Compile dynamic does not support GroupNorm in module

[PeterL1n]

🐛 Describe the bug

If my module simply wraps group norm, it does not compile with dynamic=True.

import torch
from torch import nn

class MyModule(nn.Module):
    def __init__(self):
        super().__init__()
        self.norm = nn.GroupNorm(32, 512)
    
    def forward(self, x):
        return self.norm(x)

model = MyModule().cuda().eval()
model = torch.compile(model, dynamic=True)

x = torch.randn([1, 512, 32, 32], device="cuda")
y = model(x)

Error:

Traceback (most recent call last):
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/output_graph.py", line 670, in call_user_compiler
    compiled_fn = compiler_fn(gm, self.fake_example_inputs())
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/debug_utils.py", line 1055, in debug_wrapper
    compiled_gm = compiler_fn(gm, example_inputs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/__init__.py", line 1390, in __call__
    return compile_fx(model_, inputs_, config_patches=self.config)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_inductor/compile_fx.py", line 401, in compile_fx
    return compile_fx(
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_inductor/compile_fx.py", line 455, in compile_fx
    return aot_autograd(
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/backends/common.py", line 48, in compiler_fn
    cg = aot_module_simplified(gm, example_inputs, **kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_functorch/aot_autograd.py", line 2805, in aot_module_simplified
    compiled_fn = create_aot_dispatcher_function(
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/utils.py", line 163, in time_wrapper
    r = func(*args, **kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_functorch/aot_autograd.py", line 2498, in create_aot_dispatcher_function
    compiled_fn = compiler_fn(flat_fn, fake_flat_args, aot_config)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_functorch/aot_autograd.py", line 1713, in aot_wrapper_dedupe
    return compiler_fn(flat_fn, leaf_flat_args, aot_config)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_functorch/aot_autograd.py", line 2087, in aot_dispatch_autograd
    fx_g = make_fx(joint_forward_backward, aot_config.decompositions)(
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/fx/experimental/proxy_tensor.py", line 714, in wrapped
    t = dispatch_trace(wrap_key(func, args, fx_tracer), tracer=fx_tracer, concrete_args=tuple(phs))
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/eval_frame.py", line 209, in _fn
    return fn(*args, **kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/fx/experimental/proxy_tensor.py", line 443, in dispatch_trace
    graph = tracer.trace(root, concrete_args)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/eval_frame.py", line 209, in _fn
    return fn(*args, **kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/fx/_symbolic_trace.py", line 778, in trace
    (self.create_arg(fn(*args)),),
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/fx/_symbolic_trace.py", line 652, in flatten_fn
    tree_out = root_fn(*tree_args)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/fx/experimental/proxy_tensor.py", line 459, in wrapped
    out = f(*tensors)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_functorch/aot_autograd.py", line 1156, in traced_joint
    return functionalized_f_helper(primals, tangents)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_functorch/aot_autograd.py", line 1108, in functionalized_f_helper
    f_outs = flat_fn_no_input_mutations(fn, f_primals, f_tangents, meta, keep_input_mutations)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_functorch/aot_autograd.py", line 1076, in flat_fn_no_input_mutations
    outs = flat_fn_with_synthetic_bases_expanded(fn, primals, primals_after_cloning, maybe_tangents, meta, keep_input_mutations)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_functorch/aot_autograd.py", line 1048, in flat_fn_with_synthetic_bases_expanded
    outs = forward_or_joint(fn, primals_before_cloning, primals, maybe_tangents, meta, keep_input_mutations)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_functorch/aot_autograd.py", line 1017, in forward_or_joint
    backward_out = torch.autograd.grad(
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/autograd/__init__.py", line 269, in grad
    return handle_torch_function(
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/overrides.py", line 1534, in handle_torch_function
    result = mode.__torch_function__(public_api, types, args, kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_inductor/overrides.py", line 38, in __torch_function__
    return func(*args, **kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/autograd/__init__.py", line 303, in grad
    return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/utils/_stats.py", line 20, in wrapper
    return fn(*args, **kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/fx/experimental/proxy_tensor.py", line 487, in __torch_dispatch__
    return self.inner_torch_dispatch(func, types, args, kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/fx/experimental/proxy_tensor.py", line 512, in inner_torch_dispatch
    out = proxy_call(self, func, args, kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/fx/experimental/proxy_tensor.py", line 248, in proxy_call
    r = CURRENT_DECOMPOSITION_TABLE[func](*args, **kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_decomp/decompositions.py", line 70, in inner
    r = f(*tree_map(increase_prec, args), **tree_map(increase_prec, kwargs))
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_decomp/decompositions.py", line 1171, in native_group_norm_backward
    cpg, _rem = divmod(C, group)
TypeError: unsupported operand type(s) for divmod(): 'SymInt' and 'int'

