submission 125598

"""
Submission v6: Warp-Specialized CuTeDSL NVFP4 GEMM Kernel

This submission implements warp specialization for better parallelism:
- TMA warp (5): Handles all TMA loads from GMEM to SMEM
- MMA warp (4): Executes scale factor S2T copy and MMA operations
- Epilogue warps (0-3): Handle T2R copy and stores to GMEM

Based on the persistent GEMM pattern from dense_blockscaled_gemm_persistent.py.
"""
import torch
from task import input_t, output_t

import cutlass
import cutlass.cute as cute
import cutlass.utils as utils
import cutlass.pipeline as pipeline
from cutlass.cute.nvgpu import cpasync, tcgen05
import cutlass.utils.blackwell_helpers as sm100_utils
import cutlass.utils.blockscaled_layout as blockscaled_utils
from cutlass.cute.runtime import make_ptr

# Kernel configuration parameters
# Tile sizes for M, N, K dimensions
mma_tiler_mnk = (128, 128, 256)
# Shape of the K dimension for the MMA instruction
mma_inst_shape_k = 64
# FP4 data type for A and B
ab_dtype = cutlass.Float4E2M1FN
# FP8 data type for scale factors
sf_dtype = cutlass.Float8E4M3FN
# FP16 output type
c_dtype = cutlass.Float16
# Scale factor block size (16 elements share one scale)
sf_vec_size = 16

# Warp specialization configuration
# 6 warps total = 192 threads
epilog_warp_ids = (0, 1, 2, 3)  # 4 warps for epilogue
mma_warp_id = 4                  # 1 warp for MMA
tma_warp_id = 5                  # 1 warp for TMA
threads_per_cta = 192            # 6 warps * 32 threads

# Pipeline stage counts
num_acc_stage = 1
num_ab_stage = 4  # Best configuration: good latency hiding with try-wait pattern

# Total number of columns in tmem
num_tmem_alloc_cols = 512


# Helper function for ceiling division
def ceil_div(a, b):
    return (a + b - 1) // b


# The CuTe reference implementation for NVFP4 block-scaled GEMM with warp specialization
@cute.kernel
def kernel(
    tiled_mma: cute.TiledMma,
    tma_atom_a: cute.CopyAtom,
    mA_mkl: cute.Tensor,
    tma_atom_b: cute.CopyAtom,
    mB_nkl: cute.Tensor,
    tma_atom_sfa: cute.CopyAtom,
    mSFA_mkl: cute.Tensor,
    tma_atom_sfb: cute.CopyAtom,
    mSFB_nkl: cute.Tensor,
    mC_mnl: cute.Tensor,
    a_smem_layout_staged: cute.ComposedLayout,
    b_smem_layout_staged: cute.ComposedLayout,
    sfa_smem_layout_staged: cute.Layout,
    sfb_smem_layout_staged: cute.Layout,
    num_tma_load_bytes: cutlass.Constexpr[int],
):
    """
    GPU device kernel performing the batched GEMM computation with warp specialization.
    """
    warp_idx = cute.arch.warp_idx()
    warp_idx = cute.arch.make_warp_uniform(warp_idx)
    tidx = cute.arch.thread_idx()

    #
    # Setup cta/thread coordinates
    #
    # Coords inside cluster
    bidx, bidy, bidz = cute.arch.block_idx()

    # Coords outside cluster
    cta_coord = (bidx, bidy, bidz)
    mma_tile_coord_mnl = (
        cta_coord[0] // cute.size(tiled_mma.thr_id.shape),
        cta_coord[1],
        cta_coord[2],
    )
    # Coord inside cta
    tidx, _, _ = cute.arch.thread_idx()

    #
    # Define shared storage for kernel
    #
    @cute.struct
    class SharedStorage:
        ab_mbar_ptr: cute.struct.MemRange[cutlass.Int64, num_ab_stage * 2]
        acc_mbar_ptr: cute.struct.MemRange[cutlass.Int64, num_acc_stage * 2]
        tmem_holding_buf: cutlass.Int32

    smem = utils.SmemAllocator()
    storage = smem.allocate(SharedStorage)
    # (MMA, MMA_M, MMA_K, STAGE)
    sA = smem.allocate_tensor(
        element_type=ab_dtype,
        layout=a_smem_layout_staged.outer,
        byte_alignment=128,
        swizzle=a_smem_layout_staged.inner,
    )
    # (MMA, MMA_N, MMA_K, STAGE)
    sB = smem.allocate_tensor(
        element_type=ab_dtype,
        layout=b_smem_layout_staged.outer,
        byte_alignment=128,
        swizzle=b_smem_layout_staged.inner,
    )
    # (MMA, MMA_M, MMA_K, STAGE)
    sSFA = smem.allocate_tensor(
        element_type=sf_dtype,
        layout=sfa_smem_layout_staged,
        byte_alignment=128,
    )
    # (MMA, MMA_N, MMA_K, STAGE)
    sSFB = smem.allocate_tensor(
        element_type=sf_dtype,
        layout=sfb_smem_layout_staged,
        byte_alignment=128,
    )

