tinygrad/test/test_schedule.py

# this will be the new test_ops for the next level
# schedule confirms the right things are capable of fusing
# NOTE: this has overlap with external_test_opt.py

import unittest
from typing import List, Optional, Union
from tinygrad.tensor import Tensor
from tinygrad.ops import BinaryOps, LoadOps, ReduceOps
from tinygrad.helpers import DEBUG, flatten
from tinygrad.codegen.linearizer import Linearizer
from tinygrad.features.graph import print_tree
from tinygrad.engine.schedule import create_schedule
from tinygrad import nn, dtypes
from test.helpers import is_dtype_supported

def check_schedule(t:Union[Tensor, List[Tensor]], allowed:int, to_prerealize:Optional[List[Tensor]]=None, filter_loadops=True):
  if isinstance(t, Tensor): t = [t]
  seen = set()
  if to_prerealize:
    for pre in to_prerealize:
      for s in pre.schedule(seen=seen.copy()):
        for i,out in enumerate(s.outputs):
          seen.add(out)
  sched = create_schedule(flatten([r.lazydata.lbs for r in t]), seen)
  if filter_loadops: sched = [s for s in sched if s.ast[0].op not in LoadOps]
  if len(sched) != allowed: print(f"SCHEDULE ISSUE, expecting {allowed} got {len(sched)}")
  if len(sched) != allowed or DEBUG >= 3:
    for i, s in enumerate(sched):
      print("kernel", i+1)
      for op in s.ast: print_tree(op)
  assert len(sched) == allowed, f"{len(sched)} != {allowed}"
  # test the (non loadops) ops linearize
  for s in sched:
    if s.ast[0].op in LoadOps: continue
    l = Linearizer(*s.ast)
    l.hand_coded_optimizations()
    l.linearize()
  return sched

class TestSchedule(unittest.TestCase):
  def test_basic_binop_fusion(self):
    a = Tensor.empty(10)
    b = Tensor.empty(10)
    c = Tensor.empty(10)
    d = a+b+c
    check_schedule(d, 1)

  def test_basic_binop_fusion_deep(self):
    a = Tensor.empty(10)
    b = Tensor.empty(10)
    c = Tensor.empty(10)
    d = Tensor.empty(10)
    e = a+b+c+d
    check_schedule(e, 1)

  def test_mulacc_fusion(self):
    a = Tensor.empty(10)
    b = Tensor.empty(10)
    c = (a*b).sum()
    check_schedule(c, 1)

  def test_mulacc_relu_fusion(self):
    a = Tensor.empty(10)
    b = Tensor.empty(10)
    c = (a*b).sum().relu()
    check_schedule(c, 1)

  def test_binop_reshape_fusion(self):
    a = Tensor.empty(10)
    b = Tensor.empty(10)
    c = Tensor.empty(5,2)
    d = (a+b).reshape(5,2)+c
    check_schedule(d, 1)

  def test_binop_permute_fusion(self):
    a = Tensor.empty(2,5)
    b = Tensor.empty(2,5)
    c = Tensor.empty(5,2)
    d = (a+b).permute(1,0)+c
    check_schedule(d, 1)

  def test_constants_are_embedded(self):
    a = Tensor.empty(3,3) * 2
    check_schedule(a, 2, filter_loadops=False)

  def test_binop_elu_fusion(self):
    a = Tensor.empty(10)
    b = a.elu()
    check_schedule(b, 1)

  def test_binop_reshape_reduce_fusion(self):
    a = Tensor.empty(100)
    b = Tensor.empty(100)
    c = (a+b).reshape(10, 10).sum(axis=0, keepdim=True)
    check_schedule(c, 1)

  def test_reduce_reshape_binop_fusion(self):
    a = Tensor.empty(10,10)
    b = Tensor.empty(10)
    c = a.sum(axis=0) + b
    check_schedule(c, 1)

