Originaly published here.
“Alone we can do so little; together we can do so much.” – Helen Keller
When faster Release to Market (RTM) means a better competitive advantage, it becomes imperative for each stage of the semiconductor life-cycle to contribute to this objective. The Post-Silicon Validation stage (that consumes about 60% of the entire SC product cycle) can make a very significant impact if the productivity of its software processes is improved. And in-house open source communities might just be the way to get this done.
Soliton, with 12+ years of experience in Post Silicon Validation, has helped semiconductor giants in similar missions of standardization drives by providing a validation framework. This framework is integrated with a platform that allows sharing measurement, instrument and test procedure code, thereby enabling in-house open source communities.
Engineers initially resist this concept because they:
- Think their code is unique due to their diverse nature of work
- Are not able to develop shareable code because of their non-software background
- Don’t have a proper platform that helps them share code with less effort
- Are geographically disconnected
The resistance, however, is eliminated when they see how an in-house open source community will help with faster RTM, in addition to better code quality:
Weeks of development effort saved through code reuse
Because of geographically distributed teams, engineers are not often aware of what software assets already exist within the system. This leads to every engineer and team developing and maintaining their own software assets. By having an integrated code sharing platform, engineers can freely download instrument drivers and measurement code as easily as they download an ebook from Amazon. This saves weeks of development effort. After all, everyone loves free goods; especially the ones that save us a lot of time.
Getting the most out of validation engineers
Validation engineers are responsible for measuring, analyzing and calibrating the DUT to ensure the highest quality products – at least, that’s what Google says. Engineers need not be software experts to do validation. But instead of debugging and characterizing the DUT, engineers spend significant time developing and debugging software modules. With the code sharing platform, they can perform validation by downloading best-in-class modules within the company. Therefore the organization benefits more from the validation engineers spending time and efforts on their core functions – characterizing the DUT.
Share and breed ideas
Why do engineers share their code and work? When the community thrives, the user group builds and becomes larger than its non-open counterpart. With inputs from diverse engineering groups across the organization, a problem can be solved in different ways and innovative ideas come out of it. The knowledge and best practices gathered through this process are documented and passed along to future developments.
Expertise retained when experts leave the company
When an engineer leaves the company, the company loses all his experience and work. The time and efforts to train a new engineer results in loss of productivity for the team. With a community platform, the tools, code, knowledge, best practices and all the contributions from an engineer are documented and readily available for anyone to leverage. This way, companies can always retain its expertise and have a higher starting point for new engineer training.
Quickly train and adopt engineers during the reorg
Reorganization of teams and merger of companies takes anywhere from weeks to years of settlement time and to achieve its highest productive state due to the varied software habits of engineers. In such situations, the availability of utilities, tools and shared software assets in the repository can greatly help new engineers to quickly ramp up and start contributing, rather than spending their valuable time on developing new software.
High quality and faster software development
Open source communities are what enabled major technologies to grow at a breath-taking pace. Just as a scientist relies on the scientific community to challenge theories, get feedback and improve, engineers can rely on these open source communities to improve their code. Engineers can quickly identify issues and the debug cycle of the development phase is made faster with a wide number of users providing quick feedback and fixes. This improves the code quality as well. After all, there is a difference between 5 engineers testing the code compared to 100 users testing the code!
By making validation engineers efficient and productive, teams are able to provide better analysis reports quickly to the design team, that in-turn helps in faster RTM. Remember, code alone does not cause this to happen; it takes people working as a community.
Share and Subscribe to my future blogs
Leave a Reply