import math import einops import pytest import torch import torch.nn.functional as F from einops import rearrange, repeat from flash_attn_interface import ( _flash_attn_forward, flash_attn_func, flash_attn_varlen_func, ) from tests.test_util import ( attention_ref, construct_local_mask, generate_qkv, generate_random_padding_mask, ) ABS_TOL = 5e-3 REL_TOL = 1e-1 def print_diffs(out, out_ref): out_1d = out.flatten() out_ref_1d = out_ref.flatten() for idx, (e_o, e_o_ref) in enumerate(zip(out_1d, out_ref_1d)): diff = e_o - e_o_ref abs_diff = abs(diff) abs_ref = abs(e_o_ref + 1e-5) relative_diff = abs_diff / abs_ref if abs_diff > ABS_TOL or relative_diff > REL_TOL: print(f"==== diff ==== {idx}, test: {e_o}, ref: {e_o_ref}") @pytest.mark.parametrize("dtype", [torch.float8_e4m3fn]) @pytest.mark.parametrize("mha_type", ["mha", "mqa", "gqa"]) @pytest.mark.parametrize("causal", [False, True]) @pytest.mark.parametrize("local", [False, True]) @pytest.mark.parametrize("deterministic", [True]) @pytest.mark.parametrize("gqa_parallel", [False, True]) @pytest.mark.parametrize("d", [64, 128, 256]) # @pytest.mark.parametrize("descale", [1.0]) @pytest.mark.parametrize("descale", [1.0, 2.0, 3.0]) @pytest.mark.parametrize( "seqlen_q,seqlen_k", [ (1, 1), (64, 128), (128, 128), (256, 256), (113, 203), (128, 217), (113, 211), (108, 256), (256, 512), (384, 256), (640, 128), (512, 256), (1024, 1024), (1023, 1024), (1024, 1023), (4096, 4096), (4224, 4224), ], ) def test_flash_attn_output_fp8( seqlen_q, seqlen_k, d, causal, local, deterministic, mha_type, dtype, descale, gqa_parallel, ): device = "cuda" dtype_init = torch.bfloat16 print(dtype) print("causal", causal) print("local", local) print("gqa_parallel", gqa_parallel) # set seed torch.random.manual_seed(42) # batch_size = 40 # nheads = 16 batch_size = 4 nheads = 6 nheads_kv = 6 if mha_type == "mha" else (2 if mha_type == "gqa" else 1) # nheads_kv = 1 # batch_size = 9 # nheads = 6 window_size = (-1, -1) if not local else torch.randint(0, seqlen_k, (2,)) q = torch.randn( batch_size, seqlen_q, nheads, d, device=device, dtype=dtype_init, requires_grad=True, ) k = torch.randn( batch_size, seqlen_k, nheads_kv, d, device=device, dtype=dtype_init, requires_grad=True, ) v = torch.randn( batch_size, seqlen_k, nheads_kv, d, device=device, dtype=dtype_init, requires_grad=True, ) q = q.to(dtype) k = k.to(dtype) v = v.to(dtype) softmax_scale = q.shape[-1] ** (-0.5) descale_q = torch.tensor([descale], dtype=torch.float32, device="cuda") descale_k = torch.tensor([descale], dtype=torch.float32, device="cuda") descale_v = torch.tensor([descale], dtype=torch.float32, device="cuda") out, lse = flash_attn_func( q, k, v, causal=causal, window_size=window_size, deterministic=deterministic, gqa_parallel=gqa_parallel, descale_q=descale_q, descale_k=descale_k, descale_v=descale_v, ) q = q.to(dtype_init) k = k.to(dtype_init) v = v.to(dtype_init) descale_q = descale_q.to(dtype_init) descale_k = descale_k.to(dtype_init) descale_v = descale_v.to(dtype_init) q = q * descale_q k = k * descale_k v = v * descale_v out_ref, attn_ref = attention_ref( q, k, v, None, None, causal=causal, window_size=window_size, ) out_pt, attn_pt = attention_ref( q, k, v, None, None, causal=causal, window_size=window_size, upcast=False, reorder_ops=True, ) # qk = torch.einsum('bshd,bthd->bhst', q, k).float() # m = qk.amax(-1, keepdim=True) # s_tmp = torch.exp((qk - m) / math.sqrt(d)) # exp_sum = s_tmp.sum(-1) # qk = torch.einsum('bthd,bshd->bhts', q.float() / math.sqrt(d), k.float()) # lse_ref = torch.logsumexp(qk, dim=-1) print(f"Output max diff: {(out - out_ref).abs().max().item()}") print(f"Output mean diff: {(out - out_ref).abs().mean().item()}") print(f"Pytorch max diff: {(out_pt - out_ref).abs().max().item()}") print(f"Pytorch