llama.cpp
9a5c2a16 - cuda : switch to F16 scalars + tune warps for RTX 2060
Commit
View On GitHub
Commit
1 year ago
cuda : switch to F16 scalars + tune warps for RTX 2060
Author
ggerganov ggerganov
Committer
ggerganov ggerganov
Parents
2c04beeb
  • File
    ggml-cuda.cu
  • tests
    • File
      test-backend-ops.cpp
FAQ Terms Privacy Refunds Impressum