For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).
Mojo struct
AMDWarpSharedMemoryBarrier
struct AMDWarpSharedMemoryBarrier[size: Int]
Implements a per-warp shared memory barrier for AMD GPUs.
Maintains a per-warp counter tuple so that each warp can independently track progress, and consumers wait on the summed value across all warps.
Parameters
- size (
Int): Number of warps tracked by the barrier.
Implemented traits
AnyType,
Copyable,
ImplicitlyCopyable,
ImplicitlyDeletable,
Movable,
RegisterPassable,
TrivialRegisterPassable
Methods
initialize
def initialize(ref[MutUnsafeAnyOrigin, AddressSpace._value] self)
value
increment
def increment(ref[MutUnsafeAnyOrigin, AddressSpace._value] self, warp_id: Int)
wait_until_greater_or_equal_to
def wait_until_greater_or_equal_to(ref[AddressSpace._value] self, v: Int32)
Was this page helpful?
Thank you! We'll create more content like this.
Thank you for helping us improve!