    #
    # Initialize mainloop ab_pipeline, acc_pipeline and their states
    #
    # AB pipeline: TMA warp produces, MMA warp consumes
    ab_pipeline_producer_group = pipeline.CooperativeGroup(pipeline.Agent.Thread)
    ab_pipeline_consumer_group = pipeline.CooperativeGroup(pipeline.Agent.Thread, 1)
    ab_pipeline = pipeline.PipelineTmaUmma.create(
        barrier_storage=storage.ab_mbar_ptr.data_ptr(),
        num_stages=num_ab_stage,
        producer_group=ab_pipeline_producer_group,
        consumer_group=ab_pipeline_consumer_group,
        tx_count=num_tma_load_bytes,
    )

    # ACC pipeline: MMA warp produces, epilogue warps consume
    acc_pipeline_producer_group = pipeline.CooperativeGroup(pipeline.Agent.Thread)
    acc_pipeline_consumer_group = pipeline.CooperativeGroup(
        pipeline.Agent.Thread,
        len(epilog_warp_ids) * 32,  # 4 epilogue warps = 128 threads
    )
    acc_pipeline = pipeline.PipelineUmmaAsync.create(
        barrier_storage=storage.acc_mbar_ptr.data_ptr(),
        num_stages=num_acc_stage,
        producer_group=acc_pipeline_producer_group,
        consumer_group=acc_pipeline_consumer_group,
    )

    #
    # Named barriers for synchronization
    #
    epilog_sync_barrier = pipeline.NamedBarrier(
        barrier_id=1,
        num_threads=32 * len(epilog_warp_ids),  # 128 threads
    )
    tmem_alloc_barrier = pipeline.NamedBarrier(
        barrier_id=2,
        num_threads=32 * (len(epilog_warp_ids) + 1),  # MMA + epilogue = 160 threads
    )

    #
    # Local_tile partition global tensors
    #
    # (bM, bK, RestM, RestK, RestL)
    gA_mkl = cute.local_tile(
        mA_mkl, cute.slice_(mma_tiler_mnk, (None, 0, None)), (None, None, None)
    )
    # (bN, bK, RestN, RestK, RestL)
    gB_nkl = cute.local_tile(
        mB_nkl, cute.slice_(mma_tiler_mnk, (0, None, None)), (None, None, None)
    )
    gSFA_mkl = cute.local_tile(
        mSFA_mkl, cute.slice_(mma_tiler_mnk, (None, 0, None)), (None, None, None)
    )
    gSFB_nkl = cute.local_tile(
        mSFB_nkl, cute.slice_(mma_tiler_mnk, (0, None, None)), (None, None, None)
    )
    # (bM, bN, RestM, RestN, RestL)
    gC_mnl = cute.local_tile(
        mC_mnl, cute.slice_(mma_tiler_mnk, (None, None, 0)), (None, None, None)
    )
    k_tile_cnt = cute.size(gA_mkl, mode=[3])

    #
    # Partition global tensor for TiledMMA_A/B/SFA/SFB/C
    #
    # (MMA, MMA_M, MMA_K, RestK)
    thr_mma = tiled_mma.get_slice(0)
    # (MMA, MMA_M, MMA_K, RestM, RestK, RestL)
    tCgA = thr_mma.partition_A(gA_mkl)
    # (MMA, MMA_N, MMA_K, RestN, RestK, RestL)
    tCgB = thr_mma.partition_B(gB_nkl)
    # (MMA, MMA_M, MMA_K, RestM, RestK, RestL)
    tCgSFA = thr_mma.partition_A(gSFA_mkl)
    # (MMA, MMA_N, MMA_K, RestN, RestK, RestL)
    tCgSFB = thr_mma.partition_B(gSFB_nkl)
    # (MMA, MMA_M, MMA_N, RestM, RestN, RestL)
    tCgC = thr_mma.partition_C(gC_mnl)

