ram memory 4gb Can the HDD manufacturing process still work like this?
ram memory 4gb Can the HDD manufacturing process still work like this?
harddriveSSD OEMram memory 4gbSeagate Edge RX platform applies AI inference to factory manufacturing with NVIDIA EGX.
Seagate Technology delivers tens of millions of hard drives every quarter. Therefore ram memory 4gb, it is critical to guarantee the quality of every drive ram memory 4gb, but this is not an easy task.
hard disk drives are particularly involved in the production process. For example, it takes 1400 processes to manufacture the head alone, and some very small errors can lead to product defects.
Seagate manufactures more hard disk drives than any other company in the same industry. Raghavan Srinivasan, Seagate’s director of global markets, said: “Making mistakes is expensive. Once any abnormality occurs, it will directly lead to the disappointment of subsequent benefits, and we cannot know if there is an anomaly until the entire production process is completed. “
To solve this problem, Seagate is performing anomaly monitoring on recording head slider images through GPU-based AI and machine learning to detect potential problems in a more timely manner.
At first, Seagate was looking for a product that could help with image-based anomaly detection in high-volume manufacturing environments, but there was no analogue on the market. Seagate then began discussing the business with NVIDIA and Hewlett Packard Enterprise.
the first step companies take is to understand the amount of data that is valid. Putting this idea into practice requires a factory capable of processing 17 million images per day and reasoning at 20 positions per second.
“That’s like a Boeing 747 flying a few inches above the ground at 100 times the speed of sound.”
over the course of a year, the companies built a solution that could capture and analyze images of magnetic heads that were simply molecules on the surface of spinning disks.
Seagate deployed the solution on a machine vision defect inspection system in one of its factories, which is primarily used in the manufacture of hard drive read-write heads. Seagate also plans to expand the solution to other sites.
Because the pattern of anomalies is constantly changing, it is impossible to adopt traditional rule-based AI systems and requires the blessing of deep learning.
In addition, Seagate understood the high bandwidth cost of moving so much data back and forth between factories, so it developed a parallel reference architecture, Seagate Edge RX. This architecture builds each factory on the computing “edge,” allowing the inference process to be completed with fewer computing resources.
the data scientist is training an offline deep learning model and will provide an updated model later.
Seagate gives endless possibilities to precision manufacturing
As this technology is integrated into all of Seagate’s manufacturing processes, Seagate hopes to improve efficiency and quality. Reduce cleanroom investment by 20%, manufacturing throughput time by 10%, and achieve ROI of up to 300%.
Srinivasan said there are every indication that the project will eventually lead to significant improvements in quality, yielding considerable benefits for efforts across all of Seagate’s manufacturing facilities.
technology brought by NVIDIA and HPE will play an important role in Seagate’s future development. These technologies include the HPE Edgeline system with NVIDIA T4 GPUs, which enable data acquisition and real-time AI inference in the factory field, and the HPE Apollo system with NVIDIA V100 Tensor Core GPUs, which can be used for AI training. The NVIDIA EGX platform also runs on these systems, helping enterprises securely deploy and manage AI workloads at the edge or in the data center.
Srinivasan said: “NVIDIA has become synonymous with AI and deep learning. “
Seagate sees great potential for doing this work in different directions. Currently, the company is committed to promoting the intelligent manufacturing platform to all its manufacturing sites. Seagate also plans to explore how the solution will impact other manufacturing processes. Notably, Seagate will also explore predictive maintenance based on IoT sensor process recording tools.
The good news for other manufacturers is that Seagate doesn’t want to monopolize all innovations. Seagate has released the Seagate Edge RX Parallel Reference Architecture so that other manufacturers can build similar image anomaly detection solutions in their own manufacturing environments.
Srinivasan said: “We want to drive collaboration. Because we’ve found that by using data well, there are many opportunities to improve the overall manufacturing environment. ”harddriveSSD OEMram memory 4gb