  @unittest.skip("not pushing permutes through reduces")
  def test_reduce_permute_binop_fusion(self):
    a = Tensor.empty(10,10,10)
    b = Tensor.empty(10,10,1)
    c = a.sum(axis=0, keepdim=True).permute(2,1,0) + b
    check_schedule(c, 1)

  def test_binop_early_reshape_reduce_fusion(self):
    a = Tensor.empty(100)
    b = Tensor.empty(100)
    c = Tensor.empty(10,10)
    d = ((a+b).reshape(10,10) + c).sum(axis=0)
    check_schedule(d, 1)

  def test_diamond_folded(self):
    a = Tensor.empty(10)
    b = Tensor.empty(10)
    c = Tensor.empty(10)
    d = Tensor.empty(10)
    ab = a+b
    e = (ab+c) + (ab+d)
    check_schedule(e, 1)

  def test_cache_binaryop(self):
    a = Tensor.empty(10)
    b = Tensor.empty(10)
    c = a+b
    d = a+b
    check_schedule(d, 0, [c])

  @unittest.skip("failing in old lazy")
  def test_cache_binaryop_reshaped(self):
    a = Tensor.empty(10)
    b = Tensor.empty(10)
    c = a+b
    d = a.reshape(10,1)+b.reshape(10,1)
    check_schedule(d, 0, [c])

  @unittest.skip("failing in new lazy")
  def test_cache_binaryop_transpose(self):
    a = Tensor.empty(10,10)
    b = Tensor.empty(10,10)
    c = (a.T*b.T).T #.contiguous()
    d = a*b
    check_schedule(d, 0, [c])

  def test_cache_two_reduceops(self):
    a = Tensor.empty(10)
    b = a.sum()
    c = a.sum()
    bc = b+c
    check_schedule(bc, 1)

  def test_cache_reduce_parent(self):
    x = Tensor.empty(32)
    r0 = x.mean(axis=0, keepdim=True)
    r1 = (x - r0).sum(axis=0).div(2)
    out = r0 + r1
    schedule = check_schedule(out, 2)
    reduceops = [x for si in schedule for out in si.ast for x in out.lazyops if x.op in ReduceOps]
    assert len(reduceops) == 2

  def test_cache_reduce_multiple_children(self):
    x = Tensor.empty(32)
    y = Tensor.empty(4, 4)
    r0 = x.mean(axis=0, keepdim=True)
    r1 = (x - r0).sum(axis=0).div(2)
    out0 = r0 + y
    out1 = r1 + y
    schedule = check_schedule([out0, out1], 4)
    reduceops = [x for si in schedule for out in si.ast for x in out.lazyops if x.op in ReduceOps]
    assert len(reduceops) == 2

  def test_fold_double_unary(self):
    y = Tensor.empty(2)
    out = y.sum(keepdim=True).sqrt().__neg__()
    check_schedule(out, 1)

  #@unittest.skip("may want to reconsider this")
  def test_fold_batchnorm(self):
    with Tensor.train():
      img = Tensor.empty(1,32,4,4)
      bn = nn.BatchNorm2d(32, track_running_stats=False)
      out = bn(img)
      check_schedule(out, 3)

  def test_fold_conv_batchnorm_notrain(self):
    with Tensor.train(False):
      img = Tensor.empty(1,3,8,8)
      c1 = nn.Conv2d(3,32,3)
      bn = nn.BatchNorm2d(32, track_running_stats=False)
      out = bn(c1(img)).relu()
      check_schedule(out, 1, [c1.weight, c1.bias])

  def test_fold_conv_batchnorm(self):
    with Tensor.train():
      img = Tensor.empty(1,3,8,8)
      c1 = nn.Conv2d(3,32,3)
      bn = nn.BatchNorm2d(32, track_running_stats=False)
      out = bn(c1(img)).relu()
      check_schedule(out, 4, [c1.weight, c1.bias])

  def test_fold_conv_batchnorm_optim(self):
    # this is too high
    for optim, cnt in [(nn.optim.Adam, 19), (nn.optim.SGD, 17)]:
      with self.subTest(optim=optim.__name__):
        with Tensor.train():
          img = Tensor.ones(1,3,4,4)
          c1 = nn.Conv2d(3,32,3)
          bn = nn.BatchNorm2d(32, track_running_stats=False)
          opt = optim(nn.state.get_parameters([c1, bn]))
          img_bn = bn(c1(img)).elu().sum()
          opt.zero_grad()
          img_bn.backward()
          check_schedule(opt.schedule_step(), cnt)

  def test_fold_conv_relu_backward(self):
    c1 = nn.Conv2d(3,16,3, bias=False)
    c1.weight.requires_grad = True