mean diff: {(out_pt - out_ref).abs().mean().item()}") # if not causal: # print(f"LSE max diff: {(lse - lse_ref).abs().max().item()}") # breakpoint() # dS = torch.einsum('bthd,bshd->bhts', g.float(), v.float()) # P = torch.softmax(qk, -1) # dP = P * (dS - do_o.unsqueeze(1)) # dQ = torch.einsum('bhts,bshd->bthd', dP, k.float()) # dV = torch.einsum('bhts,bthd->bshd', P, g.float()) # dK = torch.einsum('bhts,bthd->bshd', dP, q.float()) # breakpoint() # assert (out - out_ref).abs().max().item() <= 4 * (out_pt - out_ref).abs().max().item() + 1e-2 atol = 4 * (out_pt - out_ref).abs().max().item() + 1e-2 torch.testing.assert_close(out, out_ref, rtol=1e-2, atol=atol, check_dtype=False) @pytest.mark.parametrize("dtype", [torch.float16, torch.bfloat16]) # @pytest.mark.parametrize("dtype", [torch.float8_e4m3fn]) @pytest.mark.parametrize("mha_type", ["mha", "mqa", "gqa"]) # @pytest.mark.parametrize("mha_type", ["mha"]) @pytest.mark.parametrize("causal", [False, True]) # @pytest.mark.parametrize("causal", [False]) @pytest.mark.parametrize("local", [False, True]) # @pytest.mark.parametrize("local", [True]) @pytest.mark.parametrize("deterministic", [False, True]) # @pytest.mark.parametrize("deterministic", [True]) @pytest.mark.parametrize("gqa_parallel", [False, True]) # @pytest.mark.parametrize("gqa_parallel", [False]) # @pytest.mark.parametrize("d", [32, 64, 96, 128, 160, 192, 224, 256]) # @pytest.mark.parametrize('d', [32, 40, 64, 80, 96, 128, 160, 192]) # @pytest.mark.parametrize('d', [32, 64, 96, 128, 160, 192]) # @pytest.mark.parametrize('d', [56, 80]) # @pytest.mark.parametrize("d", [64, 128, 256]) # @pytest.mark.parametrize("d", [64, 96, 128]) # @pytest.mark.parametrize("d", [64]) @pytest.mark.parametrize("d", [64, 128, 256]) @pytest.mark.parametrize("descale", [1.0]) # @pytest.mark.parametrize("descale", [1.0, 2.0, 3.0, 4.0]) @pytest.mark.parametrize( "seqlen_q,seqlen_k", [ (1, 1), (64, 128), (128, 128), (256, 256), (113, 203), (128, 217), (113, 211), (108, 256), (256, 512), (384, 256), (640, 128), (512, 256), (1024, 1024), (1023, 1024), (1024, 1023), (4096, 4096), (4224, 4224), ], ) # @pytest.mark.parametrize('seqlen_q,seqlen_k', [(128, 128)]) def test_flash_attn_output( seqlen_q, seqlen_k, d, causal, local, deterministic, mha_type, dtype, descale, gqa_parallel, ): device = "cuda" if dtype == torch.float8_e4m3fn: dtype_init = torch.bfloat16 else: dtype_init = dtype print(dtype) print("causal", causal) print("local", local) print("gqa_parallel", gqa_parallel) # set seed torch.random.manual_seed(42) # batch_size = 40 # nheads = 16 batch_size = 4 nheads = 6 nheads_kv = 6 if mha_type == "mha" else (2 if mha_type == "gqa" else 1) # nheads_kv = 1 # batch_size = 9 # nheads = 6 window_size = (-1, -1) if not local else torch.randint(0, seqlen_k, (2,)) q = torch.randn( batch_size, seqlen_q, nheads, d, device=device, dtype=dtype_init, requires_grad=True, ) k = torch.randn( batch_size, seqlen_k, nheads_kv, d, device=device, dtype=dtype_init, requires_grad=True, ) v = torch.randn( batch_size, seqlen_k, nheads_kv, d, device=device, dtype=dtype_init, requires_grad=True, ) q = q.to(dtype) k = k.to(dtype) v = v.to(dtype) softmax_scale = q.shape[-1] ** (-0.5) descale_q = torch.tensor([descale], dtype=torch.float32, device="cuda") descale_k = torch.tensor([descale], dtype=torch.float32, device="cuda") descale_v = torch.tensor([descale], dtype=torch.float32, device="cuda") if dtype != torch.float8_e4m3fn: out, lse = flash_attn_func( q, k, v, causal=causal, window_size=window_size, deterministic=deterministic, gqa_parallel=gqa_parallel, ) else: out, lse = flash_attn_func( q, k, v, causal=causal, window_size=window_size, deterministic=deterministic, gqa_parallel=gqa_parallel, descale_q=descale_q, descale_k=descale_k, descale_v=descale_v, ) q = q.to(dtype_init) k = k.to(dtype_init) v = v.to(dtype_init) if dtype == torch.float8_e4m3fn: descale_q = descale_q.to(dtype_init) descale_k = descale_k.to(dtype_init) descale_v = descale_v.to(dtype_init) q = q * descale_q k = k * descale_k v = v * descale_v out_ref, attn_ref = attention_ref( q, k, v, None, None, causal=causal, window_size=window_size, ) out_pt, attn_pt = attention_ref( q, k, v, None, None, causal=causal, window_size=window_size, upcast=False, reorder_ops=True, ) # qk = torch.einsum('bshd,bthd->bhst', q, k).float() # m = qk.amax(-1, keepdim=True) # s_tmp = torch.exp((qk - m) / math.sqrt(d)) # exp_sum = s_tmp.sum(-1) # qk = torch.einsum('bthd,bshd->bhts', q.float() / math.sqrt(d), k.float()) # lse_ref = torch.logsumexp(qk, dim=-1) print(f"Output max diff: {(out - out_ref).abs().max().item()}") print(f"Output mean diff: {(out - out_ref).abs().mean().item()}") print(f"Pytorch max diff: {(out_pt - out_ref).abs().max().item()}") print(f"Pytorch mean diff: {(out_pt - out_ref).abs().mean().item()}") # if not causal: # print(f"LSE max diff: {(lse - lse_ref).abs().max().item()}") # breakpoint() if d <= 128 and dtype != torch.float8_e4m3fn: g = torch.randn_like(out) do_o = (g.float() * out.float()).sum(-1) dq, dk, dv = torch.autograd.grad(out, (q, k, v), g) dq_ref, dk_ref, dv_ref = torch.autograd.grad(out_ref, (q, k, v), g) dq_pt, dk_pt, dv_pt = torch.autograd.grad(out_pt, (q, k, v), g) print(f"dQ max diff: {(dq - dq_ref).abs().max().item()}") print(f"dK max diff: {(dk - dk_ref).abs().max().item()}") print(f"dV max diff: {(dv - dv_ref).abs().max().item()}") print(f"dQ mean diff: {(dq - dq_ref).abs().mean().item()}") print(f"dK mean diff: {(dk - dk_ref).abs().mean().item()}") print(f"dV mean diff: {(dv - dv_ref).abs().mean().item()}") print(f"dQ Pytorch max diff: {(dq_pt - dq_ref).abs().max().item()}") print(f"dK Pytorch max diff: {(dk_pt - dk_ref).abs().max().item()}") print(f"dV Pytorch max diff: {(dv_pt - dv_ref).abs().max().item()}") print(f"dQ Pytorch mean diff: {(dq_pt - dq_ref).abs().mean().item()}") print(f"dK Pytorch mean diff: {(dk_pt - dk_ref).abs().mean().item()}") print(f"dV Pytorch mean diff: {(dv_pt - dv_ref).abs().mean().item()}") # dS = torch.einsum('bthd,bshd->bhts', g.float(), v.float()) # P = torch.softmax(qk, -1) # dP = P * (dS - do_o.unsqueeze(1)) # dQ = torch.einsum('bhts,bshd->bthd', dP, k.float()) # dV = torch.einsum('bhts,bthd->bshd', P, g.float()) # dK = torch.einsum('bhts,bthd->bshd', dP, q.float()) # breakpoint() # Check that FlashAttention's numerical error is at most twice the numerical error # of a Pytorch implementation. # breakpoint() if dtype != torch.float8_e4m3fn: assert (out - out_ref).abs().max().item() <= 2 * ( out_pt - out_ref ).abs().max().item() + 3e-5 else: # just test correctness of fp8 kernel w/o further quantization techniques assert (out - out_ref).abs().max().item() <= 4 * ( out_pt - out_ref ).abs().max().item() + 2e-2 if d <= 128 and dtype != torch.float8_e4m3fn: assert (dq - dq_ref).abs().max().item() <= 2 * ( dq_pt - dq_ref ).abs().max().item() + 3e-5 assert (dk - dk_ref).abs().max().item() <= 2 * ( dk_pt - dk_ref ).abs().max().item() + 3e-5 assert (dv - dv_ref).abs().max().item() <= 2 * ( dv_pt - dv_ref ).abs().max().item() + 3e-5 @pytest.mark.parametrize("dtype", [torch.float16, torch.bfloat16]) # @pytest.mark.parametrize("dtype", [torch.float16]) @pytest.mark.parametrize("mha_type", ["mha", "mqa", "gqa"]) # @pytest.mark.parametrize("mha_type", ["mha"]) @pytest.mark.parametrize("causal", [False, True]) # @pytest.mark.parametrize("causal", [True]) @pytest.mark.parametrize("local", [False, True]) # @pytest.mark.parametrize("local", [False]) @pytest.mark.parametrize("deterministic", [False, True]) # @pytest.mark.parametrize("deterministic", [False]) @pytest.mark.parametrize("add_unused_qkv", [False, True]) # @pytest.mark.parametrize("add_unused_qkv", [True]) # @pytest.mark.parametrize("d", [32, 