I tried running any model on my 1070 and it instantly crashes my old tower, probably time to get off windows and run linux on it.
replyUnderstated how much of a boon for Linux that AI development has been.
replyThere isn’t any benefit to running a windows machine.
Au contraire, I run models on WSL and my desktop reliably wakes up from sleep. Best of both worlds.
replyWatch out for VRAM/RAM. When programs on WSL tired to take more than there was available it was crashing my WSL hard and even corrupting files on the virtual drive attached to WSL in a very strange manner that made recent files just disappear. I had at leat two projects that had to be rebuilt from past conversation with AI because all their files were gone after few crashes.
replyI asked codex to write WSL config to prevent crashes. It put some limits on WSL and the situation stabilized, but I lost all trust for WSL anyways.
What are the config settings you needed?
replyTIA
Sounds like a hardware issue, though NVIDIA driver issues can't be ruled out, they're much rarer these days
replyMind sharing your llama.cpp settings for that?
reply .\llama-server.exe -m ..\Qwen3.6-35B-A3B-UD-Q4_K_M.gguf -ngl 999 --n-cpu-moe 41 -c 262144 --port 8081 --flash-attn on --cache-type-k turbo4 --cache-type-v turbo3 --no-mmap --mlock --host 0.0.0.0 -t 8 -tb 8 -np 1
Haven't had much time to test it other than asking a few questions & changing some HTML in cline so it might be thick as a brick for all I know, but still worth trying
I just tested it with some risc-v code and it wrote down a "mov" instruction several times.. yeah something needs tuning maybe
reply