    # run
    img = Tensor.rand(2,3,64,64, requires_grad=True)
    c1(img).relu().mean().backward()
    # TODO: this should be 4, not 5
    # img.grad is requiring two reduces
    check_schedule([img.grad, c1.weight.grad], 5)

  def test_fold_batchnorm_backward(self):
    with Tensor.train():
      x = Tensor.empty((2, 16, 8, 8)).contiguous()
      bn = nn.BatchNorm2d(16)
      bn.weight.requires_grad = bn.bias.requires_grad = x.requires_grad = True
      fw = bn(x).contiguous_backward().relu().contiguous()
      fw.sum().backward()
      # TODO: this is too many
      check_schedule([x.grad, bn.weight.grad, bn.bias.grad, fw], 10)

  def test_fold_conv_relu(self):
    c1 = nn.Conv2d(3,16,3)

    # run
    img = Tensor.ones(2,3,64,64)
    out = c1(img).relu()
    check_schedule(out, 1, [c1.weight, c1.bias])

  def test_fold_conv_relu_nobias(self):
    img = Tensor.ones(1,4,8,8)
    c1 = nn.Conv2d(4, 4, kernel_size=3, bias=False)
    c2 = nn.Conv2d(4, 4, kernel_size=3, bias=False)
    out = img.sequential([c1, Tensor.relu, c2, Tensor.relu])
    check_schedule(out, 2, [c1.weight, c2.weight, img])

  def test_fold_conv_elu(self):
    c1 = nn.Conv2d(3,16,3)

    # run
    img = Tensor.rand(2,3,64,64)
    out = c1(img).elu()
    check_schedule(out, 1, [c1.weight, c1.bias, img])

  def test_two_sum(self):
    img = Tensor.empty(64,64)
    x = (img.sum(0) + img.sum(1))
    out = x.relu()
    del x    # is 3 without this
    check_schedule(out, 2)

  #@unittest.skip("failing in old lazy")
  def test_push_permute_through_reshape(self):
    a = Tensor.empty(16,16)
    b = Tensor.empty(16,16)
    c = (a+b).reshape(4,4,4,4).permute(2,3,0,1).contiguous()
    check_schedule(c, 1)

  #@unittest.skip("failing in old lazy")
  def test_push_permute_through_reshape_alt(self):
    a = Tensor.empty(4,4,4,4)
    b = Tensor.empty(4,4,4,4)
    c = (a+b).reshape(16,16).permute(1,0).contiguous()
    check_schedule(c, 1)

  def test_no_binop_rerun(self):
    a = Tensor.empty(16)
    b = Tensor.empty(16)
    c = a+b
    d = (a+b).reshape(16,1)
    check_schedule(d, 0, [c])

  def test_multi_permute_should_collapse(self):
    a = Tensor.empty(4,4,4,4)
    b = Tensor.empty(16)
    c = a.sum((0,1)).cast(dtypes.float16).permute(1,0).reshape(4,4,1).permute(1,0,2).reshape(16) + b
    check_schedule(c, 1)

  @unittest.skip("failing in old lazy")
  def test_fancy_reshape_fusion(self):
    a = Tensor.empty(10)
    b = Tensor.empty(10)
    c = a+b
    d = a.reshape(10,1)+b.reshape(10,1)
    out = c.sum() + d.sum()
    check_schedule(out, 1)

  # NOTE: for this to pass, LazyViews must be children of LazyBuffers so the (a+b) runs first
  @unittest.skip("not real world")
  def test_children_dont_push(self):
    a = Tensor.empty(10, 10, 1)
    b = Tensor.empty(10, 10, 1)
    d = (a+b).expand(10, 10, 10)
    e = (a+b).permute(2,1,0)
    f = d+e
    check_schedule(f, 2)

  @unittest.skip("failing in new lazy")
  def test_dont_fuse_binops_with_children(self):
    a = Tensor.empty(10)
    b = Tensor.empty(10)
    c = Tensor.empty(10)
    keep_me = a+b
    e = keep_me.sum() # noqa: F841 give keep_me a child (NOTE: BinaryOps won't be a child since it will instant fuse)
    d = keep_me+c
    check_schedule(d, 2)
    check_schedule(keep_me, 0, [d])