    #
    # Partition global/shared tensor for TMA load A/B/SFA/SFB
    #
    # TMA Partition_S/D for A
    # ((atom_v, rest_v), STAGE)
    # ((atom_v, rest_v), RestM, RestK, RestL)
    tAsA, tAgA = cpasync.tma_partition(
        tma_atom_a,
        0,
        cute.make_layout(1),
        cute.group_modes(sA, 0, 3),
        cute.group_modes(tCgA, 0, 3),
    )
    # TMA Partition_S/D for B
    # ((atom_v, rest_v), STAGE)
    # ((atom_v, rest_v), RestN, RestK, RestL)
    tBsB, tBgB = cpasync.tma_partition(
        tma_atom_b,
        0,
        cute.make_layout(1),
        cute.group_modes(sB, 0, 3),
        cute.group_modes(tCgB, 0, 3),
    )
    #  TMA Partition_S/D for SFA
    # ((atom_v, rest_v), STAGE)
    # ((atom_v, rest_v), RestM, RestK, RestL)
    tAsSFA, tAgSFA = cpasync.tma_partition(
        tma_atom_sfa,
        0,
        cute.make_layout(1),
        cute.group_modes(sSFA, 0, 3),
        cute.group_modes(tCgSFA, 0, 3),
    )
    tAsSFA = cute.filter_zeros(tAsSFA)
    tAgSFA = cute.filter_zeros(tAgSFA)
    # TMA Partition_S/D for SFB
    # ((atom_v, rest_v), STAGE)
    # ((atom_v, rest_v), RestN, RestK, RestL)
    tBsSFB, tBgSFB = cpasync.tma_partition(
        tma_atom_sfb,
        0,
        cute.make_layout(1),
        cute.group_modes(sSFB, 0, 3),
        cute.group_modes(tCgSFB, 0, 3),
    )
    tBsSFB = cute.filter_zeros(tBsSFB)
    tBgSFB = cute.filter_zeros(tBgSFB)

    #
    # Partition shared/tensor memory tensor for TiledMMA_A/B/C
    #
    # (MMA, MMA_M, MMA_K, STAGE)
    tCrA = tiled_mma.make_fragment_A(sA)
    # (MMA, MMA_N, MMA_K, STAGE)
    tCrB = tiled_mma.make_fragment_B(sB)
    # (MMA, MMA_M, MMA_N)
    acc_shape = tiled_mma.partition_shape_C(mma_tiler_mnk[:2])
    # (MMA, MMA_M, MMA_N)
    tCtAcc_fake = tiled_mma.make_fragment_C(acc_shape)

    #
    # Slice to per mma tile index
    #
    # ((atom_v, rest_v), RestK)
    tAgA = tAgA[(None, mma_tile_coord_mnl[0], None, mma_tile_coord_mnl[2])]
    # ((atom_v, rest_v), RestK)
    tBgB = tBgB[(None, mma_tile_coord_mnl[1], None, mma_tile_coord_mnl[2])]
    # ((atom_v, rest_v), RestK)
    tAgSFA = tAgSFA[(None, mma_tile_coord_mnl[0], None, mma_tile_coord_mnl[2])]
    # ((atom_v, rest_v), RestK)
    tBgSFB = tBgSFB[(None, mma_tile_coord_mnl[1], None, mma_tile_coord_mnl[2])]

    #
    # TMA Warp (warp_idx == 5): Handles all TMA loads
    #
    if warp_idx == tma_warp_id:
        # Prefetch TMA descriptors
        cpasync.prefetch_descriptor(tma_atom_a)
        cpasync.prefetch_descriptor(tma_atom_b)
        cpasync.prefetch_descriptor(tma_atom_sfa)
        cpasync.prefetch_descriptor(tma_atom_sfb)

        ab_producer_state = pipeline.make_pipeline_state(
            pipeline.PipelineUserType.Producer, num_ab_stage
        )

        # Peek (try_wait) AB buffer empty for first k_tile
        ab_producer_state.reset_count()
        peek_ab_empty_status = cutlass.Boolean(1)
        if ab_producer_state.count < k_tile_cnt:
            peek_ab_empty_status = ab_pipeline.producer_try_acquire(
                ab_producer_state
            )

        # Main TMA loop - load all k tiles
        for k_tile in cutlass.range(0, k_tile_cnt, 1, unroll=1):
            # Conditionally wait for AB buffer empty
            ab_pipeline.producer_acquire(ab_producer_state, peek_ab_empty_status)

            # TMA load A/B/SFA/SFB to shared memory
            cute.copy(
                tma_atom_a,
                tAgA[(None, ab_producer_state.count)],
                tAsA[(None, ab_producer_state.index)],
                tma_bar_ptr=ab_pipeline.producer_get_barrier(ab_producer_state),
            )
            cute.copy(
                tma_atom_b,
                tBgB[(None, ab_producer_state.count)],
                tBsB[(None, ab_producer_state.index)],
                tma_bar_ptr=ab_pipeline.producer_get_barrier(ab_producer_state),
            )
            cute.copy(
                tma_atom_sfa,
                tAgSFA[(None, ab_producer_state.count)],
                tAsSFA[(None, ab_producer_state.index)],
                tma_bar_ptr=ab_pipeline.producer_get_barrier(ab_producer_state),
            )
            cute.copy(
                tma_atom_sfb,
                tBgSFB[(None, ab_producer_state.count)],
                tBsSFB[(None, ab_producer_state.index)],
                tma_bar_ptr=ab_pipeline.producer_get_barrier(ab_producer_state),
            )

