Home Articles Main The Cost Of Verification


The Cost Of Verification PDF Print E-mail
User Rating: / 5
Thursday, 24 December 2009 15:59

We spend about one third of our lives in bed, right? doesn’t it make sense then to buy a good bed and not a cheap one? The same can be said for your verification tools as you spend so much of your project time on verification. Buying cheaper tools will only cost you more money at the end of the project.


Michael C. Kavcak is a design and verification expert with almost 20 years of experience and today we're very excited to have him on board. So without further ado, here’s Mike’s important article about the cost of verification:


Verification is the most time consuming and costly part of ASIC design. It can usually take anywhere from 40-80% of the design cycle time and resources to properly verify a chip. This should encourage companies to closely examine their verification methodology, and look at ways to improve it to achieve maximum results. Quite frequently though, the monetary cost of the toolset is the only factor that is considered when deciding on a verification product or methodology. The company chooses the tool which costs the least, which then leads to poor results from their verification effort. In verification, as with most things, you get what you pay for. In this article, I’ll show how spending more money up front can actually save you money in the end.


The most precious resource on any ASIC design is an engineer’s time. All other resources (such as CPUs) can be added or increased, but an individual engineer’s time is finite. Thus, the goal should be to utilize each engineer’s time to perform the highest value tasks they can contribute to the design. This will lead to having the best possible verification done on the design in the time allotted. It’s a waste for engineers to be spending their time performing mundane tasks, or tasks which can be more easily completed in a different toolset.


One of the most important tasks in any verification effort is the creation of a good verification plan. The verification plan lays out all the functionality in the design, and dictates how that functionality will be tested. It takes a significant amount of time to put the plan together, but even more time to keep the plan updated throughout the verification. Some of today’s verification tools implement the concept of a self-updating verification plan. The plan is entered in text, and then items are hyperlinked to coverage, assertions, or other ways to cover the item. Once the item is activated by running a test that covers it, it’s automatically filled in, so the plan keeps itself updated all throughout the verification process. This eliminates constant human updating, saving upwards of several weeks of engineers’ time.

There are many questions to ask when deciding on a toolset that can point you in the right direction. There are several seemingly miniscule features that can dramatically speed up the verification process. A user-friendly and easily searchable set of documentation helps dramatically when building the environment. The ability to easily experiment with new/different code, without having to check out an entire copy of the base environment, saves copious amounts of time. The ability to quickly jump between a high level performance model and an RTL-level model, or the ability to possibly intermix components of different levels will allow parts of the design to move forward while others are lagging behind. Having a full-featured set of basic functionality so that you don’t have to write special libraries or functions save tremendous time and energy.


As I stated, the biggest cost on a project is the engineer’s time. Taking an estimate of an engineer making $2K/week, every work week he spends creating something that’s available in a better tool costs the company $2000. If an engineer ends up having to spend 2 weeks creating an interface for his current verification tool that would’ve been included in the more expensive tool, that’s $4000 and two weeks of the engineer’s time that the company will not get back. Little things add up quickly, like having to spend half a day scouring the web for a solution to a problem that is easily solved by a better tool. These ˜unseen’ expenses can be easily removed by purchasing the tool which works best for the engineers.

I once was employed by a company where I had created the entire verification environment in one tool. When it came time to renew licenses, the company decided to switch tools due to a discount given by another tools vendor. I had to basically throw out my entire suite and start over again, thus wasting much of the work I had done before. I also had to adapt some of the work to the new tool, thus wasting even more time and money. In the end, I’m fairly certain the company saved nothing, and probably cost themselves money by switching to the ˜cheaper’ tools.

Someone once said to me that since you spend about one-third of your life in your bed, you shouldn’t buy a cheap one. The same can be said for your verification tools. You spend up to 75% of your project time on verification. Buying cheap, inferior verification tools (or ones that the engineers are not familiar/comfortable with) will only cost more money at the end of the project.




More articles :

» To Do List 2010

Introducing Philip Americus - a new guest blogger here on Think Verification. Phil is an ASIC veteran who's worked with every phase of ASIC design - from initial concept to tapeout, with an emphasis on verification, including management of both HW...

» Verification Documents - Love Them, Hate Them, But You Can't Ignore Them

Verification Plan (or Test Plan) and Coverage Plan are two documents that specify the features to be tested in the verification process. The first document usually lists the DUT features that need to be covered and the latter - the coverage points...

» Let The New Game Begin

Things are changing. The EDA industry is changing, and the verification world is changing (check out Janick Bergeron's inspiring at SNUG San Jose for a glimpse of the future of verification). One of the major challenges we’re already facing today...

» To Randomize Or Not To Randomize

One of my former colleagues once revealed the fact that she had no less than 70 pairs of shoes. That’s right, seventy! She had been very good at her job and by no means had any plans to start her own shoe business so I asked myself why on earth...

» Top Level Verification - What's The Big Deal?

How to attack your chip from the top? Why is it so difficult to put together a good top level verification plan? Here are a few ideas.

Add comment

Security code

Copyright © 2018 Think Verification - Tips & Insights on ASIC Verification. All Rights Reserved.
Joomla! is Free Software released under the GNU/GPL License.