  #@unittest.skip("failing in old lazy")
  def test_permute_breaks_fusion(self):
    a = Tensor.empty(10, 10, 10)
    b = Tensor.empty(10, 10)
    c = (a.sum(axis=2) + b).permute(1,0)
    d = c.permute(1,0)
    check_schedule(d, 1)

  def test_some_permute_fusion(self):
    a = Tensor.empty(8192, 16)
    b = Tensor.empty(1, 16)
    d = (a.T + b.expand(8192, 16).T)
    c = a + b.expand(8192, 16)
    e = d.T
    check_schedule(c, 1)
    check_schedule(e, 1)

  def test_shrink_fuse(self):
    a = Tensor.empty(8192, 16)
    b = Tensor.empty(8192, 16)
    c = a * b
    d = Tensor.empty(1, 16)
    e = c[0] * d
    check_schedule(e, 1)

  def test_expand_nofuse(self):
    a = Tensor.empty(1, 16)
    b = Tensor.empty(1, 16)
    c = a * b
    d = Tensor.empty(8192, 16)
    e = c * d
    check_schedule(e, 2)

  # this is the failing case in openpilot...it's very simple like this
  @unittest.skip("failing in old lazy")
  def test_image_conv_fusion(self):
    from tinygrad.features.image import image_conv2d
    w1 = Tensor.empty(16, 16, 1, 1)
    b1 = Tensor.empty(16)
    w2 = Tensor.empty(16, 16, 1, 1)
    b2 = Tensor.empty(16)
    w3 = Tensor.empty(16, 16, 1, 1)
    b3 = Tensor.empty(16)

    x = Tensor.empty(1, 16, 32, 32)
    x = base = image_conv2d(x, w1, b1)
    x = image_conv2d(x, w2, b2) + base
    x = image_conv2d(x, w3, b3)

    # NOOP, 3 convs, contiguous
    check_schedule(x, 5)

  def test_image_conv_fusion_minimal(self):
    b1 = Tensor.empty(16)
    b2 = Tensor.empty(16)
    def p(x): return x.permute(1,0).contiguous().reshape(32,16,1).expand(32,16,16).sum(axis=2).permute(1,0)

    x = Tensor.empty(16, 32)
    x = base = p(x) + b1.reshape(16,1)
    x = p(x)
    x = x + b2.reshape(16,1)
    x = x + base
    del base
    x = p(x)
    check_schedule(x, 4)

  def test_image_conv_fusion_more_minimal(self):
    b1 = Tensor.empty(16)
    def p(x): return x.permute(1,0).contiguous().reshape(32,16,1).expand(32,16,16).sum(axis=2).permute(1,0)

    x = Tensor.empty(16, 32)
    x = base = p(x) + b1.reshape(16,1)
    x = p(x)
    del base
    check_schedule(x, 3)

  def test_resnet_block(self):
    Tensor.training = False

    in_planes, planes = 64, 64
    conv1 = nn.Conv2d(in_planes, planes, kernel_size=3, stride=1, padding=1, bias=False)
    bn1 = nn.BatchNorm2d(planes)
    conv2 = nn.Conv2d(planes, planes, kernel_size=3, padding=1, stride=1, bias=False)
    bn2 = nn.BatchNorm2d(planes)

    x = Tensor.empty(1, 64, 32, 32)
    out = bn1(conv1(x)).relu()
    out = bn2(conv2(out))
    out = (out + x).relu()
    check_schedule(out, 2, [conv1.weight, conv2.weight])

  def test_contiguous_while_contiguous(self):
    x = Tensor.empty(1, 64, 32, 32)
    out = x.contiguous()
    check_schedule(out, 1, filter_loadops=False)

  def test_contiguous_while_not_contiguous(self):
    x = Tensor.empty(1, 64, 32, 32)
    out = x.permute(0,2,3,1).contiguous()
    check_schedule(out, 2, filter_loadops=False)

  def test_double_from(self):
    x = Tensor([1,2,3,4])
    out = x.to('npy')
    check_schedule(out, 0, filter_loadops=False)

  def test_pow_const_tensor_simplified(self):
    x = Tensor([1,2,3,4])
    # NOTE: this does not test ** Tensor(2) is simpler in ast than ** Tensor(2.5)
    out = x ** Tensor(2)
    check_schedule(out, 1)

  def test_pow_const_tensor_to_zero(self):
    x = Tensor([1,2,3,4])
    out = x ** Tensor(0)
    # NOTE: this is ConstBuffer 0 + ConstBuffer 1
    check_schedule(out, 0)

  def test_zero_size(self):
    x = Tensor.empty(2, 3, 0)
    out = x + 1
    check_schedule(out, 0, filter_loadops=False)