            # Peek (try_wait) AB buffer empty for next k_tile
            ab_producer_state.advance()
            peek_ab_empty_status = cutlass.Boolean(1)
            if ab_producer_state.count < k_tile_cnt:
                peek_ab_empty_status = ab_pipeline.producer_try_acquire(
                    ab_producer_state
                )

        # Wait for all TMA loads to complete
        ab_pipeline.producer_tail(ab_producer_state)

    #
    # MMA Warp (warp_idx == 4): Handles S2T copy and MMA operations
    #
    if warp_idx == mma_warp_id:
        #
        # Alloc tensor memory buffer (done by epilogue warp 0)
        # MMA warp waits for allocation
        #
        tmem = utils.TmemAllocator(
            storage.tmem_holding_buf,
            barrier_for_retrieve=tmem_alloc_barrier,
        )
        tmem.wait_for_alloc()
        acc_tmem_ptr = tmem.retrieve_ptr(cutlass.Float32)
        tCtAcc = cute.make_tensor(acc_tmem_ptr, tCtAcc_fake.layout)

        #
        # Make SFA/SFB tmem tensor
        #
        # Get SFA tmem ptr
        sfa_tmem_ptr = cute.recast_ptr(
            acc_tmem_ptr + tcgen05.find_tmem_tensor_col_offset(tCtAcc),
            dtype=sf_dtype,
        )
        # (MMA, MMA_M, MMA_K)
        tCtSFA_layout = blockscaled_utils.make_tmem_layout_sfa(
            tiled_mma,
            mma_tiler_mnk,
            sf_vec_size,
            cute.slice_(sfa_smem_layout_staged, (None, None, None, 0)),
        )
        tCtSFA = cute.make_tensor(sfa_tmem_ptr, tCtSFA_layout)
        # Get SFB tmem ptr
        sfb_tmem_ptr = cute.recast_ptr(
            acc_tmem_ptr
            + tcgen05.find_tmem_tensor_col_offset(tCtAcc)
            + tcgen05.find_tmem_tensor_col_offset(tCtSFA),
            dtype=sf_dtype,
        )
        # (MMA, MMA_N, MMA_K)
        tCtSFB_layout = blockscaled_utils.make_tmem_layout_sfb(
            tiled_mma,
            mma_tiler_mnk,
            sf_vec_size,
            cute.slice_(sfb_smem_layout_staged, (None, None, None, 0)),
        )
        tCtSFB = cute.make_tensor(sfb_tmem_ptr, tCtSFB_layout)

        #
        # Partition for S2T copy of SFA/SFB
        #
        # Make S2T CopyAtom
        copy_atom_s2t = cute.make_copy_atom(
            tcgen05.Cp4x32x128bOp(tcgen05.CtaGroup.ONE),
            sf_dtype,
        )
        # (MMA, MMA_MN, MMA_K, STAGE)
        tCsSFA_compact = cute.filter_zeros(sSFA)
        # (MMA, MMA_MN, MMA_K)
        tCtSFA_compact = cute.filter_zeros(tCtSFA)
        tiled_copy_s2t_sfa = tcgen05.make_s2t_copy(copy_atom_s2t, tCtSFA_compact)
        thr_copy_s2t_sfa = tiled_copy_s2t_sfa.get_slice(0)
        # ((ATOM_V, REST_V), Rest_Tiler, MMA_MN, MMA_K, STAGE)
        tCsSFA_compact_s2t_ = thr_copy_s2t_sfa.partition_S(tCsSFA_compact)
        # ((ATOM_V, REST_V), Rest_Tiler, MMA_MN, MMA_K, STAGE)
        tCsSFA_compact_s2t = tcgen05.get_s2t_smem_desc_tensor(
            tiled_copy_s2t_sfa, tCsSFA_compact_s2t_
        )
        # ((ATOM_V, REST_V), Rest_Tiler, MMA_MN, MMA_K)
        tCtSFA_compact_s2t = thr_copy_s2t_sfa.partition_D(tCtSFA_compact)

