Motivation: ACS Real-Time Data Infrastructure (RTDI) enables fast access to semiconductor test data and optimized control of test operations in real time. A significant component of ACS RTDI is ACS EdgeTM, a high-performance and highly secure platform in the test cell that can host any computation-intensive and memory-intensive workloads without porting the code to the tester-native software environment. MATLAB is a widely used tool in the industry for developing complex computational solutions in engineering applications, including those in semiconductor design and testing. Chip development teams would naturally like the ability to execute their MATLAB applications on the automatic test equipment (ATE) for test development, ramp-to-production, and production test purposes. However, deploying MATLAB applications on the ACS EdgeTM platform poses two significant challenges. First, MATLAB scripts can only run on limited tools such as MATLAB and Octave. Additionally, using MATLAB requires a valid license, while the compatibility of advanced MATLAB libraries with Octave may be limited. Second, deploying a MATLAB application developed on Windows/macOS to a Linux-powered ACS EdgeTM system without compromising efficiency and reliability is challenging.
Methods: To overcome the challenges described above, we present a cross-platform deployment solution that utilizes several tools, including the R2019b MATLAB library compiler or later, .NET 5 SDK or later, Docker, V93000 SmarTest platform, and the ACS EdgeTM MATLAB .NET Linux runtime image. The solution: · Packaging the MATLAB source code with the MATLAB library compiler to generate a standalone .NET library. · Integrating the library with an ACS EdgeTM compatible .NET service program. · Using the .NET command-line interface to generate a Linux-compatible binary. · The resulting files are packaged into a Docker image and pushed to the Advantest Cloud Solutions Container Hub. · The SmarTest client SDK is then used to create the container on ACS EdgeTM and perform the MATLAB runtime.
Results: Our solution allows the native deployment of MATLAB applications onto the ACS Real-Time Data Infrastructure, irrespective of the platform used during development. This approach provides higher runtime efficiency and reliability than other tools like mono. Our solution does not require the installation of any third-party tools.
