Vast.ai VS Golem

There are already ways to get around this. For example, renting compute from people who aren't in datacenters. Which is already a thing: https://vast.ai. - Source: Hacker News / 3 months ago

By "SETI" I assume you mean the SETI@Home distributed computing project. There's a two-way market where you can rent out your GPU here: https://vast.ai/. - Source: Hacker News / 4 months ago

- https://vast.ai/ (linked by gchadwick above). - Source: Hacker News / 5 months ago

Have you considered running on a cloud machine instead? You can rent machines on https://vast.ai/ for under $1 an hour that should work for small/medium models (I've mostly been playing with stable diffusion so I don't know what you'd need for an LLM off hand). Good GPUs and Apple hardware is pricey. Get a bit of automation setup with some cloud storage (e.g backblaze B2) and you can have a machine ready to run... - Source: Hacker News / 5 months ago

I have heard vast.ai is cheap but I haven't tried it out. https://websiteinvesting.com/reviews/vast-ai-review/. Source: 5 months ago

Golem, develop Docker applications and make use of their (now) very limited features. It's best suited for heavy calculations, or calculations you can split up between dozens or hundreds of nodes through sharding. A fork is working on bringing GPU & internet access, but it can be hard otherwise. They have a GLM Rewards Program that - generously rewards up to 20 users per month under regular conditions. Source: over 1 year ago

For compute, my experience has been the best with Akash, then Golem, then I have been unsuccessful with any other project as of yet. Both of these supports Docker images, but Golem is painfully thorough with securing providers with sandboxing in both networking and workloads. This makes Akash easier to use right now when wanting to run something more advanced such as a custom backend or a Minecraft Server. Source: over 1 year ago

If you want to run scientific calculations or similar, I highly recommend Golem. Right now, its best applications are ones that can scale by sharding, to use parallel computations. Think doing 100 similar small jobs on 100 computers instead of 1 large job on 1 computer. One average CPU-month costs $3.17, or you can rent 100 CPU-hours for $0.44. Notable examples are blender_cuda which runs on a GPU, and the... Source: over 1 year ago

If you're not using your computer, you can consider letting other people use it! Come checkout golem, a distributed super computer similar to Folding@Home, but for all kinds of computation not just protein research. You even earn some money and it's really easy to get started. Source: over 2 years ago

This is where the math of VPS on demand for testing vs home starts to matter. OR higher buy in but lower ongoing is SBC boards. Raspberry pi, turingpi, ION whatever boards from nvidia. All have higher cost, more limited abilities (in some ways) but FOR SURE are way lower power/heat than traditional low initial cost/higher ongoing. It's a common issue. Getting yourself a NAS or ESOS or SAN or whatever as an always... Source: over 2 years ago