        # (MMA, MMA_MN, MMA_K, STAGE)
        tCsSFB_compact = cute.filter_zeros(sSFB)
        # (MMA, MMA_MN, MMA_K)
        tCtSFB_compact = cute.filter_zeros(tCtSFB)
        tiled_copy_s2t_sfb = tcgen05.make_s2t_copy(copy_atom_s2t, tCtSFB_compact)
        thr_copy_s2t_sfb = tiled_copy_s2t_sfb.get_slice(0)
        # ((ATOM_V, REST_V), Rest_Tiler, MMA_MN, MMA_K, STAGE)
        tCsSFB_compact_s2t_ = thr_copy_s2t_sfb.partition_S(tCsSFB_compact)
        # ((ATOM_V, REST_V), Rest_Tiler, MMA_MN, MMA_K, STAGE)
        tCsSFB_compact_s2t = tcgen05.get_s2t_smem_desc_tensor(
            tiled_copy_s2t_sfb, tCsSFB_compact_s2t_
        )
        # ((ATOM_V, REST_V), Rest_Tiler, MMA_MN, MMA_K)
        tCtSFB_compact_s2t = thr_copy_s2t_sfb.partition_D(tCtSFB_compact)

        # Create pipeline states
        ab_consumer_state = pipeline.make_pipeline_state(
            pipeline.PipelineUserType.Consumer, num_ab_stage
        )
        acc_producer_state = pipeline.make_pipeline_state(
            pipeline.PipelineUserType.Producer, num_acc_stage
        )

        # Peek (try_wait) AB buffer full for first k_tile
        ab_consumer_state.reset_count()
        peek_ab_full_status = cutlass.Boolean(1)
        if ab_consumer_state.count < k_tile_cnt:
            peek_ab_full_status = ab_pipeline.consumer_try_wait(
                ab_consumer_state
            )

        # Acquire accumulator buffer
        acc_pipeline.producer_acquire(acc_producer_state)

        # Set ACCUMULATE field to False for the first k_tile iteration
        tiled_mma.set(tcgen05.Field.ACCUMULATE, False)

        # Execute k_tile loop
        for k_tile in cutlass.range(0, k_tile_cnt, 1, unroll=1):
            # Conditionally wait for AB buffer full
            ab_pipeline.consumer_wait(ab_consumer_state, peek_ab_full_status)

            # Copy SFA/SFB from shared memory to TMEM
            s2t_stage_coord = (None, None, None, None, ab_consumer_state.index)
            tCsSFA_compact_s2t_staged = tCsSFA_compact_s2t[s2t_stage_coord]
            tCsSFB_compact_s2t_staged = tCsSFB_compact_s2t[s2t_stage_coord]
            cute.copy(
                tiled_copy_s2t_sfa,
                tCsSFA_compact_s2t_staged,
                tCtSFA_compact_s2t,
            )
            cute.copy(
                tiled_copy_s2t_sfb,
                tCsSFB_compact_s2t_staged,
                tCtSFB_compact_s2t,
            )

            # tCtAcc += tCrA * tCrSFA * tCrB * tCrSFB
            num_kblocks = cute.size(tCrA, mode=[2])
            for kblock_idx in cutlass.range(num_kblocks, unroll_full=True):
                kblock_coord = (
                    None,
                    None,
                    kblock_idx,
                    ab_consumer_state.index,
                )

                # Set SFA/SFB tensor to tiled_mma
                sf_kblock_coord = (None, None, kblock_idx)
                tiled_mma.set(
                    tcgen05.Field.SFA,
                    tCtSFA[sf_kblock_coord].iterator,
                )
                tiled_mma.set(
                    tcgen05.Field.SFB,
                    tCtSFB[sf_kblock_coord].iterator,
                )

                cute.gemm(
                    tiled_mma,
                    tCtAcc,
                    tCrA[kblock_coord],
                    tCrB[kblock_coord],
                    tCtAcc,
                )
                # Enable accumulate on tCtAcc after first kblock
                tiled_mma.set(tcgen05.Field.ACCUMULATE, True)

            # Release AB buffer
            ab_pipeline.consumer_release(ab_consumer_state)

            # Peek (try_wait) AB buffer full for next k_tile
            ab_consumer_state.advance()
            peek_ab_full_status = cutlass.Boolean(1)
            if ab_consumer_state.count < k_tile_cnt:
                peek_ab_full_status = ab_pipeline.consumer_try_wait(
                    ab_consumer_state
                )

        # Commit accumulator buffer
        acc_pipeline.producer_commit(acc_producer_state)
        acc_pipeline.producer_tail(acc_producer_state)

    #
    # Epilogue Warps (warp_idx < 4): Handle T2R copy and GMEM stores
    #
    if warp_idx < mma_warp_id:
        #
        # Alloc tensor memory buffer
        #
        tmem = utils.TmemAllocator(
            storage.tmem_holding_buf,
            barrier_for_retrieve=tmem_alloc_barrier,
            allocator_warp_id=epilog_warp_ids[0],  # Only warp 0 allocates
        )
        tmem.allocate(num_tmem_alloc_cols)
        tmem.wait_for_alloc()
        acc_tmem_ptr = tmem.retrieve_ptr(cutlass.Float32)
        tCtAcc = cute.make_tensor(acc_tmem_ptr, tCtAcc_fake.layout)

