Building deep learning workstation for $2200

Part 1 (2018)
22

Gave the examples/text notebook a go for AWD-LSTM, but the model gave an error when I tried to move it to fp16:
“Expected tensor for argument #1 ‘indices’ to have scalar type Long; but got torch.cuda.HalfTensor instead (while checking arguments for embedding)”
Points to torch/nn/functional.py so appears to be upstream of fastai.

1 Like
23

Sound quite related to fp16…

Keep us posted, should you manage to get it working!

24

@sovann,
Big relief to read you says “28 PCIE lanes should be ok for three GPUs”.
Have you built your system – how is it?

I’m building a system with an i7-7800X and two RTX 2080Ti’s, but got worried that since the CPU can’t do full 16x twice, that I’d hit a bottleneck. (I’m used to working with Xeon CPUs which have lots of PCIE lanes)
-Scott

25

According to Tim Dettmers, who ran a series of systematic experiments, x8 gen3 does actually provide sufficient bandwidth for 2-3 cards. Note also that a single xeon cannot provide the 48 lanes necessary to run 3 cards at full x16 (skylake xeon-w could, but 4-8 lanes are always used for other stuff).

1 Like
26

Thanks, I saw that too (after I posted here)!

Just posted a “Completed Build” on PCPartpicker: https://pcpartpicker.com/b/j7J8TW

The budget for this one was $4000, so I got extra RAM and a second GPU.

27

Hey, guys
I am not tech savvy and building a PC looks pretty complicated for me as well as lack of tech support in case of issues. Does it worth it to buy a prebuilt deep learning desktop? I’ve found a few companies focused on this.



Any thoughts on this?

28

Thoughts are that it depends on your time and resources. What I did (two years ago) was to buy a used gaming computer with GTX 1070 off Craigslist for $800. I see now on portland.craigslist.org several with suitable nvidia GPUs (1070/1080) for $500-$1000.

Because I am a PC and Linux ignoramus, I hired a Linux expert to assess the purchase and get me through a couple of support issues. He works by phone, charges per hour, and was also found via craigslist.

So if you have the time and a little expertise you can trade it for saving money.

Other thoughts:

  • Used last generation nvida GPUs are quite cheap. Miners quitting and gamers upgrading to the latest generation.

  • Cloud computing looks like a very reasonable option right now, and the one I would choose today. You can get started for free with Google Colab, and later upgrade to a faster paid service. $3000 will pay for a lot of hours, plus you are freed from hardware failures, depreciation, and system maintenance. And you can easily work from any location, without the complications of port forwarding, WOL, dynamic DNS, VNC, and SSH.

Hope these comments are more clarifying than complicating!

1 Like
29

Hey DavidBankom,

Great build for a first deep learning PC! A few quick tips:

  1. GPU: Starting with one GPU is good; consider an NVIDIA RTX 2070/2080 for balance.
  2. Motherboard: Ensure it has 3 PCIe x16 slots for future GPUs.
  3. :Power Supply: For 3 GPUs later, a 1000W PSU would be ideal; 750W is fine for now.

Here are some Custom Build Workstations I would suggest for Deep Learning From ProX PC www.proxpc.com

30

In 2026, I’d approach a $2200 deep learning workstation differently than many older build guides suggest. GPU VRAM should be the first priority, not simply the fastest GPU available.

For this budget, a 16GB NVIDIA GPU such as the RTX 5070 Ti can make sense, paired with a Ryzen 7/Ryzen 9 CPU, 32–64GB RAM, and a fast 1–2TB NVMe SSD. The exact balance depends on whether you are training models, running local LLMs, or mainly experimenting with computer vision.

I’d also leave room for a quality PSU, cooling, and future memory upgrades. For deep learning, a properly balanced workstation is usually more valuable than overspending on one component. If you are buying a complete system, companies such as ANT PC can also configure an AI workstation around your specific frameworks and workloads.