[topi] Add arm_cpu specific pooling schedules

[FranklandJack]

This commit:

• Adds specialized arm_cpu pooling schedules for both fixed width and salable vectors.
• Enables topi testing of new arm_cpu schedules.

[tvm-bot]

Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from Reviewers by @-ing them in a comment.

• No users to tag found in teams: topi _{See #10317 for details}

_{Generated by tvm-bot}

[lhutton1]

Thanks @FranklandJack, this looks very nice addition! I left some small comments and questions below

[lhutton1]

Will leave open for a bit longer to give chance for any final comments

[FranklandJack]

Will leave open for a bit longer to give chance for any final comments

Thanks @lhutton1. This has been open for almost two months now, so hopefully everyone who it might affect has had a chance to see it. I haven't got write permissions, so feel free to merge when you think its ready.

[junrushao]

NOTE: this PR causes a regression in importing order:

Traceback (most recent call last):
  File "/root/Projects/tvm-dev/main.py", line 10, in <module>
    from tvm.testing import local_run
  File "/root/Projects/tvm-dev/python/tvm/testing/__init__.py", line 21, in <module>
    from . import auto_scheduler, autotvm
  File "/root/Projects/tvm-dev/python/tvm/testing/auto_scheduler.py", line 20, in <module>
    from tvm import auto_scheduler, te, topi
  File "/root/Projects/tvm-dev/python/tvm/topi/__init__.py", line 51, in <module>
    from . import nn
  File "/root/Projects/tvm-dev/python/tvm/topi/nn/__init__.py", line 35, in <module>
    from .conv3d_transpose import *
  File "/root/Projects/tvm-dev/python/tvm/topi/nn/conv3d_transpose.py", line 21, in <module>
    from tvm import relay
  File "/root/Projects/tvm-dev/python/tvm/relay/__init__.py", line 29, in <module>
    from . import prelude
  File "/root/Projects/tvm-dev/python/tvm/relay/prelude.py", line 21, in <module>
    from tvm.relay.transform import ToANormalFormExpr
  File "/root/Projects/tvm-dev/python/tvm/relay/transform/__init__.py", line 22, in <module>
    from . import fake_quantization_to_integer, mixed_precision
  File "/root/Projects/tvm-dev/python/tvm/relay/transform/fake_quantization_to_integer.py", line 25, in <module>
    from tvm.relay.qnn.op import canonicalizations
  File "/root/Projects/tvm-dev/python/tvm/relay/qnn/__init__.py", line 20, in <module>
    from . import op
  File "/root/Projects/tvm-dev/python/tvm/relay/qnn/op/__init__.py", line 21, in <module>
    from .qnn import *
  File "/root/Projects/tvm-dev/python/tvm/relay/qnn/op/qnn.py", line 26, in <module>
    from tvm.relay.op.nn.utils import get_pad_tuple2d
  File "/root/Projects/tvm-dev/python/tvm/relay/op/__init__.py", line 35, in <module>
    from . import strategy
  File "/root/Projects/tvm-dev/python/tvm/relay/op/strategy/__init__.py", line 24, in <module>
    from . import arm_cpu
  File "/root/Projects/tvm-dev/python/tvm/relay/op/strategy/arm_cpu.py", line 29, in <module>
    from ....topi.arm_cpu.mprofile import dsp
  File "/root/Projects/tvm-dev/python/tvm/topi/arm_cpu/__init__.py", line 33, in <module>
    from . import qnn_alter_op
  File "/root/Projects/tvm-dev/python/tvm/topi/arm_cpu/qnn_alter_op.py", line 25, in <module>
    from ..nn import qnn_conv2d_alter_layout, add_alter_layout, qnn_requantize_alter_layout
ImportError: cannot import name 'qnn_conv2d_alter_layout' from partially initialized module 'tvm.topi.nn' (most likely due to a circular import) (/root/Projects/tvm-dev/python/tvm/topi/nn/__init__.py)

[lhutton1]

Thanks @junrushao for the quick catch. I had a quick look into this but I wasn't able to reproduce the issue. Do you have a small script I could try out?