Note
Click here to download the full example code
Layer NormalizationΒΆ
Out:
layer-norm-backward:
N Triton Torch Apex
0 1024.0 356.173905 96.755900 299.707322
1 1536.0 400.695643 133.083026 338.201833
2 2048.0 486.653476 159.584422 321.254900
3 2560.0 455.111129 179.649115 323.368411
4 3072.0 508.468972 191.005181 320.556515
5 3584.0 551.384634 206.769233 310.527060
6 4096.0 558.545450 219.919464 297.890900
7 4608.0 493.714279 231.364016 285.767436
8 5120.0 518.481012 241.414550 283.133649
9 5632.0 534.260858 241.803217 288.820505
10 6144.0 540.131844 247.409397 285.214712
11 6656.0 525.473708 255.182111 284.748652
12 7168.0 503.017523 259.084340 283.881181
13 7680.0 482.513091 262.190612 275.104486
14 8192.0 463.698115 265.686491 283.296835
15 8704.0 406.412440 266.789264 283.826081
16 9216.0 419.703988 270.727053 287.251954
17 9728.0 428.388977 279.942444 288.950501
18 10240.0 437.295395 285.104413 286.767793
19 10752.0 424.421071 245.292781 289.291486
20 11264.0 423.724120 244.426754 286.069848
21 11776.0 418.702211 247.915800 288.097854
22 12288.0 414.784810 253.578674 293.737063
23 12800.0 409.327110 252.631590 287.371378
24 13312.0 408.030638 252.360194 289.653667
25 13824.0 403.620451 256.197690 291.287079
26 14336.0 394.116833 253.921779 286.242939
27 14848.0 384.829370 257.108233 289.246765
28 15360.0 380.041240 257.071134 284.884090
29 15872.0 370.913333 260.731015 289.679087
import torch
import triton
import triton.language as tl
try:
# This is https://github.com/NVIDIA/apex, NOT the apex on PyPi, so it
# should not be added to extras_require in setup.py.
import apex
HAS_APEX = True
except ModuleNotFoundError:
HAS_APEX = False
# Forward Pass
@triton.jit
def _layer_norm_fwd_fused(X, Y, W, B, M, V, stride, N, eps,
BLOCK_SIZE: tl.constexpr):
# position of elements processed by this program
row = tl.program_id(0)
cols = tl.arange(0, BLOCK_SIZE)
mask = cols < N
# offset data pointers to start at the row of interest
X += row * stride
Y += row * stride
# load data and cast to float32
x = tl.load(X + cols, mask=mask, other=0).to(tl.float32)
# compute mean
mean = tl.sum(x, axis=0) / N
# compute std
xmean = tl.where(mask, x - mean, 0.)
var = tl.sum(xmean * xmean, axis=0) / N
rstd = 1 / tl.sqrt(var + eps)
xhat = xmean * rstd
# write-back mean/rstd
tl.store(M + row, mean)
tl.store(V + row, rstd)
# multiply by weight and add bias
w = tl.load(W + cols, mask=mask)
b = tl.load(B + cols, mask=mask)
y = xhat * w + b
# write-back
tl.store(Y + cols, y, mask=mask)
# Backward pass (DX + partial DW + partial DB)
@triton.jit
def _layer_norm_bwd_dx_fused(DX, DY, DW, DB, X, W, B, M, V, Lock, stride, N, eps,
GROUP_SIZE_M: tl.constexpr, BLOCK_SIZE_N: tl.constexpr):
# position of elements processed by this program
row = tl.program_id(0)
cols = tl.arange(0, BLOCK_SIZE_N)
mask = cols < N
# offset data pointers to start at the row of interest
X += row * stride
DY += row * stride
DX += row * stride
# offset locks and weight/bias gradient pointer
# each kernel instance accumulates partial sums for
# DW and DB into one of GROUP_SIZE_M independent buffers
# these buffers stay in the L2, which allow this kernel
# to be fast
lock_id = row % GROUP_SIZE_M
Lock += lock_id
Count = Lock + GROUP_SIZE_M
DW = DW + lock_id * N + cols
DB = DB + lock_id * N + cols
# load data to SRAM
x = tl.load(X + cols, mask=mask, other=0).to(tl.float32)
dy = tl.load(DY + cols, mask=mask, other=0).to(tl.float32)
w = tl.load(W + cols, mask=mask).to(tl.float32)
mean = tl.load(M + row)
rstd = tl.load(V + row)
# compute dx
xhat = (x - mean) * rstd
wdy = w * dy
xhat = tl.where(mask, xhat, 0.)
