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 function

get_num_partitions

def get_num_partitions[micro_kernel_height: Int, micro_kernel_f_size: Int](num_threads: Int, conv_shape: ConvShape) -> IndexList[Int(4)]

Partition the workload in (batch&group, C, F, H) dimensions. HOWO is the combination of HO and WO dimensions. The actual number of tasks are the product of return num_partitions.

Parameters:

micro_kernel_height (Int): Height of the micro kernel in the output spatial dimension.
micro_kernel_f_size (Int): Size of the micro kernel along the output channel dimension.

Args:

num_threads (Int): Number of parallel threads available for partitioning.
conv_shape (ConvShape): Shape descriptor of the transposed convolution.

Returns:

IndexList[Int(4)]