[Benchmark] Add benchmark scripts for block sparse attention (#114)

* Add DeepSeek MLA decode example with Flash Attention implementation

* Add GEMM SplitK and StreamK example implementations

This commit introduces two new example scripts demonstrating advanced GEMM (matrix multiplication) techniques:
- `example_tilelang_gemm_splitk.py`: Implements a Split-K GEMM kernel using TileLang
- `example_tilelang_gemm_streamk.py`: Implements a Stream-K GEMM kernel using TileLang

Both examples showcase different parallel computation strategies for matrix multiplication, with comprehensive testing using PyTorch reference implementations.

* Refactor GEMM SplitK and StreamK example implementations

Clean up and improve code formatting for the SplitK and StreamK GEMM example scripts:
- Remove unused import (Profiler) in splitk example
- Simplify line breaks and improve code readability
- Standardize indentation and remove unnecessary whitespace
- Optimize atomic add and copy operations for better clarity

* Add block sparse attention benchmarks for multiple libraries

This commit introduces comprehensive block sparse attention benchmarks for different libraries:
- TileLang block sparse FMHA implementation
- Triton block sparse FMHA implementation
- PyTorch reference block sparse FMHA implementation
- FlashAttention dense FMHA reference implementation

The benchmarks include:
- Configurable benchmark parameters (batch size, heads, sequence length, etc.)
- Sparse mask generation using top-k and threshold methods
- Performance measurement for different sparse attention configurations
- Utility functions for mask generation and benchmarking

* Refactor block sparse attention benchmarks with code style improvements

- Add Ruff linter ignore comments to benchmark files
- Improve code formatting and line breaks
- Remove unused imports
- Standardize print statement formatting
- Enhance code readability across multiple library benchmarks

* lint fix

benchmark/blocksparse_attention/benchmark_configs.py

+# BATCH, N_HEADS, SEQ_LEN, D_HEAD, TOPK, BLOCK
+configs = [[4, 2, 256, 64, 2, 64]]

benchmark/blocksparse_attention/benchmark_library_dense_fmha.py

+# Copyright (c) Microsoft Corporation.
+# Licensed under the MIT License.
+# ruff: noqa
+import torch
+from tilelang.profiler import do_bench
+
+
+def get_sparse_attn_mask_from_topk(x, topk, use_dense_for_last_block=False):
+    bsz, num_head, downsample_len, _ = x.shape
+    # N_CTX = downsample_len * BLOCK
+    sparse_index = torch.topk(x, topk, dim=-1).indices
+    dense_mask = torch.full([bsz, num_head, downsample_len, downsample_len],
+                            False,
+                            dtype=torch.bool,
+                            device=x.device)
+    dense_mask.scatter_(-1, sparse_index, True)
+    if use_dense_for_last_block:
+        dense_mask[:, :, -2:, :] = True
+    dense_mask.tril_()
+    return dense_mask
+
+
+def get_sparse_attn_mask_from_threshold(x, threshold, use_dense_for_last_block=False):
+    dense_mask = x > threshold
+    if use_dense_for_last_block:
+        dense_mask[:, :, -2:, :] = True
+    dense_mask.tril_()
+    return dense_mask
+
+
+def benchmark_topk_sparse_attention():
+    from benchmark_configs import configs
+    torch.manual_seed(0)
+
+    # Config
+    for BATCH, N_HEADS, SEQ_LEN, D_HEAD, TOPK, BLOCK in configs:
+
+        # Create inputs
+        q = torch.randn(BATCH, N_HEADS, SEQ_LEN, D_HEAD, device='cuda', dtype=torch.float16)
+        k = torch.randn(BATCH, N_HEADS, SEQ_LEN, D_HEAD, device='cuda', dtype=torch.float16)
+        v = torch.randn(BATCH, N_HEADS, SEQ_LEN, D_HEAD, device='cuda', dtype=torch.float16)
+
+        import flash_attn
+
+        def benchmark_fn():
+            flash_attn.flash_attn_func(q, k, v, causal=True)
+
+        ref_latency = do_bench(
+            benchmark_fn,
+            warmup=10,
+            rep=100,
+        )
+        print(
+            f"BATCH: {BATCH}, N_HEADS: {N_HEADS}, SEQ_LEN: {SEQ_LEN}, D_HEAD: {D_HEAD}, TOPK: {TOPK}, BLOCK: {BLOCK}, ref_latency: {ref_latency}"
+        )
+
+
+if __name__ == "__main__":
+    benchmark_topk_sparse_attention()

