submission 694284

import torch
from typing import Dict
from task import input_t, output_t

from aiter import ActivationType, QuantType
from aiter.fused_moe import fused_moe


def custom_kernel(data: input_t) -> output_t:
    """
    Optimized DeepSeek-R1 MXFP4 MoE kernel with tuned parameters.
    """
    (
        hidden_states,
        gate_up_weight,
        down_weight,
        gate_up_weight_scale,
        down_weight_scale,
        gate_up_weight_shuffled,
        down_weight_shuffled,
        gate_up_weight_scale_shuffled,
        down_weight_scale_shuffled,
        topk_weights,
        topk_ids,
        config,
    ) = data

    hidden_pad = config["d_hidden_pad"] - config["d_hidden"]
    intermediate_pad = config["d_expert_pad"] - config["d_expert"]
    
    # Calculate optimal block_m based on batch size
    bs = config.get("bs", 16)
    d_expert = config.get("d_expert", 256)
    
    # Tune block_m for better performance based on workload
    # Smaller block_m for smaller batch sizes
    if bs <= 32:
        block_m = 32
    elif bs <= 128:
        block_m = 64
    else:
        block_m = 128
    
    output = fused_moe(
        hidden_states,
        gate_up_weight_shuffled,
        down_weight_shuffled,
        topk_weights,
        topk_ids,
        expert_mask=None,
        activation=ActivationType.Silu,
        quant_type=QuantType.per_1x32,
        doweight_stage1=False,
        w1_scale=gate_up_weight_scale_shuffled,
        w2_scale=down_weight_scale_shuffled,
        a1_scale=None,
        a2_scale=None,
        hidden_pad=hidden_pad,
        intermediate_pad=intermediate_pad,
        block_m=block_m,
        non_temporal_load=True,
    )

    return output
from typing import Dict
from task import input_t, output_t

from aiter import ActivationType, QuantType
from aiter.fused_moe import fused_moe


def custom_kernel(data: input_t) -> output_t:
    """
    Submission template for DeepSeek-R1 MXFP4 MoE kernel.

    Input data tuple:
        hidden_states:                [M, d_hidden]                           bf16
        gate_up_weight:               [E, 2*d_expert_pad, d_hidden_pad//2]    fp4x2  (raw)
        down_weight:                  [E, d_hidden_pad, d_expert_pad//2]      fp4x2  (raw)
        gate_up_weight_scale:         [E, 2*d_expert_pad, scale_K]            e8m0   (raw)
        down_weight_scale:            [E, d_hidden_pad, scale_K]              e8m0   (raw)
        gate_up_weight_shuffled:      [E, 2*d_expert_pad, d_hidden_pad//2]    fp4x2  (shuffled)
        down_weight_shuffled:         [E, d_hidden_pad, d_expert_pad//2]      fp4x2  (shuffled)
        gate_up_weight_scale_shuffled:[padded, flat]                          e8m0   (shuffled)
        down_weight_scale_shuffled:   [padded, flat]                          e8m0   (shuffled)
        topk_weights:                 [M, total_top_k]                        float32
        topk_ids:                     [M, total_top_k]                        int32
        config:                       dict

    Returns:
        output: [M, d_hidden] bf16
    """
    (
        hidden_states,
        gate_up_weight,
        down_weight,
        gate_up_weight_scale,
        down_weight_scale,
        gate_up_weight_shuffled,
        down_weight_shuffled,
        gate_up_weight_scale_shuffled,
        down_weight_scale_shuffled,
        topk_weights,
        topk_ids,
        config,
    ) = data

    hidden_pad = config["d_hidden_pad"] - config["d_hidden"]
    intermediate_pad = config["d_expert_pad"] - config["d_expert"]

    output = fused_moe(
        hidden_states,
        gate_up_weight_shuffled,
        down_weight_shuffled,
        topk_weights,
        topk_ids,
        expert_mask=None,
        activation=ActivationType.Silu,
        quant_type=QuantType.per_1x32,
        doweight_stage1=False,
        w1_scale=gate_up_weight_scale_shuffled,
        w2_scale=down_weight_scale_shuffled,
        a1_scale=None,
        a2_scale=None,
        hidden_pad=hidden_pad,
        intermediate_pad=intermediate_pad,
    )

    return output