pub fn load<T: RudaType, U: RudaType + Copy, I: RudaRead<T>, W: RudaWrite<U>, O: RudaUnaryOp<T, U>>(
input: &I,
output: &mut W,
convert: &O,
offset: usize,
valid: usize,
lane: usize,
padding: U,
lanes: usize,
items: usize,
striped: bool,
warp_striped: bool,
width: usize,
pad: bool,
)Expand description
Iterator load with explicit conversion and tail semantics. lane is a block lane or logical-warp lane, respectively; input/output must not overlap.