Splitting Qwen3.5-397B across 20 rented GPUs that can only reach each other through a relay: what limits the speed, what two days of testing fixed, and the real numbers.
0 comments
No comments yet.
Read the full thread on Hacker News →
Related stories
- Hacker News · 4 points · 10 days ago
- DEV Community · 0 points · about 14 hours ago
- Gmux: Tmux for GPUsgithub.comHacker News · 1 points · 6 days ago
- Tile-based Rasterization in Nvidia GPUsrealworldtech.comLobsters · 1 points · about 10 years ago
- Hacker News · 1 points · 4 days ago
- T. rex teeth indicate it ran as warm as an elephantarstechnica.comArs Technica · 0 points · 11 days ago