benchmark/blocksparse_attention/benchmark_tilelang_block_sparse_fmha.py

+# Copyright (c) Microsoft Corporation.
+# Licensed under the MIT License.
+# ruff: noqa
+import math
+import torch
+
+import tilelang
+from tilelang import language as T
+from tilelang.profiler import do_bench
+
+
+def is_hip():
+    return False
+
+
+def get_sparse_attn_mask_from_topk(x, topk, use_dense_for_last_block=False):
+    bsz, num_head, downsample_len, _ = x.shape
+    # N_CTX = downsample_len * BLOCK
+    sparse_index = torch.topk(x, topk, dim=-1).indices
+    dense_mask = torch.full([bsz, num_head, downsample_len, downsample_len],
+                            False,
+                            dtype=torch.bool,
+                            device=x.device)
+    dense_mask.scatter_(-1, sparse_index, True)
+    if use_dense_for_last_block:
+        dense_mask[:, :, -2:, :] = True
+    dense_mask.tril_()
+    return dense_mask
+
+
+def get_sparse_attn_mask_from_threshold(x, threshold, use_dense_for_last_block=False):
+    dense_mask = x > threshold
+    if use_dense_for_last_block:
+        dense_mask[:, :, -2:, :] = True
+    dense_mask.tril_()
+    return dense_mask
+
+
+def blocksparse_flashattn(batch, heads, seq_len, dim, downsample_len, is_causal):
+    block_M = 64
+    block_N = 64
+    num_stages = 0
+    threads = 128
+    scale = (1.0 / dim)**0.5 * 1.44269504  # log2(e)
+    shape = [batch, heads, seq_len, dim]
+    block_mask_shape = [batch, heads, downsample_len, downsample_len]
+
+    dtype = "float16"
+    accum_dtype = "float"
+    block_mask_dtype = "int8"
+
+    def kernel_func(block_M, block_N, num_stages, threads):
+
+        @T.macro
+        def MMA0(
+            K: T.Buffer(shape, dtype),
+            Q_shared: T.Buffer([block_M, dim], dtype),
+            K_shared: T.Buffer([block_N, dim], dtype),
+            acc_s: T.Buffer([block_M, block_N], accum_dtype),
+            k: T.int32,
+            bx: T.int32,
+            by: T.int32,
+            bz: T.int32,
+        ):
+            T.copy(K[bz, by, k * block_N:(k + 1) * block_N, :], K_shared)
+            if is_causal:
+                for i, j in T.Parallel(block_M, block_N):
+                    acc_s[i, j] = T.if_then_else(bx * block_M + i >= k * block_N + j, 0,
+                                                 -T.infinity(acc_s.dtype))
+            else:
+                T.clear(acc_s)
+            T.gemm(Q_shared, K_shared, acc_s, transpose_B=True, policy=T.GemmWarpPolicy.FullRow)
+
+        @T.macro
+        def MMA1(
+                V: T.Buffer(shape, dtype),
+                V_shared: T.Buffer([block_M, dim], dtype),
+                acc_s_cast: T.Buffer([block_M, block_N], dtype),
+                acc_o: T.Buffer([block_M, dim], accum_dtype),
+                k: T.int32,
+                by: T.int32,
+                bz: T.int32,
+        ):
+            T.copy(V[bz, by, k * block_N:(k + 1) * block_N, :], V_shared)
+            T.gemm(acc_s_cast, V_shared, acc_o, policy=T.GemmWarpPolicy.FullRow)
+
+        @T.macro
+        def Softmax(
+                acc_s: T.Buffer([block_M, block_N], accum_dtype),
+                acc_s_cast: T.Buffer([block_M, block_N], dtype),
+                scores_max: T.Buffer([block_M], accum_dtype),
+                scores_max_prev: T.Buffer([block_M], accum_dtype),
+                scores_scale: T.Buffer([block_M], accum_dtype),
+                scores_sum: T.Buffer([block_M], accum_dtype),
+                logsum: T.Buffer([block_M], accum_dtype),
+        ):
+            T.copy(scores_max, scores_max_prev)
+            T.fill(scores_max, -T.infinity(accum_dtype))
+            T.reduce_max(acc_s, scores_max, dim=1, clear=False)
+            # To do causal softmax, we need to set the scores_max to 0 if it is -inf
+            # This process is called Check_inf in FlashAttention3 code, and it only need to be done
+            # in the first ceil_div(kBlockM, kBlockN) steps.
+            # for i in T.Parallel(block_M):
+            #     scores_max[i] = T.if_then_else(scores_max[i] == -T.infinity(accum_dtype), 0, scores_max[i])
+            for i in T.Parallel(block_M):
+                scores_scale[i] = T.exp2(scores_max_prev[i] * scale - scores_max[i] * scale)
+            for i, j in T.Parallel(block_M, block_N):
+                # Instead of computing exp(x - max), we compute exp2(x * log_2(e) -
+                # max * log_2(e)) This allows the compiler to use the ffma
+                # instruction instead of fadd and fmul separately.
+                acc_s[i, j] = T.exp2(acc_s[i, j] * scale - scores_max[i] * scale)
+            T.reduce_sum(acc_s, scores_sum, dim=1)
+            for i in T.Parallel(block_M):
