Mojo function

broadcast

broadcast[val_type: DType, simd_width: Int, //](val: SIMD[val_type, simd_width]) -> SIMD[val_type, simd_width]

Broadcasts a SIMD value from lane 0 to all lanes in the warp.

This function takes a SIMD value from lane 0 and copies it to all other lanes in the warp, effectively broadcasting the value across the entire warp. This is useful for sharing data between threads in a warp without using shared memory.

Parameters:

val_type (DType): The data type of the SIMD elements (e.g. float32, int32).
simd_width (Int): The number of elements in the SIMD vector.

Args:

val (SIMD[val_type, simd_width]): The SIMD value to broadcast from lane 0.

Returns:

A SIMD value where all lanes contain a copy of the input value from lane 0.

broadcast(val: Int) -> Int

Broadcasts an integer value from lane 0 to all lanes in the warp.

This function takes an integer value from lane 0 and copies it to all other lanes in the warp. It provides a convenient way to share scalar integer data between threads without using shared memory.

Args:

val (Int): The integer value to broadcast from lane 0.

Returns:

The broadcast integer value, where all lanes receive a copy of the input from lane 0.

broadcast(val: UInt) -> UInt

Broadcasts an unsigned integer value from lane 0 to all lanes in the warp.

This function takes an unsigned integer value from lane 0 and copies it to all other lanes in the warp. It provides a convenient way to share scalar unsigned integer data between threads without using shared memory.

Args:

val (UInt): The unsigned integer value to broadcast from lane 0.

Returns:

The broadcast unsigned integer value, where all lanes receive a copy of the input from lane 0.