search

shamanDevel / quick-mlp

Fused Multi-layer-perceptrons in CUDA
MIT License
15 stars 1 forks source link

Fix compatibility with nvcc & programs working without CUDA #1

Closed chrismile closed 1 year ago

chrismile commented 1 year ago

Moved spdlog include to source files. Added check for return value of CUDA call in fetchSharedMemory to avoid errors on startup of programs linking to this library when CUDA is not available. Added check whether nlohman_json was already added by another project (like tiny-cuda-nn).