  def test_reduce_permute_nofuse(self):
    x = Tensor.empty(32, 32, 32)
    y = Tensor.empty(32, 32)
    out = x.sum(axis=2).T+y
    check_schedule(out, 2)

  def test_two_elus_sum(self):
    x = Tensor.empty(32, 32)
    y = Tensor.empty(32, 32)
    out = x.sum(1).relu().elu() + y.sum(1).relu().elu()
    check_schedule(out, 2)

  def test_multistage_reduce(self):
    x = Tensor.empty(32, 32, 32)
    out = x.sum(2).relu().sum(1)
    check_schedule(out, 2)

  def test_multistage_reduce_fork(self):
    x = Tensor.empty(32, 32, 32)
    x = x.sum(2)
    out2 = x + 1
    out = x.relu().sum(1) + out2[0]
    check_schedule(out, 2)

  def test_example_matmul(self):
    x = Tensor.eye(64, requires_grad=True)
    y = Tensor.eye(64, requires_grad=True)
    z = y.matmul(x).sum()
    z.backward()
    out = x.grad.contiguous()
    check_schedule(out, 2)

  def test_contiguous_add(self):
    x = Tensor.empty(32)
    y = Tensor.empty(32)
    z = Tensor.empty(32)
    out = (x+y).contiguous()+z
    check_schedule(out, 2)

  def test_double_sum_ref(self):
    x = Tensor.empty(32, 32, 32)
    x = x.sum(2)
    out = x + x[:, 4]
    check_schedule(out, 2)

  def test_reduce_shrink(self):
    x = Tensor.empty(32, 32)
    y = Tensor.empty(16)
    x = x.sum(1)
    x = x[:16]
    out = x + y
    check_schedule(out, 2)  # TODO: this should be 1

  @unittest.skip("broken due to const folding and two contiguous are different kernels")
  def test_const_no_recompute(self):
    x = Tensor(2) + Tensor(2)
    y = Tensor(2) + Tensor(2)
    out = x.contiguous() + y.contiguous()
    check_schedule(out, 2)

  def test_reduce_same_size(self):
    a = Tensor.empty(4, 4)
    out0 = a.sum() + 2
    out1 = a.sum() + 4
    out2 = out0 * out1
    check_schedule([out0, out1, out2], 1)

  def test_reduce_multiple_paths(self):
    a = Tensor.empty(4, 4)
    out0 = a.sum().exp2()
    # out1 has two paths to a.sum()
    out1 = a.sum() + out0
    check_schedule([out0, out1], 1)

  def test_reduce_ext_reduce_child(self):
    a = Tensor.empty((4, 4))
    b = Tensor.empty((4, 4))
    # b.sum() is not a descendant of the fused nodes
    out0 = a.sum() + b.sum() + 2
    out1 = a.sum() + b.sum() + 4
    check_schedule([out0, out1], 4)

  def test_reduce_multiple_paths_midreduce(self):
    a = Tensor.empty(4, 4)
    r = a.sum()
    out0 = r.exp2()
    # reduce node in the indirect path from r to out2
    out1 = (a - out0).max()
    out2 = r + out1
    check_schedule([r, out0, out1, out2], 4)

  def test_reduce_multiple_paths_midreduce_fused(self):
    a = Tensor.empty(4, 4)
    b = Tensor.empty(4, 4)
    out0 = a.sum() + 4
    out1 = b.max() + out0*2
    out2 = a.sum() + out1
    check_schedule([out0, out1, out2], 4)

  def test_reduce_multiple_paths_midexpand(self):
    a = Tensor.empty(4, 4)
    b = Tensor.empty(4, 4, 4)
    r = a.sum()
    out0 = r.exp2()
    # e1 is in the indirect path from a.sum() to out1
    e = b + out0
    out1 = r + e[0][0][0]
    check_schedule([r, out0, out1, e], 4)

  def test_reduce_expand_child(self):
    a = Tensor.empty((32, 32, 32))
    b = Tensor.empty((1, 16))
    out0 = a.sum() + 2
    out1 = a.sum() + b
    check_schedule([out0, out1], 4)

  def test_reduce_shrink_child(self):
    a = Tensor.empty(100, 100)
    b = Tensor.empty(10,)
    c = a.sum() + b[0]
    d = a.sum() + 2
    check_schedule([c, d], 1)

