GitLab

We prefer the nvcuda library, which reports driver versions. When we
dropped cuda v11, we added a safety check for too-old drivers.  What
we missed was the cudart fallback discovery logic didn't have driver
version wired up.  This fixes cudart discovery to expose the driver
version as well so we no longer reject all GPUs if nvcuda didn't work.

Cleaning up go package naming

pointer deref's weren't correct on a few libraries, which explains
some crashes on older systems or miswired symlinks for discovery libraries.

This library will give us the most reliable free VRAM reporting on windows
to enable concurrent model scheduling.

Now that we call the GPU discovery routines many times to
update memory, this splits initial discovery from free memory
updating.

This change adds support for multiple concurrent requests, as well as
loading multiple models by spawning multiple runners. The default
settings are currently set at 1 concurrent request per model and only 1
loaded model at a time, but these can be adjusted by setting
OLLAMA_NUM_PARALLEL and OLLAMA_MAX_LOADED_MODELS.

Leaving the cudart library loaded kept ~30m of memory
pinned in the GPU in the main process.  This change ensures
we don't hold GPU resources when idle.

"cudart init failure: 35" isn't particularly helpful in the logs.