Summary of 3000W AMD Epyc Server Tear-Down, ft. Wendell of Level1Techs

00:00:00 - 00:20:00

The YouTube video "3000W AMD Epyc Server Tear-Down, ft. Wendell of Level1Techs" features an employee from Level1Techs tearing down an AMD Epyc server. The video discusses the features of the server and how it can be used for machine learning and other applications. The video also mentions the high price tag for the server and the level of engineering that went into it.

• 00:00:00 This YouTube video shows a tear-down of a 3000W AMD Epyc Server. The server is designed for AI and machine learning, and is capable of handling power levels up to 80 Plus Titanium. The case is also mechanically advanced, with multiple fans for reduced noise and a well-designed power supply.
• 00:05:00 This YouTube video shows how to tear down an AMD Epyc server. Wendell of Level1Techs shows how to remove the motherboard, graphics card, and CPU. The video also explains how the server is packed with density and how the fins don't start until the card is halfway down the card.
• 00:10:00 This video shows a 3000W AMD Epyc server tear-down. Level1Techs discusses how the server has four GPUs and no direct heat sinks, which allows the use of high- CFM fans. The sound capture process is completed, and the final product shows the card's connectivity and size.
• 00:15:00 The YouTube video "3000W AMD Epyc Server Tear-Down, ft. Wendell of Level1Techs" shows a Level1Techs employee tearing down an AMD Epyc server. The server is dense with features, including super micro engineering and titanium radiating power supplies. The video also mentions potential uses for the server, including research in machine learning, biology, and ai applications. The video concludes with a mention of the high price tag for the server and the level of engineering Super Micro put into it.
• 00:20:00 In this video, a Level1Techs employee, Wendell, takes down an AMD Epyc server. Wendell discusses the new features of the AMD Epyc M5 platform, and how they make it easier to port cuda applications to it. He also mentions that performance engineers at some large companies are interested in using this platform to run their applications as well as possible.