Passive Salt Water Cooling Boosts CPU Performance by Almost 33%

tomshardware.com

89 points·ohjeez··6 comments

Is this mostly datacenter tech or something that might end up in consumer desktops etc? I've certainly never heard of alternative cooling tech beyond aluminum or maybe copper heatsinks so definitely interesting. Makes sense in DCs certainly given large scale cooling costs.

The issue I see here is that it depends on some humidity in the air, which DCs usually try to avoid, and homes can vary in.

> The system is self-refilling, too: it can recharge its cooling capacity by absorbing moisture from the air.

Unsure if this recharge cycle happens in batches (requiring downtime/throttling) or acts continuously, but I have a hard time believing the latter.

When cool it absorbs water. When hot it desorbs water. So you would need to idle or throttle. So some sort of “turbo boost duty cycle” sort of thing.

The data centers that I have been around have had humidifiers included in the HVAC system to prevent the air from being too dry. Static electricity is a problem when the air is too dry.

You certainly wouldn't want salt on any circuit boards. Liquid would be extremely conductive, and dried salt crystals wouldn't be great either. Might be like glitter.

It is not NaCl, corrosion might not be an issue. I would call it open circuit water cooling, this thing is just a solid state evaporator. Anyway, as sibling comment said, you don't want the moisture anywhere near your datacenter, and runtime of 400 minutes doesn't sound like production ready product. Maybe it could make sense for burst usage patterns, but standard water-cooling or connecting huge metal blocks seems to be simpler solution. This one could compete at best with liquid nitrogen.

Reply on news.ycombinator.com