+                logsum[i] = logsum[i] * scores_scale[i] + scores_sum[i]
+            T.copy(acc_s, acc_s_cast)
+
+        @T.macro
+        def Rescale(
+                acc_o: T.Buffer([block_M, dim], accum_dtype),
+                scores_scale: T.Buffer([block_M], accum_dtype),
+        ):
+            for i, j in T.Parallel(block_M, dim):
+                acc_o[i, j] *= scores_scale[i]
+
+        @T.prim_func
+        def main(
+                Q: T.Buffer(shape, dtype),
+                K: T.Buffer(shape, dtype),
+                V: T.Buffer(shape, dtype),
+                BlockSparseMask: T.Buffer(block_mask_shape, block_mask_dtype),
+                Output: T.Buffer(shape, dtype),
+        ):
+            with T.Kernel(
+                    T.ceildiv(seq_len, block_M), heads, batch, threads=threads) as (bx, by, bz):
+                Q_shared = T.alloc_shared([block_M, dim], dtype)
+                K_shared = T.alloc_shared([block_N, dim], dtype)
+                V_shared = T.alloc_shared([block_N, dim], dtype)
+                O_shared = T.alloc_shared([block_M, dim], dtype)
+                acc_s = T.alloc_fragment([block_M, block_N], accum_dtype)
+                acc_s_cast = T.alloc_fragment([block_M, block_N], dtype)
+                acc_o = T.alloc_fragment([block_M, dim], accum_dtype)
+                scores_max = T.alloc_fragment([block_M], accum_dtype)
+                scores_max_prev = T.alloc_fragment([block_M], accum_dtype)
+                scores_scale = T.alloc_fragment([block_M], accum_dtype)
+                scores_sum = T.alloc_fragment([block_M], accum_dtype)
+                logsum = T.alloc_fragment([block_M], accum_dtype)
+                block_mask = T.alloc_local([downsample_len], block_mask_dtype)
+
+                T.copy(Q[bz, by, bx * block_M:(bx + 1) * block_M, :], Q_shared)
+                T.fill(acc_o, 0)
+                T.fill(logsum, 0)
+                T.fill(scores_max, -T.infinity(accum_dtype))
+
+                for vj in T.serial(downsample_len):
+                    block_mask[vj] = BlockSparseMask[bz, by, bx, vj]
+
+                loop_range = (
+                    T.min(T.ceildiv(seq_len, block_N), T.ceildiv(
+                        (bx + 1) * block_M, block_N)) if is_causal else T.ceildiv(seq_len, block_N))
+
+                for k in T.Pipelined(loop_range, num_stages=num_stages):
+                    if block_mask[k] != 0:
+                        MMA0(K, Q_shared, K_shared, acc_s, k, bx, by, bz)
+                        Softmax(acc_s, acc_s_cast, scores_max, scores_max_prev, scores_scale,
+                                scores_sum, logsum)
+                        Rescale(acc_o, scores_scale)
+                        MMA1(V, V_shared, acc_s_cast, acc_o, k, by, bz)
+                for i, j in T.Parallel(block_M, dim):
+                    acc_o[i, j] /= logsum[i]
+                T.copy(acc_o, O_shared)
+                T.copy(O_shared, Output[bz, by, bx * block_M:(bx + 1) * block_M, :])
+
+        return main
+
+    return kernel_func(block_M, block_N, num_stages, threads)
+
+
+def benchmark_topk_sparse_attention():
+    from benchmark_configs import configs
+    torch.manual_seed(0)
+
+    # Config
+    for BATCH, N_HEADS, SEQ_LEN, D_HEAD, TOPK, BLOCK in configs:
+
+        # Create inputs
+        q = torch.randn(BATCH, N_HEADS, SEQ_LEN, D_HEAD, device='cuda', dtype=torch.float16)
+        k = torch.randn(BATCH, N_HEADS, SEQ_LEN, D_HEAD, device='cuda', dtype=torch.float16)
+        v = torch.randn(BATCH, N_HEADS, SEQ_LEN, D_HEAD, device='cuda', dtype=torch.float16)
+
+        sm_scale = 1.0 / (D_HEAD**0.5)
+
+        # Create sparse mask (downsampled to block level)
+        downsample_factor = BLOCK
+        downsample_len = math.ceil(SEQ_LEN / downsample_factor)
+        x_ds = torch.randn([BATCH, N_HEADS, downsample_len, downsample_len],
+                           device='cuda',
+                           dtype=torch.bfloat16)
+        x_ds[:, :, :, 0] = 100
+        block_mask = get_sparse_attn_mask_from_topk(x_ds, topk=TOPK)
+        program = blocksparse_flashattn(
+            BATCH, N_HEADS, SEQ_LEN, D_HEAD, downsample_len, is_causal=True)
+        kernel = tilelang.compile(program, out_idx=4)
+
+        def benchmark_fn():
+            # Compute reference
+            # Expand block mask to full attention matrix
+            kernel(q, k, v, block_mask)
+
+        ref_latency = do_bench(
+            benchmark_fn,
+            warmup=10,
+            rep=100,
+        )
+        print(
+            f"BATCH: {BATCH}, N_HEADS: {N_HEADS}, SEQ_LEN: {SEQ_LEN}, D_HEAD: {D_HEAD}, TOPK: {TOPK}, BLOCK: {BLOCK}, ref_latency: {ref_latency}"
+        )
+
+
+if __name__ == "__main__":
+    benchmark_topk_sparse_attention()