  def test_reduce_multiple_paths_midshrink(self):
    a = Tensor.empty(4, 4)
    r = a.sum(axis=1)
    out0 = r.exp2()
    out1 = out0[0] + out0
    check_schedule([r, out0, out1], 3)

  def test_reduce_shrink_output(self):
    a = Tensor.empty(4, 4)
    r = a.sum(keepdim=True)
    out0 = r.exp2()
    out1 = out0[0] + Tensor.empty(1, )
    check_schedule([r, out0, out1], 3)

  def test_softmax_fusion(self):
    out = Tensor.empty(4, 12, 64, 64).softmax()
    check_schedule(out, 3)

  def test_layernorm_onelayer_fusion(self):
    layer = nn.LayerNorm([10, 10])
    x = Tensor.empty(20, 5, 10, 10)
    check_schedule(layer(x), 3)

  def test_scaled_dot_product_attention_fusion(self):
    x, y, z, m = (Tensor.empty(32, 8, 16, 16) for _ in range(4))
    out = Tensor.scaled_dot_product_attention(x, y, z, attn_mask=m)
    check_schedule(out, 5)

  def test_scaled_dot_product_attention_causal_fusion(self):
    x, y, z, m = (Tensor.empty(32, 8, 16, 16) for _ in range(4))
    out = Tensor.scaled_dot_product_attention(x, y, z, attn_mask=m, is_causal=True)
    check_schedule(out, 7)

  def test_adam_step_fusion(self):
    with Tensor.train():
      x = Tensor.empty(4, 64, 768)
      layer = nn.Linear(768, 768*4)
      opt = nn.optim.Adam(nn.state.get_parameters(layer), lr=1e-4)
      layer(x).relu().sum().backward()
      check_schedule(opt.schedule_step(), 11)

  def test_adam_conv_fuse(self):
    with Tensor.train():
      img = Tensor.empty(2,3,4,4)
      c1 = nn.Conv2d(3,32,3)
      opt = nn.optim.Adam(nn.state.get_parameters(c1), lr=1e-4)
      opt.zero_grad()
      c1(img).relu().sum().backward()
      check_schedule(opt.schedule_step(), 11)

  def test_adam_2convs_fuse(self):
    with Tensor.train():
      img = Tensor.empty(2,3,4,4)
      c1 = nn.Conv2d(3,16,3,bias=False)
      c2 = nn.Conv2d(16,32,3,bias=False)
      opt = nn.optim.Adam(nn.state.get_parameters([c1, c2]), lr=1e-4)
      opt.zero_grad()
      c2(c1(img).relu()).relu().sum().backward()
      check_schedule(opt.schedule_step(), 13)

  def test_sgd_conv_fuse(self):
    with Tensor.train():
      img = Tensor.empty(2,3,4,4)
      c1 = nn.Conv2d(3,32,3)
      opt = nn.optim.SGD(nn.state.get_parameters(c1))
      opt.zero_grad()
      c1(img).relu().sum().backward()
      check_schedule(opt.schedule_step(), 7)

  def test_sgd_2convs_fuse(self):
    with Tensor.train():
      img = Tensor.empty(2,3,4,4)
      c1 = nn.Conv2d(3,16,3,bias=False)
      c2 = nn.Conv2d(16,32,3,bias=False)
      opt = nn.optim.SGD(nn.state.get_parameters([c1, c2]))
      opt.zero_grad()
      c2(c1(img).relu()).relu().sum().backward()
      check_schedule(opt.schedule_step(), 7)

  def test_fold_2convs_sgd_nesterov_momentum_wd(self):
    with Tensor.train():
      img = Tensor.empty(2,3,4,4)
      c1 = nn.Conv2d(3,16,3,bias=False)
      c2 = nn.Conv2d(16,32,3,bias=False)
      opt = nn.optim.SGD(nn.state.get_parameters([c1, c2]), nesterov=True, momentum=0.9, weight_decay=0.1)
      opt.zero_grad()
      c2(c1(img).relu()).relu().sum().backward()
      check_schedule(opt.schedule_step(), 9)

  def test_sgd_4convs_fuse(self):
    with Tensor.train():
      img = Tensor.empty(2,3,64,64)
      c1 = nn.Conv2d(3,4,3,bias=False)
      c2 = nn.Conv2d(4,8,3,bias=False)
      c3 = nn.Conv2d(8,16,3,bias=False)
      c4 = nn.Conv2d(16,32,3,bias=False)
      opt = nn.optim.SGD(nn.state.get_parameters([c1, c2, c3, c4]))
      opt.zero_grad()
      c4(c3(c2(c1(img).relu()).relu()).relu()).relu().sum().backward()
      check_schedule(opt.schedule_step(), 22)

