cuda_info: Report CUDA Build and Runtime Capability

In fastPLS: Fast Partial Least Squares for High-Dimensional Data

Report CUDA Build and Runtime Capability

Description

Distinguish a functional CUDA build from a package built without CUDA and from an explicit diagnostic-only build.

Usage

cuda_info()

Details

A functional result requires both CUDA code compiled into fastPLS and at least one device reported by the CUDA runtime. fastPLS never substitutes the CPU backend for an explicit CUDA request. CUDA version integers use the CUDA Runtime API representation; for example, 13000 denotes CUDA 13.0.

Value

A named list containing status, compiled, available, diagnostic_only, device_count, runtime_version, driver_version, and no_cpu_fallback.

See Also

has_cuda, pls

Examples

cuda_info()

fastPLS documentation built on Sept. 29, 2026, 1:06 a.m.