benchmark/blocksparse_attention/benchmark_torch_block_sparse_fmha.py

+# Copyright (c) Microsoft Corporation.
+# Licensed under the MIT License.
+# ruff: noqa
+import math
+import torch
+
+import torch.nn.functional as F
+from tilelang.profiler import do_bench
+
+
+def get_sparse_attn_mask_from_topk(x, topk, use_dense_for_last_block=False):
+    bsz, num_head, downsample_len, _ = x.shape
+    # N_CTX = downsample_len * BLOCK
+    sparse_index = torch.topk(x, topk, dim=-1).indices
+    dense_mask = torch.full([bsz, num_head, downsample_len, downsample_len],
+                            False,
+                            dtype=torch.bool,
+                            device=x.device)
+    dense_mask.scatter_(-1, sparse_index, True)
+    if use_dense_for_last_block:
+        dense_mask[:, :, -2:, :] = True
+    dense_mask.tril_()
+    return dense_mask
+
+
+def get_sparse_attn_mask_from_threshold(x, threshold, use_dense_for_last_block=False):
+    dense_mask = x > threshold
+    if use_dense_for_last_block:
+        dense_mask[:, :, -2:, :] = True
+    dense_mask.tril_()
+    return dense_mask
+
+
+def benchmark_topk_sparse_attention():
+    from benchmark_configs import configs
+    torch.manual_seed(0)
+
+    # Config
+    for BATCH, N_HEADS, SEQ_LEN, D_HEAD, TOPK, BLOCK in configs:
+
+        # Create inputs
+        q = torch.randn(BATCH, N_HEADS, SEQ_LEN, D_HEAD, device='cuda', dtype=torch.float16)
+        k = torch.randn(BATCH, N_HEADS, SEQ_LEN, D_HEAD, device='cuda', dtype=torch.float16)
+        v = torch.randn(BATCH, N_HEADS, SEQ_LEN, D_HEAD, device='cuda', dtype=torch.float16)
+
+        sm_scale = 1.0 / (D_HEAD**0.5)
+
+        # Create sparse mask (downsampled to block level)
+        downsample_factor = BLOCK
+        downsample_len = math.ceil(SEQ_LEN / downsample_factor)
+        x_ds = torch.randn([BATCH, N_HEADS, downsample_len, downsample_len],
+                           device='cuda',
+                           dtype=torch.bfloat16)
+        x_ds[:, :, :, 0] = 100
+        block_mask = get_sparse_attn_mask_from_topk(x_ds, topk=TOPK)
+
+        def benchmark_fn():
+            # Compute reference
+            # Expand block mask to full attention matrix
+            full_mask = torch.kron(block_mask.float(), torch.ones(BLOCK, BLOCK, device='cuda'))
+            full_mask = full_mask[..., :SEQ_LEN, :SEQ_LEN].bool()
+            full_mask = full_mask & torch.tril(torch.ones_like(full_mask))  # Apply causal
+
+            # PyTorch reference implementation
+            attn = torch.einsum('bhsd,bhtd->bhst', q, k) * sm_scale
+            attn = attn.masked_fill(~full_mask, float('-inf'))
+            attn = F.softmax(attn, dim=-1)
+            ref_output = torch.einsum('bhst,bhtd->bhsd', attn, v)
+            return ref_output
+
+        ref_latency = do_bench(
+            benchmark_fn,
+            warmup=10,
+            rep=100,
+        )
+        print(
+            f"BATCH: {BATCH}, N_HEADS: {N_HEADS}, SEQ_LEN: {SEQ_LEN}, D_HEAD: {D_HEAD}, TOPK: {TOPK}, BLOCK: {BLOCK}, ref_latency: {ref_latency}"
+        )
+
+
+if __name__ == "__main__":
+    benchmark_topk_sparse_attention()