  @unittest.skipUnless(is_dtype_supported(dtypes.half), "need half")
  def test_prefer_half_buffer(self):
    x = Tensor.ones(4).contiguous().realize()
    # y = Tensor.ones(4).contiguous().realize()
    z = Tensor.ones(4, 4).contiguous().realize()

    # should not create extra kernel if output will be realized anyways
    dummy = x.sum().half().float()
    check_schedule(dummy, 1)
    dummy = x.sum().half().float().contiguous() + 1
    check_schedule(dummy, 2)

    # shared between two outputs
    shared = x.sum().half().float()
    a = shared * 2
    b = shared * 3
    sched = check_schedule([a, b], 1)
    for si in sched[:-2]: assert all(out.dtype is dtypes.half for out in si.outputs)

    # reduce
    a = z.sum(axis=0).half().float().sum(axis=0)
    sched = check_schedule(a, 2)
    for si in sched[:-1]: assert all(out.dtype is dtypes.half for out in si.outputs)

    # expand
    # expand will realize just after the .float(), so requires change to realize-before-expand
    # normal = (x.sum().half().float().reshape(1) * y).sum()
    # sched = check_schedule(normal, 2)
    # for si in sched[:-1]: assert all(out.dtype == dtypes.half for out in si.outputs[:-1])

    # parallel reduce
    # a = x.sum().half().float() * y.sum().half().float()
    # b = a + 1
    # c = a + 2
    # sched = check_schedule([b, c], 4)
    # doesn't store either in half because it doesn't chase

  def test_reduce_simple_chase(self):
    a = Tensor.empty(4, 4, 4)
    r = a.sum(0) + 6
    b = r.sum(0) * 4
    c = r.sum(1) * 2
    schedule = check_schedule([b, c], 3)
    assert schedule[0].ast[0].src[0].op is BinaryOps.ADD

  def test_push_permute_chase(self):
    a = Tensor.empty(4, 4, 4)
    b = Tensor.empty(4, 4)
    r = a.sum(2) + b
    d = r.T * 4
    e = r * d
    schedule = check_schedule([d, e], 3)
    assert schedule[0].ast[0].src[0].op is BinaryOps.ADD

  def test_push_shrink_chase(self):
    a = Tensor.empty(16, 16)
    b = Tensor.empty(4)
    c = Tensor.empty(16, )
    r = a.sum(1) + c
    d = r[:4] * b
    schedule = check_schedule(d, 2)
    assert schedule[0].ast[0].src[0].op is BinaryOps.ADD

  def test_midreduce_nochase(self):
    a = Tensor.empty(16, 16)
    b = (a.sum(0) + a.max(1)) + 2
    schedule = check_schedule(b, 2)
    assert schedule[0].ast[0].src[0].op is ReduceOps.MAX

  # pattern in test_transformer
  def test_partial_fuse1(self):
    a = Tensor.empty(16, 16)
    b = Tensor.empty(16, 16)
    c = a.sum() + 2
    d = (a.sum() - b.sum()) * 4
    check_schedule([c, d], 3)

  # pattern in conv
  def test_partial_fuse2(self):
    a = Tensor.empty(16, 16)
    b = Tensor.empty(16, 16)
    c = a.sum() + 2
    d = b.sum() - c
    check_schedule([c, d], 2)

  # pattern in adam
  def test_partial_fuse3(self):
    a = Tensor.empty(16, 16)
    b = Tensor.empty(16, 16)
    c = a.sum() + 2
    d = a.sum() * 2
    e = c * d
    f = b.sum() - e
    check_schedule([c, d, e, f], 2)

  def test_partial_fuse4(self):
    a = Tensor.empty(16, 16)
    b = Tensor.empty(16, 16)
    c = a.sum() + 2
    d = a.sum() * 2
    e = c * d
    f = (b - d).sum() - e
    check_schedule([c, d, e, f], 3)

if __name__ == '__main__':
  unittest.main(verbosity=2)