        #
        # Partition for epilogue
        #
        op = tcgen05.Ld32x32bOp(tcgen05.Repetition.x128, tcgen05.Pack.NONE)
        copy_atom_t2r = cute.make_copy_atom(op, cutlass.Float32)
        tiled_copy_t2r = tcgen05.make_tmem_copy(copy_atom_t2r, tCtAcc)
        thr_copy_t2r = tiled_copy_t2r.get_slice(tidx)
        # (T2R_M, T2R_N, EPI_M, EPI_M)
        tTR_tAcc = thr_copy_t2r.partition_S(tCtAcc)
        # (T2R_M, T2R_N, EPI_M, EPI_N, RestM, RestN, RestL)
        tTR_gC = thr_copy_t2r.partition_D(tCgC)
        # (T2R_M, T2R_N, EPI_M, EPI_N）
        tTR_rAcc = cute.make_rmem_tensor(
            tTR_gC[None, None, None, None, 0, 0, 0].shape, cutlass.Float32
        )
        # (T2R_M, T2R_N, EPI_M, EPI_N）
        tTR_rC = cute.make_rmem_tensor(
            tTR_gC[None, None, None, None, 0, 0, 0].shape, c_dtype
        )
        # STG Atom
        simt_atom = cute.make_copy_atom(cute.nvgpu.CopyUniversalOp(), c_dtype)
        tTR_gC = tTR_gC[(None, None, None, None, *mma_tile_coord_mnl)]

        # Create pipeline state
        acc_consumer_state = pipeline.make_pipeline_state(
            pipeline.PipelineUserType.Consumer, num_acc_stage
        )

        # Wait for accumulator buffer full
        acc_pipeline.consumer_wait(acc_consumer_state)

        # Copy accumulator to register
        cute.copy(tiled_copy_t2r, tTR_tAcc, tTR_rAcc)
        acc_vec = tTR_rAcc.load().to(c_dtype)
        tTR_rC.store(acc_vec)
        # Store C to global memory
        cute.copy(simt_atom, tTR_rC, tTR_gC)

        # Release accumulator buffer
        acc_pipeline.consumer_release(acc_consumer_state)

        # Sync all epilogue warps before deallocating TMEM
        epilog_sync_barrier.arrive_and_wait()

        # Deallocate TMEM (only warp 0)
        if warp_idx == epilog_warp_ids[0]:
            tmem.free(acc_tmem_ptr)

    return


@cute.jit
def my_kernel(
    a_ptr: cute.Pointer,
    b_ptr: cute.Pointer,
    sfa_ptr: cute.Pointer,
    sfb_ptr: cute.Pointer,
    c_ptr: cute.Pointer,
    problem_size: tuple,
):
    """
    Host-side JIT function to prepare tensors and launch GPU kernel.
    """
    m, n, k, l = problem_size

    # Setup attributes that depend on gemm inputs
    a_tensor = cute.make_tensor(
        a_ptr,
        cute.make_layout(
            (m, cute.assume(k, 32), l),
            stride=(cute.assume(k, 32), 1, cute.assume(m * k, 32)),
        ),
    )
    b_tensor = cute.make_tensor(
        b_ptr,
        cute.make_layout(
            (n, cute.assume(k, 32), l),
            stride=(cute.assume(k, 32), 1, cute.assume(n * k, 32)),
        ),
    )
    c_tensor = cute.make_tensor(
        c_ptr, cute.make_layout((cute.assume(m, 32), n, l), stride=(n, 1, m * n))
    )
    # Setup sfa/sfb tensor by filling A/B tensor to scale factor atom layout
    # ((Atom_M, Rest_M),(Atom_K, Rest_K),RestL)
    sfa_layout = blockscaled_utils.tile_atom_to_shape_SF(a_tensor.shape, sf_vec_size)
    sfa_tensor = cute.make_tensor(sfa_ptr, sfa_layout)

    # ((Atom_N, Rest_N),(Atom_K, Rest_K),RestL)
    sfb_layout = blockscaled_utils.tile_atom_to_shape_SF(b_tensor.shape, sf_vec_size)
    sfb_tensor = cute.make_tensor(sfb_ptr, sfb_layout)

    mma_op = tcgen05.MmaMXF4NVF4Op(
        sf_dtype,
        (mma_tiler_mnk[0], mma_tiler_mnk[1], mma_inst_shape_k),
        tcgen05.CtaGroup.ONE,
        tcgen05.OperandSource.SMEM,
    )
    tiled_mma = cute.make_tiled_mma(mma_op)

    cluster_layout_vmnk = cute.tiled_divide(
        cute.make_layout((1, 1, 1)),
        (tiled_mma.thr_id.shape,),
    )

