Back to News
Advertisement
Advertisement

⚡ Community Insights

Discussion Sentiment

100% Positive

Analyzed from 295 words in the discussion.

Trending Topics

#heatsink#using#heat#fins#few#hours#fan#should#mining#card

Discussion (8 Comments)Read Original on HackerNews

RALaBarge•about 2 hours ago
I've yet to really run mine since I dont have a great solution for the heat -- I hear it pulls 60w straight so it would be mega hot I presume.
galvan•17 minutes ago
Peeling the center fins of the heatsink (takes a few hours of patience with needle nose pliers/medical forceps) and zip-tieing a Arctic P12 Pro (or any other 120mm high static pressure fan) to it works pretty well. Note that you should probably take the heatsink off while doing this to prevent metal shards from shorting out your board (removing the heatsink might necessitate redoing the thermal pads/putty).
Joel_Mckay•about 2 hours ago
Most just 3D print (or buy) a manifold for a standard PC 12v fan.

I assume duct-tape and cardboard would also work if things are dire. =3

latchkey•25 minutes ago
"The ASRock BC-250 is interesting. Sold as a mining card during the crypto boom of 2021/2022, then killed by Ethereum’s proof-of-work migration, leaving many mining farms out of business with many BC 250s never used or even opened. Dumped as e-waste."

I worked for the farm. We had 20k of them deployed and from what I understand, another 30k in boxes.

aleksejs•about 2 hours ago
Is your heat sink entirely unmodified? Opening up the fins should make it a lot easier to cool.
tobi_bsf•about 2 hours ago
Hard to find these for under $300 these days.
mkarr•about 1 hour ago
You could find these under $100 yesterday on AliExpress with some referral + coupon stacking (rakuten + coupon + paypal pay-in-4 discount). The cheapest sellers often only put out a limited number of items each day, so you have to check back every few hours.
cmrdporcupine•about 2 hours ago
I have one of these I haven't gotten around to futzing with, but have two friends who bought two (each) and have been playing with them for local inference.

Notably we did some research and my one friend has had some luck using the M.2 slot with a PCIe adapter to a ConnectX 3 card and then using that to do multinode coordination (using RDMA and a llama.cpp patch) and has been getting almost 20 tok/s decode out of Qwen3.8 27b dividing it up between two BC250s and a coordinating PC. Prefill is weak though.

A lot of futzing around so not worth it if you don't enjoy that kind of thing, but remarkably cheaper than buying GPUs or a DGX Spark. (I have a Spark so I really shouldn't be bothering... but....)

Fun.