benchmark/blocksparse_attention/benchmark_triton_block_sparse_fmha.py

+# Copyright (c) Microsoft Corporation.
+# Licensed under the MIT License.
+# ruff: noqa
+import math
+import torch
+
+import triton
+import triton.language as tl
+from tilelang.profiler import do_bench
+
+
+def is_hip():
+    return False
+
+
+def get_sparse_attn_mask_from_topk(x, topk, use_dense_for_last_block=False):
+    bsz, num_head, downsample_len, _ = x.shape
+    # N_CTX = downsample_len * BLOCK
+    sparse_index = torch.topk(x, topk, dim=-1).indices
+    dense_mask = torch.full([bsz, num_head, downsample_len, downsample_len],
+                            False,
+                            dtype=torch.bool,
+                            device=x.device)
+    dense_mask.scatter_(-1, sparse_index, True)
+    if use_dense_for_last_block:
+        dense_mask[:, :, -2:, :] = True
+    dense_mask.tril_()
+    return dense_mask
+
+
+def get_sparse_attn_mask_from_threshold(x, threshold, use_dense_for_last_block=False):
+    dense_mask = x > threshold
+    if use_dense_for_last_block:
+        dense_mask[:, :, -2:, :] = True
+    dense_mask.tril_()
+    return dense_mask
+
+
+@triton.jit
+def _fwd_kernel_inner(
+    acc,
+    l_i,
+    m_i,
+    q,
+    k_block_col_idx,
+    block_mask_ptr,
+    k_ptrs,
+    v_ptrs,
+    offs_m,
+    offs_n,
+    stride_kt,
+    stride_vt,
+    stride_bmask_n,
+    sm_scale,
+    seqlen_k,
+    past_len,
+    LAST_K_BLOCK: tl.constexpr,
+    BLOCK_M: tl.constexpr,
+    BLOCK_N: tl.constexpr,
+):
+
+    mask_val = tl.load(block_mask_ptr + k_block_col_idx * stride_bmask_n)
+
+    if mask_val == True:
+        start_n = k_block_col_idx * BLOCK_N
+        # -- compute qk ----
+
+        k = tl.load(k_ptrs + start_n * stride_kt)
+
+        qk = tl.zeros([BLOCK_M, BLOCK_N], dtype=tl.float32)
+        qk += tl.dot(q, k)
+
+        qk *= sm_scale
+
+        # the following is needed only when LAST_K_BLOCK or BLOCK_M < BLOCK_N
+        if LAST_K_BLOCK:
+            qk += tl.where(offs_m[:, None] + past_len >= (start_n + offs_n[None, :]), 0,
+                           float('-inf'))
+
+        m_ij = tl.maximum(m_i, tl.max(qk, 1))
+        qk -= m_ij[:, None]
+        p = tl.exp(qk)
+        l_ij = tl.sum(p, 1)
+        alpha = tl.exp(m_i - m_ij)
+        l_i = l_i * alpha + l_ij
+        acc = acc * alpha[:, None]
+
+        # update acc
+        v = tl.load(v_ptrs + start_n * stride_vt)
+
+        p = p.to(v.type.element_ty)
+
+        acc += tl.dot(p, v)
+        # update m_i and l_i
+        m_i = m_ij
+    return acc, l_i, m_i
+
+
+@triton.jit
+def _fwd_kernel(
+    Q,
+    K,
+    V,
+    sm_scale,
+    block_mask_ptr,
+    Out,
+    stride_qz,
+    stride_qh,
+    stride_qm,
+    stride_qd,
+    stride_kz,
+    stride_kh,
+    stride_kn,
+    stride_kd,
+    stride_vz,
+    stride_vh,
+    stride_vn,
+    stride_vd,
+    stride_bmz,
