Unit tests for PointNet Model Ops

[saichandax]

Describe the bug

Input Resolution: (32,3,2500)

• 7 ttnn Conv1d ops failed.
• Among 7 failed convs, 4 Convs fails with OOM error, 2 Convs failed with the following error: RuntimeError: TT_FATAL @ ../ttnn/cpp/ttnn/operations/matmul/device/matmul_op.cpp:959: a_shape[-1] == b_shape[-2]
• The other conv fails with the error RuntimeError: shape '[32, 2500, 3]' is invalid for input of size 2560000

Input Resolution: (32,3,1024)

• 7 ttnn Conv1d ops failed.
• Among 7 failed convs, 4 Convs fails with OOM error, 2 Convs failed with the following error: RuntimeError: TT_FATAL @ ../ttnn/cpp/ttnn/operations/matmul/device/matmul_op.cpp:959: a_shape[-1] == b_shape[-2]
• The other conv fails with the error RuntimeError: shape '[32, 1024, 3]' is invalid for input of size 1048576

Input Resolution: (32,3,512)

• 6 ttnn Conv1d ops failed.
• Among 6 failed convs, 2 Convs failed with the following error: RuntimeError: TT_FATAL @ ../ttnn/cpp/ttnn/operations/matmul/device/matmul_op.cpp:959: a_shape[-1] == b_shape[-2]
• 1 conv fails with the error RuntimeError: shape '[32, 512, 3]' is invalid for input of size 524288
• 2 convs fails when output dtype is bfloat16 with RuntimeError: TT_THROW @ ../tt_metal/impl/program/program.cpp:519 tt::exception
• 1 conv fails with OOM error.

Input Resolution: (32,3,256)

• 4 ttnn Conv1d ops failed.
• Among 4 failed convs, 2 Convs failed with the following error: RuntimeError: TT_FATAL @ ../ttnn/cpp/ttnn/operations/matmul/device/matmul_op.cpp:959: a_shape[-1] == b_shape[-2]
• 1 conv fails with the error RuntimeError: shape '[32, 256, 3]' is invalid for input of size 262144
• 1 convs fails with RuntimeError: TT_THROW @ ../tt_metal/impl/program/program.cpp:519 tt::exception

Input Resolution: (32,3,128)

• 4 ttnn Conv1d ops failed.
• Among 4 failed convs, 2 Convs failed with the following error: RuntimeError: TT_FATAL @ ../ttnn/cpp/ttnn/operations/matmul/device/matmul_op.cpp:959: a_shape[-1] == b_shape[-2]
• 1 conv fails with the error RuntimeError: shape '[32, 128, 3]' is invalid for input of size 131072
• 1 convs fails when output dtype is bfloat16 with RuntimeError: TT_THROW @ ../tt_metal/impl/program/program.cpp:519 tt::exception

To Reproduce Steps to reproduce the behavior:

1. Checkout to the branch: harini/Pointnet_unit_tests
2. Run the command to unit test conv1d of resolution (32,3,2500): pytest tests/ttnn/unit_tests/operations/test_conv1d.py::test_pointnet_conv1d
3. Run the command to unit test conv1d of resolution (32,3,1024): pytest tests/ttnn/unit_tests/operations/test_conv1d.py::test_pointnet_conv1d_low_res_1024
4. Run the command to unit test conv1d of resolution (32,3,512): pytest tests/ttnn/unit_tests/operations/test_conv1d.py::test_pointnet_conv1d_low_res_512
5. Run the command to unit test conv1d of resolution (32,3,256): pytest tests/ttnn/unit_tests/operations/test_conv1d.py::test_pointnet_conv1d_low_res_256
6. Run the command to unit test conv1d of resolution (32,3,128): pytest tests/ttnn/unit_tests/operations/test_conv1d.py::test_pointnet_conv1d_low_res_128

Expected behavior Expected to pass the failing unit test configurations in conv1d.

Please complete the following environment information:

• OS: [e.g. Ubuntu 20.04]
• Device: WH n300

[dvartaniansTT]

@saichandax and @keerthana-r-mcw please also include the error message you're seeing in the issue!

[HariniMohan0102]

@dvartaniansTT Added the error message of respective input resolutions to the issue description, please check.