    # Compute A/B/SFA/SFB/C shared memory layout
    a_smem_layout_staged = sm100_utils.make_smem_layout_a(
        tiled_mma,
        mma_tiler_mnk,
        ab_dtype,
        num_ab_stage,
    )
    b_smem_layout_staged = sm100_utils.make_smem_layout_b(
        tiled_mma,
        mma_tiler_mnk,
        ab_dtype,
        num_ab_stage,
    )
    sfa_smem_layout_staged = blockscaled_utils.make_smem_layout_sfa(
        tiled_mma,
        mma_tiler_mnk,
        sf_vec_size,
        num_ab_stage,
    )
    sfb_smem_layout_staged = blockscaled_utils.make_smem_layout_sfb(
        tiled_mma,
        mma_tiler_mnk,
        sf_vec_size,
        num_ab_stage,
    )

    atom_thr_size = cute.size(tiled_mma.thr_id.shape)

    # Setup TMA for A
    a_smem_layout = cute.slice_(a_smem_layout_staged, (None, None, None, 0))
    tma_atom_a, tma_tensor_a = cute.nvgpu.make_tiled_tma_atom_A(
        cpasync.CopyBulkTensorTileG2SOp(tcgen05.CtaGroup.ONE),
        a_tensor,
        a_smem_layout,
        mma_tiler_mnk,
        tiled_mma,
        cluster_layout_vmnk.shape,
    )
    # Setup TMA for B
    b_smem_layout = cute.slice_(b_smem_layout_staged, (None, None, None, 0))
    tma_atom_b, tma_tensor_b = cute.nvgpu.make_tiled_tma_atom_B(
        cpasync.CopyBulkTensorTileG2SOp(tcgen05.CtaGroup.ONE),
        b_tensor,
        b_smem_layout,
        mma_tiler_mnk,
        tiled_mma,
        cluster_layout_vmnk.shape,
    )
    # Setup TMA for SFA
    sfa_smem_layout = cute.slice_(sfa_smem_layout_staged, (None, None, None, 0))
    tma_atom_sfa, tma_tensor_sfa = cute.nvgpu.make_tiled_tma_atom_A(
        cpasync.CopyBulkTensorTileG2SOp(tcgen05.CtaGroup.ONE),
        sfa_tensor,
        sfa_smem_layout,
        mma_tiler_mnk,
        tiled_mma,
        cluster_layout_vmnk.shape,
        internal_type=cutlass.Int16,
    )
    # Setup TMA for SFB
    sfb_smem_layout = cute.slice_(sfb_smem_layout_staged, (None, None, None, 0))
    tma_atom_sfb, tma_tensor_sfb = cute.nvgpu.make_tiled_tma_atom_B(
        cpasync.CopyBulkTensorTileG2SOp(tcgen05.CtaGroup.ONE),
        sfb_tensor,
        sfb_smem_layout,
        mma_tiler_mnk,
        tiled_mma,
        cluster_layout_vmnk.shape,
        internal_type=cutlass.Int16,
    )

    # Compute TMA load bytes
    a_copy_size = cute.size_in_bytes(ab_dtype, a_smem_layout)
    b_copy_size = cute.size_in_bytes(ab_dtype, b_smem_layout)
    sfa_copy_size = cute.size_in_bytes(sf_dtype, sfa_smem_layout)
    sfb_copy_size = cute.size_in_bytes(sf_dtype, sfb_smem_layout)
    num_tma_load_bytes = (
        a_copy_size + b_copy_size + sfa_copy_size + sfb_copy_size
    ) * atom_thr_size

    # Compute grid size
    grid = (
        cute.ceil_div(c_tensor.shape[0], mma_tiler_mnk[0]),
        cute.ceil_div(c_tensor.shape[1], mma_tiler_mnk[1]),
        c_tensor.shape[2],
    )

    # Launch the kernel
    kernel(
        # MMA (Matrix Multiply-Accumulate) configuration
        tiled_mma,  # Tiled MMA object defining NVFP4 GEMM compute pattern
        # TMA (Tensor Memory Accelerator) atoms and tensors for input matrix A
        tma_atom_a,  # TMA copy atom defining how to load A from global memory
        tma_tensor_a,  # Tensor descriptor for A matrix (m, k, l)
        # TMA atoms and tensors for input matrix B
        tma_atom_b,  # TMA copy atom defining how to load B from global memory
        tma_tensor_b,  # Tensor descriptor for B matrix (n, k, l)
        # TMA atoms and tensors for scale factor A
        tma_atom_sfa,  # TMA copy atom for loading scale factors for A
        tma_tensor_sfa,  # Tensor descriptor for SFA (block scale factors for A)
        # TMA atoms and tensors for scale factor B
        tma_atom_sfb,  # TMA copy atom for loading scale factors for B
        tma_tensor_sfb,  # Tensor descriptor for SFB (block scale factors for B)
        # Output tensor C
        c_tensor,  # Output tensor C where result will be stored (m, n, l)
        # Shared memory layouts with staging for pipelined execution
        a_smem_layout_staged,  # Staged shared memory layout for A (includes stage dimension)
        b_smem_layout_staged,  # Staged shared memory layout for B (includes stage dimension)
        sfa_smem_layout_staged,  # Staged shared memory layout for SFA (includes stage dimension)
        sfb_smem_layout_staged,  # Staged shared memory layout for SFB (includes stage dimension)
        # Pipeline synchronization parameter
        num_tma_load_bytes,  # Total bytes to load per TMA transaction (for barrier setup)
    ).launch(
        grid=grid,
        block=[threads_per_cta, 1, 1],
        cluster=(1, 1, 1),
    )
    return


