Watercooling a Deep Learning Machine

My custom loop watercooled build with 3xGtx 1080ti’s
Exhaust side of fan
Inlet side of fan
EK Laing D5 vario pump
All in one pump-res combo unit from EK. https://www.ekwb.com/shop/ek-xres-100-revo-d5-pwm-incl-pump
Barb vs Compression fittings after https://www.ekwb.com/blog/fittings-and-tubing-guide/
Shorting pins 4 and 5 to run pump without motherboard.
4x PCI-E slots but sufficienct spacing for only 2x GPU’s
Phanteks case, Xeon E5–2680 v2 CPU’s, ASRock EP2C602–4L/D16 motherboard and vertically mounted GPU.
MSI X299 Tomahawk motherboard
EVGA X299 FTW K. Note how there is plenty of room below the bottom slot to allow GPU+waterblock to seat properly. Total of approx. 8cm between top and bottom PCIe 3.0 slots).
Finally 3x GPU’s. (1x Gigabyte 1080ti Gaming OC, 2x Gigabyte 1080ti AORUS Xtreme)
Its a bit chaotic in there, but hey it’s working :-)
learn.model=torch.nn.DataParallel(learn.model, device_ids=[0,1,2])
#enable persistance mode
sudo nvidia-smi -pm 1
#set GPU 0 max power use to 220W
sudo nvidia-smi -i 0 -pl 220
Note how GPU power limit is set to 220W


Watercooling is time-consuming and moderately expensive and only really necessary if you plan to have multiple open-air GPU’s in your machine. If you are able to buy blower style GPU’s and have a motherboard with sufficient PCI-E spacing to fit more than 2x GPU’s next to each-other, I would recommend this as the easiest option.


Serial vs Parallel for GPU, pump and radiator setup is woth some consideration. See some tips in the links below:



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Adrian G

Adrian G

Geophysicist and Deep Learning Practitioner