Mojo function
mma
mma[c_register_buffer_type: RegisterBuffer, a_register_buffer_type: RegisterMMABuffer, b_buffer_type: KVBuffer, //, tensor_core_mma: TiledTensorCore[out_type, in_type, shape, group_size, transpose_b], BK: Int, prefetch_function: OptionalReg[fn() capturing -> None], swap_a_b: Bool = False, beg_iter: Int = 0, num_iters: Int = 1, prefetched_b_tile: Bool = False](c: c_register_buffer_type, mut a_tile: a_register_buffer_type, mut b_tile: b_buffer_type)
Was this page helpful?
Thank you! We'll create more content like this.
Thank you for helping us improve!