined property. If results are inconclusive the user is provided with a design-centric coverage metric that can be used in signoff.
Formal verification of an integrated circuit (IC) may require significant resources and terminate without providing the designer with conclusive results that either indicate that the design is correct or present a witness that demonstrates a defect in the design. Therefore the system and methods for reasonable formal verification remedy prior art limitations by providing the user with coverage information that can be used for verification signoff in case the results are inconclusive; and, by avoiding long run time by limiting formal verification to bounded analyses only. The guidance is achieved through formal analysis techniques where analysis is applied to drive the design to a state that is close to a failure.
To further appreciate the invention the following non-limiting example may be considered. A first-in first-out (FIFO) memory that is eight (8) stages deep may be checked for overflow. The guided analysis drives the design to the states where there are six (6) or more writes to the memory. The coverage is calculated based on formal analysis techniques and is provided to the user in terms of design-centric metrics rather than formal-centric metrics. For example, for the FIFO example above, if a FIFO overflow check is not conclusive, the method provides coverage metrics in terms of a number of writes that makes the designer confident that the FIFO does not overflow.
The method comprises of a setup for failure (SFF) function and a trigger the failure (TTF) function. During SFF formal analysis is applied in an attempt to reach a set of states close enough to a failure states. During TTF formal analysis is applied starting from the SFF states to search for a state violating a predetermined property. In case the results are inconclusive, this provides the user with design-centric coverage metric that can be used in signoff. Note that both these steps, SFF and TTF require only bounded formal analysis which will reduce run time and computation effort compared to full model checking traditionally used for verification.
Reference is now made to
A coverage metric for S160-20 can be achieved by implementing the following exemplary and non-limiting method. As an input property is received in S160-10, the designer may define a minimum number of operations (N) to be performed when formal verification takes place. During formal verification, a number of operations (n) is actually performed. The coverage metric is defined as the ratio n/N. In a non-limiting example, a designer may require the verification to perform at least N=20 write cycles into a memory. During verification, n=18 write cycles are performed. Therefore the coverage metric C=18/20, which is 90%. If the coverage metric received in S160-10 defines a, requirement for 90% or above coverage then the target is achieved. However, if the requirement is for 95% coverage, then the coverage metric target was not achieved. This metric may be part of the report of a reasonable pass according to an embodiment.
One of ordinary skill in the art would readily appreciate that the teachings herein of a reasonable formal verification will result in one of four statuses: failure of SFF, failure of the property, a reasonable pass, or a failure due to inadequate coverage. This provides the designer with a rich set of information to conclude verification. For SFF failures, the designer can study the cause of failure and either modify his design to fix defects causing SFF failures or can increase the time of analysis. For a failure of the property, the designer is provided with a witness that demonstrates a manifestation of the bug. For a coverage failure, the designer can increase coverage by increasing the run time or reduce his coverage target to get a reasonable pass. A reasonable pass status, according to an embodiment, means that a pass was reached with appropriate coverage.
Data storage 230 may further comprise storage portion 235 containing the aforementioned abstraction, as well as, but not limited to, the description of the IC, for example in RTL, for example, the RTL description received in S110, including its sub-circuits discussed hereinabove.
The principles of the invention are implemented as hardware, firmware, software or any combination thereof, including but not limited to a CAD system and software products thereof. Moreover, the software is preferably implemented as an application program tangibly embodied on a program storage unit or computer readable medium. The application program may be uploaded to, and executed by, a machine comprising any suitable architecture. Preferably, the machine is implemented on a computer platform having hardware such as one or more central processing units (“CPUs”), a memory, and input/output interfaces. The computer platform may also include an operating system and microinstruction code. The various processes and functions described herein may be either part of the microinstruction code or part of the application program, or any combination thereof, which may be executed by a CPU, whether or not such computer or processor is explicitly shown. In addition, various other peripheral units may be connected to the computer platform such as an additional data storage unit and a printing unit and/or display unit.
The present application claims priority under 35U.S.C. 119(e) from prior U.S. provisional application 61/786,668, filed on Mar. 15, 2013.
Number | Name | Date | Kind |
---|---|---|---|
4727545 | Glackemeyer et al. | Feb 1988 | A |
5465216 | Rotem et al. | Nov 1995 | A |
6178533 | Chang | Jan 2001 | B1 |
6226777 | Zhang | May 2001 | B1 |
6321186 | Yuan et al. | Nov 2001 | B1 |
6446243 | Huang et al. | Sep 2002 | B1 |
6449750 | Tsuchiya | Sep 2002 | B1 |
6553524 | Das | Apr 2003 | B1 |
6691287 | Ganesan et al. | Feb 2004 | B2 |
6925617 | Bayraktaroglu et al. | Aug 2005 | B2 |
7107553 | Lockyear et al. | Sep 2006 | B2 |
7120829 | Shigeta | Oct 2006 | B2 |
7257786 | Kukula | Aug 2007 | B1 |
7283918 | Nozuyama | Oct 2007 | B2 |
7496788 | Alfieri et al. | Feb 2009 | B1 |
7844873 | Funatsu | Nov 2010 | B2 |
7882473 | Baumgartner et al. | Feb 2011 | B2 |
7941771 | Kaszynski et al. | May 2011 | B2 |
8127261 | Auerbach et al. | Feb 2012 | B2 |
8140315 | Walter et al. | Mar 2012 | B2 |
8156462 | Moon | Apr 2012 | B2 |
8205174 | Li et al. | Jun 2012 | B2 |
8239796 | Stevens et al. | Aug 2012 | B2 |
8316332 | Goyal et al. | Nov 2012 | B1 |
8321825 | Stevens et al. | Nov 2012 | B2 |
20030144826 | Mandell et al. | Jul 2003 | A1 |
20050289485 | Willis | Dec 2005 | A1 |
20060041810 | Ishida et al. | Feb 2006 | A1 |
20060271822 | Pal et al. | Nov 2006 | A1 |
20070013403 | Nozuyama | Jan 2007 | A1 |
20080234967 | Vandewiele et al. | Sep 2008 | A1 |
20080256404 | Funatsu | Oct 2008 | A1 |
20090210764 | Nozuyama | Aug 2009 | A1 |
20090297019 | Zafar et al. | Dec 2009 | A1 |
20110175643 | Bogenberger et al. | Jul 2011 | A1 |
20110214096 | Sheeley et al. | Sep 2011 | A1 |
20110286656 | Kulkarni et al. | Nov 2011 | A1 |
20120079442 | Akar et al. | Mar 2012 | A1 |
20120131527 | Kekare | May 2012 | A1 |
Number | Date | Country | |
---|---|---|---|
61786668 | Mar 2013 | US |