wdy = tl.where(mask, wdy, 0.)
mean1 = tl.sum(xhat * wdy, axis=0) / N
mean2 = tl.sum(wdy, axis=0) / N
dx = (wdy - (xhat * mean1 + mean2)) * rstd
# write-back dx
tl.store(DX + cols, dx, mask=mask)
# accumulate partial sums for dw/db
partial_dw = (dy * xhat).to(w.dtype)
partial_db = (dy).to(w.dtype)
while tl.atomic_cas(Lock, 0, 1) == 1:
pass
count = tl.load(Count)
# first store doesn't accumulate
if count == 0:
tl.atomic_xchg(Count, 1)
else:
partial_dw += tl.load(DW, mask=mask)
partial_db += tl.load(DB, mask=mask)
tl.store(DW, partial_dw, mask=mask)
tl.store(DB, partial_db, mask=mask)
# release lock
tl.atomic_xchg(Lock, 0)
# Backward pass (total DW + total DB)
@triton.jit
def _layer_norm_bwd_dwdb(DW, DB, FINAL_DW, FINAL_DB, M, N,
BLOCK_SIZE_M: tl.constexpr, BLOCK_SIZE_N: tl.constexpr):
pid = tl.program_id(0)
cols = pid * BLOCK_SIZE_N + tl.arange(0, BLOCK_SIZE_N)
dw = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
db = tl.zeros((BLOCK_SIZE_M, BLOCK_SIZE_N), dtype=tl.float32)
for i in range(0, M, BLOCK_SIZE_M):
rows = i + tl.arange(0, BLOCK_SIZE_M)
mask = (rows[:, None] < M) & (cols[None, :] < N)
offs = rows[:, None] * N + cols[None, :]
dw += tl.load(DW + offs, mask=mask, other=0.)
db += tl.load(DB + offs, mask=mask, other=0.)
sum_dw = tl.sum(dw, axis=0)
sum_db = tl.sum(db, axis=0)
tl.store(FINAL_DW + cols, sum_dw, mask=cols < N)
tl.store(FINAL_DB + cols, sum_db, mask=cols < N)
class LayerNorm(torch.autograd.Function):
@staticmethod
def forward(ctx, x, normalized_shape, weight, bias, eps):
# allocate output
y = torch.empty_like(x)
# reshape input data into 2D tensor
x_arg = x.reshape(-1, x.shape[-1])
M, N = x_arg.shape
mean = torch.empty((M, ), dtype=torch.float32, device='cuda')
rstd = torch.empty((M, ), dtype=torch.float32, device='cuda')
# Less than 64KB per feature: enqueue fused kernel
MAX_FUSED_SIZE = 65536 // x.element_size()
BLOCK_SIZE = min(MAX_FUSED_SIZE, triton.next_power_of_2(N))
if N > BLOCK_SIZE:
raise RuntimeError("This layer norm doesn't support feature dim >= 64KB.")
# heuristics for number of warps
num_warps = min(max(BLOCK_SIZE // 256, 1), 8)
# enqueue kernel
_layer_norm_fwd_fused[(M,)](x_arg, y, weight, bias, mean, rstd,
x_arg.stride(0), N, eps,
BLOCK_SIZE=BLOCK_SIZE, num_warps=num_warps)
ctx.save_for_backward(x, weight, bias, mean, rstd)
ctx.BLOCK_SIZE = BLOCK_SIZE
ctx.num_warps = num_warps
ctx.eps = eps
return y
@staticmethod
def backward(ctx, dy):
x, w, b, m, v = ctx.saved_tensors
# heuristics for amount of parallel reduction stream for DG/DB
N = w.shape[0]
GROUP_SIZE_M = 64
if N <= 8192: GROUP_SIZE_M = 96
if N <= 4096: GROUP_SIZE_M = 128
if N <= 1024: GROUP_SIZE_M = 256
# allocate output
locks = torch.zeros(2 * GROUP_SIZE_M, dtype=torch.int32, device='cuda')
_dw = torch.empty((GROUP_SIZE_M, w.shape[0]), dtype=x.dtype, device=w.device)
_db = torch.empty((GROUP_SIZE_M, w.shape[0]), dtype=x.dtype, device=w.device)
dw = torch.empty((w.shape[0],), dtype=w.dtype, device=w.device)
db = torch.empty((w.shape[0],), dtype=w.dtype, device=w.device)
dx = torch.empty_like(dy)
# enqueue kernel using forward pass heuristics
# also compute partial sums for DW and DB
x_arg = x.reshape(-1, x.shape[-1])
M, N = x_arg.shape
_layer_norm_bwd_dx_fused[(M,)](dx, dy, _dw, _db, x, w, b, m, v, locks,
