wave_reduce_max

core::arch::amdgpu

Function wave_reduce_max 

Source 
pub fn wave_reduce_max<const STRATEGY: u32>(value: i32) -> i32
🔬This is a nightly-only experimental API. (stdarch_amdgpu #149988)
Available on target_arch=amdgpu only.
Expand description

Performs an arithmetic max reduction on the signed values provided by each lane in the wavefront.

The STRATEGY argument is a hint for the reduction strategy.

  • 0: Target default preference
  • 1: Iterative strategy
  • 2: DPP

If target does not support the DPP operations (e.g. gfx6/7), reduction will be performed using default iterative strategy.