Kode and Supermicro together for critical workloads AI solutions performances

January 30th 2024 – Kode, creators of AI-based products and solutions for over 10 years, and Supermicro, Inc. (NASDAQ: SMCI), global technology leader delivering first-to-market innovation for Enterprise, Cloud, AI, Metaverse, and 5G Telco/Edge IT Infrastructure, started a collaboration to offer together reliable solutions enabling manufacturers to easily and safely enter an AI project, guaranteeing performances.

In the dynamic landscape of modern computing, the performance and reliability of on-premise machines play a pivotal role in the seamless functioning of various applications and services. Hence, Kode and Supermicro selected the SYS-111E-FWTR family and tested it with some of Kode’s most advanced AI solutions to verify the fit.

The tests

The tests involved various types of technologies, from computer vision to Multivariate Statistical Process Control, applied in real fields (such as manufacturing and services).

In particular, our team tested:

• A software application for monitoring, controlling and predicting product quality, as well as efficiency and productivity parameters of the plant. The inclusion of Multivariate Statistical Process Control (MSPC) techniques on top of the MES system is fundamental to improving the control of the entire production process, both descriptively and predictively, but encounters significant bottlenecks, due to the computational resources and the time required to perform data preprocessing and to develop PCA models. Besides this is an operation that shall occur at different stages of production to reach the goal of reducing scrap and optimizing yield, with obvious positive effects on overall productivity.

• An advanced software solution for video data analysis in road sign detection and geolocation. Using cutting-edge deep learning algorithms, the platform performs batch analysis of recorded video streams, precisely identifying and classifying various road signs. This solution takes charge of various computationally intensive tasks, integrating object detection and classification with video streams and geolocation. The demand for robust accuracy amplifies the computational load.

• A sophisticated vision system designed to decipher gold engravings. This system seamlessly integrates an object detection model and three distinct classification models, each with the singular purpose of identifying characters within the engravings. The bottleneck in the system primarily stems from the computational resources and time required by the three classification models responsible for character identification. This becomes particularly critical in a low-resource environment, impacting the overall efficiency and timeliness of the system. In a production setting, where timely verification of the correctness of the read is paramount, this bottleneck poses significant challenges.

The testing process involved different machines to compare performances, including those where the AI applications tested have been developed or are currently up and running.

The Results

After 2 weeks of testing, the results were excellent and the machine provided by Supermicro met our predictions. The solution gathered will allow manufacturers to harness AI’s unprecedented benefits in process optimization and automation with many of the advanced technologies that deep learning and AI provide (from Computer Vision to Big Data and Multivariate Statistical Process Control).

The common mission that unites Kode and Supermicro is the desire to provide enterprises with a fast, cost-effective onramp to apply artificial intelligence (AI) and machine learning technologies to solve their most pressing business problems.

Marco Calderisi, CEO of Kode, comments: “Our main effort is always providing our clients with performing and complete solutions. That’s why we decided to partner with Supermicro and carefully test this compounded solution. And we’re glad to say that Supermicro systems are the perfect fit for our clients in the manufacturing and services field”.

Learn more about this partnership

Discover more about the tests conducted on the White Paper “Machine performance testing for rapid application of AI to manufacturers’ business needs”.

Download the White Paper