Hi folks,

I currently have a HP ML10v2 with Xeon E3-1265L v3 in it (4c/8t) and 16GB DDR3 ECC. At work, we disposed of (and I now have in my wardrobe) a dual socket Xeon E5-2680 v4 (14c/28t each) and 64GB DDR4 ECC RAM.

Now this is obviously a massive jump in horsepower, and I very much doubt that I have any use case for a total of 28c/56t, so I was considering removing one of the Xeons and keeping it spare. I currently use a 1050TI for transcodes and that will be moved across.

It’s all built on a ASUS Z10PE-D8 WS motherboard, so I’m just looking for anyone else using similar gear, and any tips you might have to reduce power consumption. I haven’t yet gone diving through the BIOS to see what options are there, but I’m hedging between trying to screw down the power consumption, and just seeing if I can sell off the bigger machine and get a more modern Core i5 or similar…

As is, on the HP I’m running Unraid with a dozen docker containers, and that’s about it. I would like to start playing with Ansible, so the bigger system would be good in that regard, but I’m not 100% sure if I want to keep this machine or not.

For reference, my power costs are AUD $0.8087/day supply charge, and AUD $0.2299/kwh. I’m renting so have no options for solar etc.

Thanks!

  • theyreplayingyou@alien.topBEnglish
    1·
    10 months ago

    I have that exact same proc in one of my nodes:

    dell r630 2x E5-2680v4 128gb ram 8x spinning SAS drives 1x quadro K1200 dual SFP+ nic with SFP+ modules

    power consumption: 168w was 108-124w before the K1200 gpu.

    I’m going to hopefully pull those 8 disks and swap in 4x SAS SSD’s this weekend, hopefully get my power consumption back closer to 100w. Then I’m going to throw in one or two T1000 8GB GPUS so I’ll probably be back up to to 175-200w but thats damn impressive for the compute capacity of that box.