x_arg.stride(0), N, ctx.eps,
BLOCK_SIZE_N=ctx.BLOCK_SIZE,
GROUP_SIZE_M=GROUP_SIZE_M,
num_warps=ctx.num_warps)
grid = lambda meta: [triton.cdiv(N, meta['BLOCK_SIZE_N'])]
# accumulate partial sums in separate kernel
_layer_norm_bwd_dwdb[grid](_dw, _db, dw, db, GROUP_SIZE_M, N,
BLOCK_SIZE_M=32,
BLOCK_SIZE_N=128)
return dx, None, dw, db, None
layer_norm = LayerNorm.apply
def test_layer_norm(M, N, dtype, eps=1e-5, device='cuda'):
# create data
x_shape = (M, N)
w_shape = (x_shape[-1], )
weight = torch.rand(w_shape, dtype=dtype, device='cuda', requires_grad=True)
bias = torch.rand(w_shape, dtype=dtype, device='cuda', requires_grad=True)
x = -2.3 + 0.5 * torch.randn(x_shape, dtype=dtype, device='cuda')
dy = .1 * torch.randn_like(x)
x.requires_grad_(True)
# forward pass
y_tri = layer_norm(x, w_shape, weight, bias, eps)
y_ref = torch.nn.functional.layer_norm(x, w_shape, weight, bias, eps).to(dtype)
# backward pass (triton)
y_tri.backward(dy, retain_graph=True)
dx_tri, dw_tri, db_tri = [_.grad.clone() for _ in [x, weight, bias]]
x.grad, weight.grad, bias.grad = None, None, None
# backward pass (torch)
y_ref.backward(dy, retain_graph=True)
dx_ref, dw_ref, db_ref = [_.grad.clone() for _ in [x, weight, bias]]
# compare
triton.testing.assert_almost_equal(y_tri, y_ref)
triton.testing.assert_almost_equal(dx_tri, dx_ref)
triton.testing.assert_almost_equal(db_tri, db_ref, decimal=1)
triton.testing.assert_almost_equal(dw_tri, dw_ref, decimal=1)
@triton.testing.perf_report(
triton.testing.Benchmark(
x_names=['N'],
x_vals=[512 * i for i in range(2, 32)],
line_arg='provider',
line_vals=['triton', 'torch'] + (['apex'] if HAS_APEX else []),
line_names=['Triton', 'Torch'] + (['Apex'] if HAS_APEX else []),
styles=[('blue', '-'), ('green', '-'), ('orange', '-')],
ylabel='GB/s',
plot_name='layer-norm-backward',
args={'M': 4096, 'dtype': torch.float16, 'mode': 'backward'}
)
)
def bench_layer_norm(M, N, dtype, provider, mode='backward', eps=1e-5, device='cuda'):
# create data
x_shape = (M, N)
w_shape = (x_shape[-1], )
weight = torch.rand(w_shape, dtype=dtype, device='cuda', requires_grad=True)
bias = torch.rand(w_shape, dtype=dtype, device='cuda', requires_grad=True)
x = -2.3 + 0.5 * torch.randn(x_shape, dtype=dtype, device='cuda')
dy = .1 * torch.randn_like(x)
x.requires_grad_(True)
# utility functions
if provider == 'triton':
y_fwd = lambda: layer_norm(x, w_shape, weight, bias, eps)
if provider == 'torch':
y_fwd = lambda: torch.nn.functional.layer_norm(x, w_shape, weight, bias, eps)
if provider == 'apex':
apex_layer_norm = apex.normalization.FusedLayerNorm(w_shape).to(x.device).to(x.dtype)
y_fwd = lambda: apex_layer_norm(x)
# forward pass
if mode == 'forward':
gbps = lambda ms: 2 * x.numel() * x.element_size() / ms * 1e-6
ms, min_ms, max_ms = triton.testing.do_bench(y_fwd, rep=500)
# backward pass
if mode == 'backward':
gbps = lambda ms: 3 * x.numel() * x.element_size() / ms * 1e-6
y = y_fwd()
ms, min_ms, max_ms = triton.testing.do_bench(lambda: y.backward(dy, retain_graph=True),
grad_to_none=[x], rep=500)
return gbps(ms), gbps(max_ms), gbps(min_ms)
bench_layer_norm.run(save_path='.', print_data=True)
Total running time of the script: ( 2 minutes 16.519 seconds)