5 types of post-silicon validation and why they matter

The Internet of Things is so pervasive it impacts everyday life for most people on the planet. Complex computing devices can be found within small household items, wearables like earbuds and watches, and automobiles. Being surrounded by these things daily contributes to improved productivity and better communication, but these effects necessitate the devices working the way they should.

Embedded computing devices usually use a system-on-chip, or SoC, desig n paradigm or an integrated circuit to connect and communicate with other components. SoC design strategies help promote quick design turnaround times. Validation is necessary after the design phase to ensure that it functions properly.

Here’s more about the types of post-silicon validation and why they’re so important. Each one plays a vital role in the development and production of various products.

Pre-silicon versus post-silicon validation

Testing processes are essential for any product type. There are two distinct phases of validation for silicon products: pre-silicon and post-silicon validation. Pre-silicon validation involves testing in a virtual environment, such as with a simulator or emulator. In contrast, post-silicon validation involves real-life scenarios using actual devices.

Pre-silicon validation is an important first step in product development and testing because it helps identify bugs earlier in the design process. It ensures the SoC has a basic level of operational functionality, and it does so with minimal development time.

Post-silicon validation is one of the most complex, expensive and important aspects of SoC design. Examiners will operate with various tests and software to validate the SoC and ensure it works under actual operating conditions.

Post-silicon validation accounts for up to 70% of the overall time and resources dedicated to SoC design and manufacturing.

Five types of post-silicon tests

The most critical and fundamental aspect of post-silicon testing is having a plan, including establishing the coverage targets, corner cases and functionality. There are five main types of post-silicon validation, and they all serve various purposes in manufacturing.

1. Functional bug hunting

Functional bug hunting is a great way to perform an encompassing, high-impact post-silicon validation. In this method, testers use various methods to run the SoC through different scenarios. These methods include:

• Using the system within the bounds of a normal usage scenario
• Operating with an alternative or nontypical workflow
• Deliberately attacking the system to find vulnerabilities

2. Random instruction testing

A random instruction test is helpful in scenarios where the quantity of testing that needs to be done exceeds the amount of time allotted. Sometimes, the complexity of the function prevents the post-silicon validation method from running every scenario. A random instruction test lets the system run a random selection of scenarios and determine if the SoC passes or fails that test.

The benefits of a random instruction test include that it finds real bugs that are relevant to the product and the tests are relatively quick to complete. However, it may only find basic bugs and requires a large volume of testing to be accurate.

3. Memory subsystem validation

A memory subsystem is a hardware and software component designed to retrieve and store a large amount of information. It should be able to support input and output data connections, but this system is extremely prone to bugs. As a result, post-silicon testing is immensely important. Bugs in the memory subsystem can have a devastating impact on the product’s operation.

4. I/O concurrency

The communication between the input and the output within an information processing system, or the I/O, is essential for proper SoC functioning. I/O concurrency testing ensures that necessary inputs and outputs can happen simultaneously without impacting the system's functionality as a whole.

Many industries have standards and regulations, as well as other systems an SoC needs to be compatible with. Post-silicon validation can ensure industry standards are met to the specified guidelines and that the system is compatible with others. This is vital so products work as intended and companies are not fined for shortcomings.

Why Is Post-Silicon Validation Important?

Post-silicon validation aims to ensure the product is ready for the market by identifying potential errors or operational issues. To do this, they must run the SoC through rigorous tests and processes that mimic how the product will be used in real life.

Even the simplest things (sponges, for example) must undergo rigorous testing. Producing any type of product requires significant time and monetary investment, so ensuring it's free from errors is essential before moving on to the next manufacturing stages, such as large-scale production and distribution. 

In addition to functional components, it’s also essential to test the physical features to ensure they’re sufficient for the type of wear they’re being produced for. A standard cellphone may not need the rigorous testing required by a satellite phone marketed toward the military.

There’s an increasing interdependence on computing devices for everything from tracking fitness goals to completing household chores to performing life-saving surgeries. Therefore, the importance of post-silicon validation – and the necessity of investing heavily into this process – will only increase.

Future trends in SoC validation

Manufacturers pour time and money into research and development for their products. Pre-silicon validation helps catch most surface-level bugs, but post-silicon validation is essential to find the more complex and microcode bugs that impact the real usage of a product.

Industries will see an uptick in the need for comprehensive post-silicon validation tests as items in the IoT play a more significant role in various tasks. For example, they can expect increasingly complex usage scenarios as technology progresses and SoC products become more advanced. If current trends remain steady, they can also expect to produce and rely more on items within the IoT.

No product will be completely free of bugs. However, ensuring they are extensively tested and fully functional before they hit the market will protect against issues from the loss of profit margins to lawsuits and litigation resulting from product failures. Any company that makes and sells products must ensure they work as they should to maintain a good reputation in their industry.