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
BlackwellProfileWarp
struct BlackwellProfileWarp[workspace_origin: MutOrigin, load_warps: UInt32, mma_warps: UInt32, scheduler_warps: UInt32, epilogue_warps: UInt32, max_entries_per_warp: UInt32, //, WorkspaceManager: BlackwellWarpProfilingWorkspaceManager[load_warps, mma_warps, scheduler_warps, epilogue_warps, max_entries_per_warp], warp_role: UInt32 = UInt32(0)]
This struct calculates execution time for a warp/s, and writes a single entry to the workspace.
Fields
- timeline (
Tuple[UInt64, UInt64]): - workspace (
Span[UInt64, workspace_origin]): - entry_idx (
UInt32):
Implemented traits
AnyType,
Copyable,
ImplicitlyCopyable,
ImplicitlyDeletable,
Movable
comptime members
enable_profiling
comptime enable_profiling = (max_entries_per_warp > UInt32(0))
Methods
__init__
def __init__(out self, workspace: Span[UInt64, workspace_origin], entry_idx: UInt32)
__enter__
def __enter__(mut self)
__exit__
def __exit__(mut self)
Was this page helpful?
Thank you! We'll create more content like this.
Thank you for helping us improve!