Skip to main content

GEMV_F32_ROWS

logicaffeine_compile::codegen_cuda::lane::kernels

Constant GEMV_F32_ROWS 

Source 
pub const GEMV_F32_ROWS: usize = 4;
Expand description

The activation rows a block of lk_gemv_f32 takes, together (its LK_F32_ROWS): few enough that a wide chunk’s rows spread over the chip, every row’s dots summed as one row’s are.