Computational Design of Dense Servers for Immersion Cooling

The growing demands for computational power in cloud computing haveled to a significant increase in the deployment of high-performance servers.The growing power consumption of servers and the heat they produce is on rack to outpace the capacity of conventional air cooling systems, necessitat-ing more efficient cooling solutions such as liquid immersion cooling. Thesuperior heat exchange capabilities of immersion cooling both eliminates theneed for bulky heat sinks, fans, and air flow channels while also unlockingthe potential go beyond conventional 2D blade servers to three-dimensionaldesigns. In this work, we present a computational framework to exploredesigns of servers in three-dimensional space, specifically targeting themaximization of server density within immersion cooling tanks. Our toolis designed to handle a variety of physical and electrical server design con-straints. We demonstrate our optimized designs can reduce server volumeby 25-52% compared to traditional flat server designs. This increased den-sity reduces land usage as well as the amount of liquid used for immersion,with significant reduction in the carbon emissions embodied in datacenterbuildings. We further create physical prototypes to simulate dense serverdesigns and perform real-world experiments in an immersion cooling tankdemonstrating they operate at safe temperatures. This approach marks acritical step forward in sustainable and efficient datacenter management.