59, 64, 80, 96, 111, 128, 160, 192, 224, 256]) # @pytest.mark.parametrize("d", [32, 64, 96, 128, 160, 192, 224, 256]) # @pytest.mark.parametrize('d', [256]) # @pytest.mark.parametrize("d", [64, 128, 256]) @pytest.mark.parametrize("d", [64, 128]) # @pytest.mark.parametrize("d", [128]) @pytest.mark.parametrize( "seqlen_q,seqlen_k", [ (1, 1), (1, 3), (2, 1), (511, 1), (3, 513), (64, 128), (113, 203), (128, 128), (128, 217), (113, 211), (108, 256), (256, 512), (384, 256), (512, 256), (640, 128), (1024, 1024), (1023, 1024), (1024, 1023), (2048, 2048), ], ) # @pytest.mark.parametrize('seqlen_q,seqlen_k', [(128, 128)]) def test_flash_attn_varlen_output( seqlen_q, seqlen_k, d, causal, local, deterministic, add_unused_qkv, mha_type, dtype ): if ( max(seqlen_q, seqlen_k) >= 2048 and torch.cuda.get_device_properties("cuda").total_memory <= 16 * 2**30 ): pytest.skip() # Reference implementation OOM device = "cuda" # set seed torch.random.manual_seed(0) # batch_size = 1 # nheads = 1 # nheads_kv = 1 batch_size = 9 nheads = 6 nheads_kv = 6 if mha_type == "mha" else (2 if mha_type == "gqa" else 1) window_size = (-1, -1) if not local else torch.randint(0, seqlen_k, (2,)) q = torch.randn( batch_size, seqlen_q, nheads, d, device=device, dtype=dtype, requires_grad=True ) k = torch.randn( batch_size, seqlen_k, nheads_kv, d, device=device, dtype=dtype, requires_grad=True, ) v = torch.randn( batch_size, seqlen_k, nheads_kv, d, device=device, dtype=dtype, requires_grad=True, ) query_padding_mask = generate_random_padding_mask( seqlen_q, batch_size, device, mode="random", zero_lengths=False ) key_padding_mask = generate_random_padding_mask( seqlen_k, batch_size, device, mode="random", zero_lengths=True ) # key_padding_mask = generate_random_padding_mask(seqlen_k, batch_size, device, mode='full') def _gen_unused_masks(padding_mask, add_unused, max_seq_len, bs, device): if add_unused: another_mask = generate_random_padding_mask(max_seq_len, bs, device) attn_mask = torch.logical_and(padding_mask, another_mask) unused_mask = torch.logical_xor( torch.logical_or(padding_mask, another_mask), attn_mask ) else: attn_mask = padding_mask unused_mask = None return attn_mask, unused_mask query_padding_mask, query_unused_mask = _gen_unused_masks( query_padding_mask, add_unused_qkv, seqlen_q, batch_size, q.device ) key_padding_mask, key_unused_mask = _gen_unused_masks( key_padding_mask, add_unused_qkv, seqlen_k, batch_size, k.device ) ( q_unpad, k_unpad, v_unpad, cu_seqlens_q, cu_seqlens_k, seqused_q, seqused_k, max_seqlen_q, max_seqlen_k, q, k, v, output_pad_fn, dq_pad_fn, dk_pad_fn, ) = generate_qkv( q, k, v, query_padding_mask, key_padding_mask, kvpacked=False, query_unused_mask=query_unused_mask, key_unused_mask=key_unused_mask, ) # print("cu_seqlens_q: ", cu_seqlens_q) # print("cu_seqlens_k: ", cu_seqlens_k) # print("q_unpad, shape: ", q_unpad.shape) # print("k_unpad, shape: ", k_unpad.shape) # print("v_unpad, shape: ", v_unpad.shape) out_unpad, sm_lse = flash_attn_varlen_func( q_unpad, k_unpad, v_unpad, cu_seqlens_q, cu_seqlens_k, max_seqlen_q, max_seqlen_k, causal=causal, deterministic=deterministic, seqused_q=seqused_q, seqused_k=seqused_k, window_size=window_size, ) out = output_pad_fn(out_unpad) if query_unused_mask is not None: q_zero_masking = rearrange(query_unused_mask, "b s -> b s 1 1") out.masked_fill_(q_zero_masking, 0.0) dropout_mask = None out_ref, attn_ref = attention_ref( q, k, v, query_padding_mask, key_padding_mask, causal=causal, window_size=window_size, ) out_pt, attn_pt = attention_ref( q, k, v, query_padding_mask, key_padding_mask, causal=causal, window_size=window_size, upcast=False, reorder_ops=True, ) print(f"Output max diff: {(out - out_ref).abs().max().item()}") print(f"Output mean diff: {(out - out_ref).abs().mean().item()}") print(f"Pytorch max diff: {(out_pt - out_ref).abs().max().item()}") print(f"Pytorch mean diff: {(out_pt - out_ref).abs().mean().item()}") g = torch.randn_like(out) if d <= 128: ( dq_unpad, dk_unpad, dv_unpad, ) = torch.autograd.grad(out, (q_unpad, k_unpad, v_unpad), g) dk = dk_pad_fn(dk_unpad) dv = dk_pad_fn(dv_unpad) if key_unused_mask is not None: k_zero_masking = rearrange(key_unused_mask, "b s -> b s 1 1") dk.masked_fill_(k_zero_masking, 0.0) dv.masked_fill_(k_zero_masking, 0.0) ( dq_ref, dk_ref, dv_ref, ) = torch.autograd.grad(out_ref, (q, k, v), g) zero_masking = rearrange( torch.logical_not(torch.any(key_padding_mask, 1)), "b -> b 1 1 1" ) dk_ref.masked_fill_(zero_masking, 0.0) dv_ref.masked_fill_(zero_masking, 0.0) ( dq_pt, dk_pt, dv_pt, ) = torch.autograd.grad(out_pt, (q, k, v), g) dk_pt.masked_fill_(zero_masking, 0.0) dv_pt.masked_fill_(zero_masking, 0.0) dq = dq_pad_fn(dq_unpad) if query_unused_mask is not None: dq.masked_fill_(q_zero_masking, 0.0) print(f"dQ max diff: {(dq - dq_ref).abs().max().item()}") print(f"dK max diff: {(dk - dk_ref).abs().max().item()}") print(f"dV max diff: {(dv - dv_ref).abs().max().item()}") print(f"dQ mean diff: {(dq - dq_ref).abs().mean().item()}") print(f"dK mean diff: {(dk - dk_ref).abs().mean().item()}") print(f"dV mean diff: {(dv - dv_ref).abs().mean().item()}") print(f"dQ Pytorch max diff: {(dq_pt - dq_ref).abs().max().item()}") print(f"dK Pytorch max diff: {(dk_pt - dk_ref).abs().max().item()}") print(f"dV Pytorch max diff: {(dv_pt - dv_ref).abs().max().item()}") print(f"dQ Pytorch mean diff: {(dq_pt - dq_ref).abs().mean().item()}") print(f"dK Pytorch mean diff: {(dk_pt - dk_ref).abs().mean().item()}") print(f"dV Pytorch mean diff: {(dv_pt - dv_ref).abs().mean().item()}") # Check that FlashAttention's numerical error is at most twice the numerical error # of a Pytorch implementation. assert (out - out_ref).abs().max().item() <= 2 * ( out_pt - out_ref ).abs().max().item() if d <= 128: assert (dq - dq_ref).abs().max().item() < 1e-4 or ( dq - dq_ref ).abs().max().item() <= 3 * (dq_pt - dq_ref).abs().max().item() assert (dk - dk_ref).abs().max().item() < 1e-4 or ( dk - dk_ref ).abs().max().item() <= 3 * (dk_pt - dk_ref).abs().max().item() assert (dv - dv_ref).abs().max().item() < 1e-4 or ( dv - dv_ref ).abs().max().item() <= 3 * (dv_pt - dv_ref).abs().max().item() @pytest.mark.parametrize("dtype", [torch.bfloat16]) # @pytest.mark.parametrize("dtype", [torch.float16]) @pytest.mark.parametrize("mha_type", ["mha", "mqa", "gqa"]) @pytest.mark.parametrize("causal", [False, True]) # @pytest.mark.parametrize("causal", [False]) @pytest.mark.parametrize("deterministic", [True, False]) # @pytest.mark.parametrize("deterministic", [False]) # @pytest.mark.parametrize("d", [32, 59, 64, 80, 96, 111, 128, 160, 192, 224, 256]) # @pytest.mark.parametrize("d", [32, 64, 96, 128, 160, 192, 224, 256]) # @pytest.mark.parametrize('d', [128]) # @pytest.mark.parametrize("d", [64, 128, 256]) @pytest.mark.parametrize("d", [128, 64]) # @pytest.mark.parametrize("d", [128]) @pytest.mark.parametrize( "seqlen_q,seqlen_k", [ # (1, 1), # (1, 3), # (2, 1), # (511, 1), # (3, 513), # (64, 128), # (113, 203), # (128, 128), # (128, 217), # (113, 211), # (108, 256), (256, 512), # (384, 256), (768, 512), # (512, 256), # (640, 128), (1024, 1024), # (1023, 1024), # (1024, 1023), # (2048, 2048), ], ) @pytest.mark.parametrize("add_unused_qkv", [False]) @pytest.mark.parametrize("shuffle_pages", [True, False]) # @pytest.mark.parametrize('seqlen_q,seqlen_k', [(128, 128)]) def