+    stride_bmh,
+    stride_bmm,
+    stride_bmn,
+    stride_oz,
+    stride_oh,
+    stride_om,
+    stride_od,
+    H,
+    N_CTX,
+    PAST_LEN,
+    BLOCK_M: tl.constexpr,
+    BLOCK_N: tl.constexpr,
+    BLOCK_DMODEL: tl.constexpr,
+):
+    Q_LEN = N_CTX - PAST_LEN
+    start_m = tl.program_id(0)
+    off_hz = tl.program_id(1)
+    off_h = off_hz % H
+    off_z = off_hz // H
+    Q += off_z * stride_qz + off_h * stride_qh
+    K += off_z * stride_kz + off_h * stride_kh
+    V += off_z * stride_vz + off_h * stride_vh
+    block_mask_ptr += off_z * stride_bmz + off_h * stride_bmh
+
+    # initialize offsets
+    offs_m = start_m * BLOCK_M + tl.arange(0, BLOCK_M)
+    offs_n = tl.arange(0, BLOCK_N)
+    offs_d = tl.arange(0, BLOCK_DMODEL)
+    off_q = offs_m[:, None] * stride_qm + offs_d[None, :] * stride_qd
+    # off_k = offs_n[:, None] * stride_kn + offs_d[None, :] * stride_kd
+    off_k = offs_n[None, :] * stride_kn + offs_d[:, None] * stride_kd
+    off_v = offs_n[:, None] * stride_vn + offs_d[None, :] * stride_vd
+    # Initialize pointers to Q, K, V
+    q_ptrs = Q + off_q
+    k_ptrs = K + off_k
+    v_ptrs = V + off_v
+    mask_ptrs = block_mask_ptr + start_m * stride_bmm
+
+    m_i = tl.zeros([BLOCK_M], dtype=tl.float32) - float('inf')
+    l_i = tl.zeros([BLOCK_M], dtype=tl.float32)
+    acc = tl.zeros([BLOCK_M, BLOCK_DMODEL], dtype=tl.float32)
+
+    q = tl.load(q_ptrs, mask=offs_m[:, None] < Q_LEN)
+
+    k_block_start = 0
+    k_block_end = tl.cdiv((start_m + 1) * BLOCK_M, BLOCK_N)
+
+    # loop over k, v and update accumulator
+    for col_idx in range(k_block_start, k_block_end):
+        acc, l_i, m_i = _fwd_kernel_inner(
+            acc,
+            l_i,
+            m_i,
+            q,
+            col_idx,
+            mask_ptrs,
+            k_ptrs,
+            v_ptrs,
+            offs_m,
+            offs_n,
+            stride_kn,
+            stride_vn,
+            stride_bmn,
+            sm_scale,
+            N_CTX,
+            PAST_LEN,
+            col_idx == k_block_end - 1,
+            BLOCK_M,
+            BLOCK_N,
+        )
+
+    m_i += tl.math.log(l_i)
+    l_recip = 1 / l_i[:, None]
+    acc = acc * l_recip
+    acc = acc.to(Out.dtype.element_ty)
+
+    off_o = off_z * stride_oz + off_h * stride_oh + offs_m[:, None] * stride_om + offs_d[
+        None, :] * stride_od
+    out_ptrs = Out + off_o
+    tl.store(out_ptrs, acc, mask=offs_m[:, None] < N_CTX)
+
+
+def _forward(ctx,
+             q,
+             k,
+             v,
+             block_sparse_mask,
+             sm_scale,
+             BLOCK_M=64,
+             BLOCK_N=64,
+             num_warps=None,
+             num_stages=1,
+             out=None):
+
+    assert q.shape[-1] == k.shape[-1] == v.shape[-1]
+    assert k.shape[2] == v.shape[2]
+    o = out if out is not None else torch.empty_like(q).contiguous()
