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
TuningConfigSM90
struct TuningConfigSM90
Tuning record for a single (M, N, K) SM90 matmul configuration.
Stores the full set of kernel hyperparameters: MMA shape, block tile shape, pipeline stages, cluster shape, consumer count, multicast flag, optional grid shape, scheduling strategy, split-K count, and raster order, corresponding to one entry in the pre-tuned BF16 lookup table.
Fieldsβ
- βM (
Int): - βN (
Int): - βK (
Int): - βmma_shape (
IndexList[Int(3)]): - βblock_tile_shape (
IndexList[Int(3)]): - βnum_pipeline_stages (
Int): - βcluster_shape (
IndexList[Int(3)]): - βnum_consumer (
Int): - βpartitioned_multicast (
Bool): - βgrid_shape (
OptionalReg[IndexList[Int(2)]]): - βschedule (
MatmulSchedule): - βsplits (
OptionalReg[Int]): - βraster_order (
OptionalReg[RasterOrder]): - βdispatch_group (
TuningGroup):
Implemented traitsβ
AnyType,
Copyable,
ImplicitlyCopyable,
ImplicitlyDeletable,
Movable,
RegisterPassable,
TrivialRegisterPassable,
TuningConfig,
Writable
Methodsβ
__init__β
def __init__(M: Int, N: Int, K: Int, mma_shape: IndexList[Int(3)], block_tile_shape: IndexList[Int(3)], num_pipeline_stages: Int, cluster_shape: IndexList[Int(3)], num_consumer: Int, partitioned_multicast: Bool, grid_shape: OptionalReg[IndexList[Int(2)]] = None, schedule: MatmulSchedule = MatmulSchedule.NONE, splits: OptionalReg[Int] = None, raster_order: OptionalReg[RasterOrder] = None, dispatch_group: TuningGroup = TuningGroup.CORE) -> Self
write_toβ
def write_to(self, mut writer: T)
Writes the tuning config as a string.
Args:
- βwriter (
T): The writer to write to.
Was this page helpful?
Thank you! We'll create more content like this.
Thank you for helping us improve!