test_flash_attn_paged1( seqlen_q, seqlen_k, d, causal, deterministic, add_unused_qkv, mha_type, dtype, shuffle_pages, ): run_conf = locals() if ( max(seqlen_q, seqlen_k) >= 2048 and torch.cuda.get_device_properties("cuda").total_memory <= 16 * 2**30 ): pytest.skip() # Reference implementation OOM device = "cuda" # set seed torch.random.manual_seed(0) # batch_size = 1 # nheads = 1 batch_size = 9 nheads = 6 nheads_kv = 6 if mha_type == "mha" else (2 if mha_type == "gqa" else 1) q = torch.randn( batch_size, seqlen_q, nheads, d, device=device, dtype=dtype, requires_grad=True ) page_size = 256 num_pages = batch_size * seqlen_k // page_size assert seqlen_k % page_size == 0, "Max seqlen must be divisible by page size" block_table = torch.reshape( torch.arange(num_pages, dtype=torch.int32, device=device), (batch_size, -1) ) k_paged = torch.randn( num_pages, page_size, nheads_kv, d, device=device, dtype=dtype, requires_grad=True, ) v_paged = torch.randn( num_pages, page_size, nheads_kv, d, device=device, dtype=dtype, requires_grad=True, ) if shuffle_pages: block_table = torch.randperm(num_pages, dtype=torch.int32, device=device).view( batch_size, -1 ) k = torch.index_select(k_paged, 0, block_table.view(-1)).view( batch_size, seqlen_k, nheads_kv, d ) v = torch.index_select(v_paged, 0, block_table.view(-1)).view( batch_size, seqlen_k, nheads_kv, d ) else: k = torch.reshape(k_paged, (batch_size, seqlen_k, nheads_kv, d)) v = torch.reshape(v_paged, (batch_size, seqlen_k, nheads_kv, d)) query_padding_mask = generate_random_padding_mask( seqlen_q, batch_size, device, mode="random" ) key_padding_mask = generate_random_padding_mask( seqlen_k, batch_size, device, mode="random" ) # key_padding_mask = generate_random_padding_mask(seqlen_k, batch_size, device, mode='full') def _gen_unused_masks(padding_mask, add_unused, max_seq_len, bs, device): if add_unused: another_mask = generate_random_padding_mask(max_seq_len, bs, device) attn_mask = torch.logical_and(padding_mask, another_mask) unused_mask = torch.logical_xor( torch.logical_or(padding_mask, another_mask), attn_mask ) else: attn_mask = padding_mask unused_mask = None return attn_mask, unused_mask query_padding_mask, query_unused_mask = _gen_unused_masks( query_padding_mask, add_unused_qkv, seqlen_q, batch_size, q.device ) key_padding_mask, key_unused_mask = _gen_unused_masks( key_padding_mask, add_unused_qkv, seqlen_k, batch_size, k.device ) ( q_unpad, k_unpad, v_unpad, cu_seqlens_q, cu_seqlens_k, seqused_q, seqused_k, max_seqlen_q, max_seqlen_k, q, k, v, output_pad_fn, dq_pad_fn, dk_pad_fn, ) = generate_qkv( q, k, v, query_padding_mask, key_padding_mask, kvpacked=False, query_unused_mask=query_unused_mask, key_unused_mask=key_unused_mask, ) # print("cu_seqlens_q: ", cu_seqlens_q) # print("cu_seqlens_k: ", cu_seqlens_k) # print("q_unpad, shape: ", q_unpad.shape) # print("k_unpad, shape: ", k_unpad.shape) # print("v_unpad, shape: ", v_unpad.shape) out_unpad, sm_lse = flash_attn_varlen_func( q_unpad, k_paged, v_paged, cu_seqlens_q, cu_seqlens_k, max_seqlen_q, max_seqlen_k, causal=causal, deterministic=deterministic, block_table=block_table, ) out = output_pad_fn(out_unpad) out_unpaged_unpad, sm_unpaged_lse = flash_attn_varlen_func( q_unpad, k_unpad, v_unpad, cu_seqlens_q, cu_seqlens_k, max_seqlen_q, max_seqlen_k, causal=causal, deterministic=deterministic, ) out_unpaged = output_pad_fn(out_unpaged_unpad) dropout_mask = None out_ref, attn_ref = attention_ref( q, k, v, query_padding_mask, key_padding_mask, causal=causal, ) out_pt, attn_pt = attention_ref( q, k, v, query_padding_mask, key_padding_mask, causal=causal, upcast=False, reorder_ops=True, ) # print(f"{k.stride()=}, {v.stride()=}, {k_paged.stride()=}, {v_paged.stride()=}, {block_table.stride()=}") print(f"Output max diff: {(out - out_ref).abs().max().item()}") print(f"Output mean diff: {(out - out_ref).abs().mean().item()}") print(f"Pytorch max diff: {(out_pt - out_ref).abs().max().item()}") print(f"Pytorch mean diff: {(out_pt - out_ref).abs().mean().item()}") print(f"Output max diff paged vs varlen: {(out - out_unpaged).abs().max().item()}") print( f"Output mean diff paged vs varlen: {(out - out_unpaged).abs().mean().item()}" ) # Check that FlashAttention's numerical error is at most twice the numerical error # of a Pytorch implementation. # import fbvscode; fbvscode.set_trace() assert (out - out_ref).abs().max().item() <= 2 * ( out_pt - out_ref ).abs().max().item() @pytest.mark.parametrize("dtype", ([torch.bfloat16])) # @pytest.mark.parametrize("dtype", [torch.bfloat16]) @pytest.mark.parametrize("local", [False]) # @pytest.mark.parametrize("local", [True]) @pytest.mark.parametrize( "d", [128, 64] ) # [32, 40, 59, 64, 80, 96, 111, 128, 160, 192, 224, 256]) # @pytest.mark.parametrize("d", [32, 64, 96, 128, 160, 192, 224, 256]) # @pytest.mark.parametrize('d', [32, 40, 64, 80, 96, 128, 160, 192]) # @pytest.mark.parametrize('d', [32, 64, 96, 128, 160, 192]) # @pytest.mark.parametrize('d', [56, 80]) # @pytest.mark.parametrize("d", [64]) @pytest.mark.parametrize("swap_sq_sk", [False, True]) # @pytest.mark.parametrize("swap_sq_sk", [True]) @pytest.mark.parametrize( "seqlen_q,seqlen_k", [ (1, 239), (3, 799), (127, 512), (127, 513), (113, 203), (128, 217), (113, 211), (108, 256), (256, 512), (1023, 1024), ], ) # TODO: add smaller page sizes when https://github.com/Dao-AILab/flash-attention/pull/824 is merged @pytest.mark.parametrize("paged_kv_block_size", [256, 512]) # @pytest.mark.parametrize("seqlen_q,seqlen_k", [(256, 128)]) def test_flash_attn_varlen_paged2( seqlen_q, seqlen_k, swap_sq_sk, d, local, paged_kv_block_size, dtype ): # Test ported from FlashAttention V2 test test_flash_attn_varlen_causal def _generate_block_kvcache( seqlen_k, paged_kv_block_size, batch_size, nheads_k, d, device, dtype ): num_blocks = math.ceil(seqlen_k / paged_kv_block_size) * batch_size * 3 k_cache_paged = torch.randn( num_blocks, paged_kv_block_size, nheads_k, d, device=device, dtype=dtype ) v_cache_paged = torch.randn( num_blocks, paged_kv_block_size, nheads_k, d, device=device, dtype=dtype ) block_table = rearrange( torch.randperm(num_blocks, dtype=torch.int32, device=device), "(b nblocks) -> b nblocks", b=batch_size, ) k_cache = rearrange( # pytorch 1.12 doesn't have indexing with int32 k_cache_paged[block_table.to(dtype=torch.long).flatten()], "(b nblocks) block_size ... -> b (nblocks block_size) ...", b=batch_size, )[:, :seqlen_k] v_cache = rearrange( v_cache_paged[block_table.to(dtype=torch.long).flatten()], "(b nblocks) block_size ... -> b (nblocks block_size) ...", b=batch_size, )[:, :seqlen_k] return k_cache, v_cache, block_table, k_cache_paged, v_cache_paged, num_blocks if ( max(seqlen_q, seqlen_k) >= 2048 and torch.cuda.get_device_properties("cuda").total_memory <= 16 * 2**30 ): pytest.skip() # Reference implementation OOM if swap_sq_sk: seqlen_q, seqlen_k = seqlen_k, seqlen_q device = "cuda" causal = True # set seed torch.random.manual_seed(0) batch_size = 8 nheads = 9 window_size = (-1, -1) if not local else torch.randint(0, seqlen_k, (2,)) q = torch.randn( batch_size, seqlen_q, nheads, d, device=device, dtype=dtype, requires_grad=True ) if paged_kv_block_size is None: k = torch.randn( batch_size, seqlen_k, nheads, d, device=device, dtype=dtype, requires_grad=True, ) v = torch.randn( batch_size, seqlen_k, nheads, d, device=device, dtype=dtype, requires_grad=True, ) block_table = None else: k, v, block_table, k_cache_paged, v_cache_paged, num_blocks = ( _generate_block_kvcache( seqlen_k, paged_kv_block_size, batch_size, nheads, d, device, dtype ) ) query_padding_mask = generate_random_padding_mask( seqlen_q, batch_size, device, mode="random" ) key_padding_mask = generate_random_padding_mask( seqlen_k, batch_size, device, mode="random" ) def _gen_unused_masks(padding_mask, add_unused, max_seq_len, bs, device): if add_unused: another_mask = generate_random_padding_mask(max_seq_len, bs, device) attn_mask = torch.logical_and(padding_mask, another_mask) unused_mask = torch.logical_xor( torch.logical_or(padding_mask, another_mask), attn_mask ) else: attn_mask = padding_mask unused_mask = None return attn_mask, unused_mask query_padding_mask, query_unused_mask = _gen_unused_masks( query_padding_mask, False, seqlen_q, batch_size, q.device ) key_padding_mask, key_unused_mask = _gen_unused_masks( key_padding_mask, False, seqlen_k, batch_size, k.device ) ( q_unpad, k_unpad, v_unpad, cu_seqlens_q, cu_seqlens_k, seqused_q, seqused_k, max_seqlen_q, max_seqlen_k, q, k, v, output_pad_fn, dq_pad_fn, dk_pad_fn, ) = generate_qkv(q, k, v, query_padding_mask, key_padding_mask, kvpacked=False) out_unpad, sm_lse = flash_attn_varlen_func( q_unpad, k_unpad if paged_kv_block_size is None else k_cache_paged, v_unpad if paged_kv_block_size is None else v_cache_paged, cu_seqlens_q, cu_seqlens_k, max_seqlen_q, max_seqlen_k, causal=causal, block_table=block_table, ) out = output_pad_fn(out_unpad) out_ref, attn_ref = attention_ref( q, k, v, query_padding_mask, key_padding_mask, None, 0.0, None, causal=causal, window_size=window_size, ) out_pt, attn_pt = attention_ref( q, k, v, query_padding_mask, key_padding_mask, None, 0.0, None, causal=causal, window_size=window_size, upcast=False, reorder_ops=True, ) print(f"Output max diff: {(out - out_ref).abs().max().item()}") print(f"Output mean diff: {(out - out_ref).abs().mean().item()}") print(f"Pytorch max diff: {(out_pt - out_ref).abs().max().item()}") print(f"Pytorch mean diff: {(out_pt - out_ref).abs().mean().item()}") g = torch.randn_like(out) do_o = (g.float() * out.float()).sum(-1) test_backward = block_table is None if test_backward: ( dq_unpad, dk_unpad, dv_unpad, ) = torch.autograd.grad(out, (q_unpad, k_unpad, v_unpad), g) dq = dq_pad_fn(dq_unpad) dk = dk_pad_fn(dk_unpad) dv = dk_pad_fn(dv_unpad) ( dq_ref, dk_ref, dv_ref, ) = torch.autograd.grad(out_ref, (q, k, v), g) ( dq_pt, dk_pt, dv_pt, ) = torch.autograd.grad(out_pt, (q, k, v), g) print(f"dQ max diff: {(dq - dq_ref).abs().max().item()}") print(f"dK max diff: {(dk - dk_ref).abs().max().item()}") print(f"dV max diff: {(dv - dv_ref).abs().max().item()}") print(f"dQ mean diff: {(dq - dq_ref).abs().mean().item()}") print(f"dK mean diff: {(dk - dk_ref).abs().mean().item()}") print(f"dV mean diff: {(dv - dv_ref).abs().mean().item()}") print(f"dQ Pytorch max diff: {(dq_pt - dq_ref).abs().max().item()}") print(f"dK Pytorch max diff: {(dk_pt - dk_ref).abs().max().item()}") print(f"dV Pytorch max diff: {(dv_pt - dv_ref).abs().max().item()}") print(f"dQ Pytorch mean diff: {(dq_pt - dq_ref).abs().mean().item()}") print(f"dK Pytorch mean diff: {(dk_pt - dk_ref).abs().mean().item()}") print(f"dV Pytorch mean diff: {(dv_pt - dv_ref).abs().mean().item()}") # Check that FlashAttention's numerical error is at most twice the numerical error # of a Pytorch implementation. assert (out - out_ref).abs().max().item() <= 2 * ( out_pt - out_ref ).abs().max().item() + 1e-5 if test_backward: assert (dq - dq_ref).abs().max().item() <= 2 * ( dq_pt - dq_ref ).abs().max().item() + 1e-5 assert (dk - dk_ref).abs().max().item() <= 2 * ( dk_pt - dk_ref ).abs().max().item() + 1e-5 assert (dv - dv_ref).abs().max().item() <= 2 * ( dv_pt - dv_ref ).abs().max().item() + 1e-5 if __name__ == "__main__": test_flash_attn_varlen_causal(512, 768, False, 128, False, 256, torch.bfloat16)