+    grid = (triton.cdiv(q.shape[2], BLOCK_M), q.shape[0] * q.shape[1])
+
+    assert q.shape[-1] in [64, 128]
+    BLOCK_DMODEL = q.shape[-1]
+
+    if is_hip():
+        num_warps, num_stages = 8, 1
+    else:
+        num_warps, num_stages = 4, 2
+
+    N_CTX = k.shape[2]
+    PAST_LEN = N_CTX - q.shape[2]
+
+    H = q.shape[1]
+
+    _fwd_kernel[grid](
+        q,
+        k,
+        v,
+        sm_scale,
+        block_sparse_mask,
+        o,
+        *q.stride(),
+        *k.stride(),
+        *v.stride(),
+        *block_sparse_mask.stride(),
+        *o.stride(),
+        H,
+        N_CTX,
+        PAST_LEN,
+        BLOCK_M,
+        BLOCK_N,
+        BLOCK_DMODEL,
+        num_warps=num_warps,
+        num_stages=num_stages,
+    )
+
+    return o
+
+
+class _sparse_attention(torch.autograd.Function):
+
+    @staticmethod
+    def forward(ctx, q, k, v, block_sparse_dense, sm_scale):
+        # shape constraints
+        return _forward(ctx, q, k, v, block_sparse_dense, sm_scale)
+
+    @staticmethod
+    def backward(ctx, do):
+        # No gradient propagation.
+        raise NotImplementedError("It does not support gradient propagation yet")
+        return None, None, None, None, None
+
+
+block_sparse_triton_fn = _sparse_attention.apply
+
+
+def benchmark_topk_sparse_attention():
+    from benchmark_configs import configs
+    torch.manual_seed(0)
+
+    # Config
+    for BATCH, N_HEADS, SEQ_LEN, D_HEAD, TOPK, BLOCK in configs:
+
+        # Create inputs
+        q = torch.randn(BATCH, N_HEADS, SEQ_LEN, D_HEAD, device='cuda', dtype=torch.float16)
+        k = torch.randn(BATCH, N_HEADS, SEQ_LEN, D_HEAD, device='cuda', dtype=torch.float16)
+        v = torch.randn(BATCH, N_HEADS, SEQ_LEN, D_HEAD, device='cuda', dtype=torch.float16)
+
+        sm_scale = 1.0 / (D_HEAD**0.5)
+
+        # Create sparse mask (downsampled to block level)
+        downsample_factor = BLOCK
+        downsample_len = math.ceil(SEQ_LEN / downsample_factor)
+        x_ds = torch.randn([BATCH, N_HEADS, downsample_len, downsample_len],
+                           device='cuda',
+                           dtype=torch.bfloat16)
+        x_ds[:, :, :, 0] = 100
+        block_mask = get_sparse_attn_mask_from_topk(x_ds, topk=TOPK)
+
+        def benchmark_fn():
+            # Compute reference
+            # Expand block mask to full attention matrix
+            block_sparse_triton_fn(q, k, v, block_mask, sm_scale)  # noqa: B023
+
+        ref_latency = do_bench(
+            benchmark_fn,
+            warmup=10,
+            rep=100,
+        )
+        print(
+            f"BATCH: {BATCH}, N_HEADS: {N_HEADS}, SEQ_LEN: {SEQ_LEN}, D_HEAD: {D_HEAD}, TOPK: {TOPK}, BLOCK: {BLOCK}, ref_latency: {ref_latency}"
+        )
+
+
+if __name__ == "__main__":
+    benchmark_topk_sparse_attention()

benchmark/blocksparse_attention/requirements.txt

+flash-attn