Mojo module
profiler
Aliases
MatmulProfileWarp
alias MatmulProfileWarp[warp_role: UInt32, max_entries_per_warp: UInt32] = BlackwellProfileWarp[BlackwellWarpProfilingWorkspaceManager[1, 1, 1, 4, max_entries_per_warp](), warp_role]
Parameters
MatmulWarpSpecializationWorkSpaceManager
alias MatmulWarpSpecializationWorkSpaceManager[max_entries_per_warp: UInt32] = BlackwellWarpProfilingWorkspaceManager[1, 1, 1, 4, max_entries_per_warp]
Parameters
- max_entries_per_warp (
UInt32):
Structs
-
BlackwellProfileWarp: This struct calculates execution time for a warp/s, and writes a single entry to the workspace. -
BlackwellWarpProfilingWorkspaceManager: This struct manages the profiling workspace. The workspaces consists of equal sized chunks, the total number of which is equal to the total number of active SMs. Each SM chunk consists of sequences of entries, with a maximum number of entries per warp role.
Was this page helpful?
Thank you! We'll create more content like this.
Thank you for helping us improve!