# Global cache for compiled kernel
_compiled_kernel_cache = None


# This function is used to compile the kernel once and cache it and then allow users to
# run the kernel multiple times to get more accurate timing results.
def compile_kernel():
    """
    Compile the kernel once and cache it.
    This should be called before any timing measurements.

    Returns:
        The compiled kernel function
    """
    global _compiled_kernel_cache

    if _compiled_kernel_cache is not None:
        return _compiled_kernel_cache

    # Create CuTe pointers for A/B/C/SFA/SFB via torch tensor data pointer
    a_ptr = make_ptr(ab_dtype, 0, cute.AddressSpace.gmem, assumed_align=16)
    b_ptr = make_ptr(ab_dtype, 0, cute.AddressSpace.gmem, assumed_align=16)
    c_ptr = make_ptr(c_dtype, 0, cute.AddressSpace.gmem, assumed_align=16)
    sfa_ptr = make_ptr(sf_dtype, 0, cute.AddressSpace.gmem, assumed_align=32)
    sfb_ptr = make_ptr(sf_dtype, 0, cute.AddressSpace.gmem, assumed_align=32)

    # Compile the kernel
    _compiled_kernel_cache = cute.compile(
        my_kernel, a_ptr, b_ptr, sfa_ptr, sfb_ptr, c_ptr, (0, 0, 0, 0)
    )

    return _compiled_kernel_cache


def custom_kernel(data: input_t) -> output_t:
    """
    Execute the block-scaled GEMM kernel.

    This is the main entry point called by the evaluation framework.
    It converts PyTorch tensors to CuTe tensors, launches the kernel,
    and returns the result.

    Args:
        data: Tuple of (a, b, sfa_ref, sfb_ref, sfa_permuted, sfb_permuted, c) PyTorch tensors
            a: [m, k, l] - Input matrix in float4e2m1fn
            b: [n, k, l] - Input vector in float4e2m1fn
            sfa_ref: [m, k, l] - Scale factors in float8_e4m3fn, used by reference implementation
            sfb_ref: [n, k, l] - Scale factors in float8_e4m3fn, used by reference implementation
            sfa_permuted: [32, 4, rest_m, 4, rest_k, l] - Scale factors in float8_e4m3fn
            sfb_permuted: [32, 4, rest_n, 4, rest_k, l] - Scale factors in float8_e4m3fn
            c: [m, n, l] - Output vector in float16

    Returns:
        Output tensor c with computed results
    """
    a, b, _, _, sfa_permuted, sfb_permuted, c = data

    # Ensure kernel is compiled (will use cached version if available)
    # To avoid the compilation overhead, we compile the kernel once and cache it.
    compiled_func = compile_kernel()

    # Get dimensions from MxKxL layout
    m, k, l = a.shape
    n, _, _ = b.shape
    # Torch use e2m1_x2 data type, thus k is halved
    k = k * 2

    # Create CuTe pointers for A/B/C/SFA/SFB via torch tensor data pointer
    a_ptr = make_ptr(ab_dtype, a.data_ptr(), cute.AddressSpace.gmem, assumed_align=16)
    b_ptr = make_ptr(ab_dtype, b.data_ptr(), cute.AddressSpace.gmem, assumed_align=16)
    c_ptr = make_ptr(c_dtype, c.data_ptr(), cute.AddressSpace.gmem, assumed_align=16)
    sfa_ptr = make_ptr(
        sf_dtype, sfa_permuted.data_ptr(), cute.AddressSpace.gmem, assumed_align=32
    )
    sfb_ptr = make_ptr(
        sf_dtype, sfb_permuted.data_ptr(), cute.AddressSpace.gmem, assumed_align=32
    )

    # Execute the compiled kernel
    compiled_func(a_ptr, b_ptr, sfa_ptr, sfb_ptr, c_ptr, (m, n, k, l))

    return c