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "/opt/tiger/code/test.py", line 16, in <module>
    y = model(x)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
    return forward_call(*args, **kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/eval_frame.py", line 82, in forward
    return self.dynamo_ctx(self._orig_mod.forward)(*args, **kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/eval_frame.py", line 209, in _fn
    return fn(*args, **kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/eval_frame.py", line 337, in catch_errors
    return callback(frame, cache_size, hooks)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/convert_frame.py", line 404, in _convert_frame
    result = inner_convert(frame, cache_size, hooks)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/convert_frame.py", line 104, in _fn
    return fn(*args, **kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/convert_frame.py", line 262, in _convert_frame_assert
    return _compile(
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/utils.py", line 163, in time_wrapper
    r = func(*args, **kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/convert_frame.py", line 324, in _compile
    out_code = transform_code_object(code, transform)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/bytecode_transformation.py", line 445, in transform_code_object
    transformations(instructions, code_options)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/convert_frame.py", line 311, in transform
    tracer.run()
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/symbolic_convert.py", line 1726, in run
    super().run()
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/symbolic_convert.py", line 576, in run
    and self.step()
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/symbolic_convert.py", line 540, in step
    getattr(self, inst.opname)(inst)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/symbolic_convert.py", line 1792, in RETURN_VALUE
    self.output.compile_subgraph(
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/output_graph.py", line 517, in compile_subgraph
    self.compile_and_call_fx_graph(tx, list(reversed(stack_values)), root)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/output_graph.py", line 588, in compile_and_call_fx_graph
    compiled_fn = self.call_user_compiler(gm)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/utils.py", line 163, in time_wrapper
    r = func(*args, **kwargs)
  File "/home/tiger/.local/lib/python3.9/site-packages/torch/_dynamo/output_graph.py", line 675, in call_user_compiler
    raise BackendCompilerFailed(self.compiler_fn, e) from e
torch._dynamo.exc.BackendCompilerFailed: debug_wrapper raised TypeError: unsupported operand type(s) for divmod(): 'SymInt' and 'int'

Set torch._dynamo.config.verbose=True for more information


You can suppress this exception and fall back to eager by setting:
    torch._dynamo.config.suppress_errors = True

Versions

Collecting environment information...
PyTorch version: 2.0.0+cu117
Is debug build: False
CUDA used to build PyTorch: 11.7
ROCM used to build PyTorch: N/A

OS: Debian GNU/Linux 11 (bullseye) (x86_64)
GCC version: (Debian 10.2.1-6) 10.2.1 20210110
Clang version: Could not collect
CMake version: version 3.25.0
Libc version: glibc-2.31

Python version: 3.9.2 (default, Feb 28 2021, 17:03:44)  [GCC 10.2.1 20210110] (64-bit runtime)
Python platform: Linux-5.4.143.bsk.7-amd64-x86_64-with-glibc2.31
Is CUDA available: True
CUDA runtime version: 11.7.99
CUDA_MODULE_LOADING set to: LAZY
GPU models and configuration: GPU 0: A100-SXM-80GB
Nvidia driver version: 450.191.01
cuDNN version: Probably one of the following:
/usr/lib/x86_64-linux-gnu/libcudnn.so.8.5.0
/usr/lib/x86_64-linux-gnu/libcudnn_adv_infer.so.8.5.0
/usr/lib/x86_64-linux-gnu/libcudnn_adv_train.so.8.5.0
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_infer.so.8.5.0
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_train.so.8.5.0
/usr/lib/x86_64-linux-gnu/libcudnn_ops_infer.so.8.5.0
/usr/lib/x86_64-linux-gnu/libcudnn_ops_train.so.8.5.0
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True

CPU:
Architecture:                    x86_64
CPU op-mode(s):                  32-bit, 64-bit
Byte Order:                      Little Endian
Address sizes:                   46 bits physical, 57 bits virtual
CPU(s):                          128
On-line CPU(s) list:             0-127
Thread(s) per core:              2
Core(s) per socket:              32
Socket(s):                       2
NUMA node(s):                    2
Vendor ID:                       GenuineIntel
CPU family:                      6
Model:                           106
Model name:                      Intel(R) Xeon(R) Platinum 8336C CPU @ 2.30GHz
Stepping:                        6
CPU MHz:                         3000.000
CPU max MHz:                     3500.0000
CPU min MHz:                     800.0000
BogoMIPS:                        4600.00
Virtualization:                  VT-x
L1d cache:                       3 MiB
L1i cache:                       2 MiB
L2 cache:                        80 MiB
L3 cache:                        108 MiB
NUMA node0 CPU(s):               0-31,64-95
NUMA node1 CPU(s):               32-63,96-127
Vulnerability Itlb multihit:     Not affected
Vulnerability L1tf:              Not affected
Vulnerability Mds:               Not affected
Vulnerability Meltdown:          Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl and seccomp
Vulnerability Spectre v1:        Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:        Mitigation; Enhanced IBRS, IBPB conditional, RSB filling
Vulnerability Srbds:             Not affected
Vulnerability Tsx async abort:   Not affected
Flags:                           fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 invpcid_single ssbd mba ibrs ibpb stibp ibrs_enhanced tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb intel_pt avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local wbnoinvd dtherm ida arat pln pts hwp hwp_act_window hwp_epp hwp_pkg_req avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg tme avx512_vpopcntdq rdpid md_clear pconfig flush_l1d arch_capabilities

Versions of relevant libraries:
[pip3] byted-torch==2.0.0.post0
[pip3] byted-torch-monitor==0.0.1
[pip3] numpy==1.24.2
[pip3] pytorch-triton==2.0.0+0d7e753227
[pip3] torch==2.0.0
[pip3] torchaudio==2.0.1
[pip3] torchvision==0.15.1
[pip3] triton==2.0.0
[conda] Could not collect

cc @ezyang @soumith @msaroufim @wconstab @ngimel @bdhirsh

[nima10khodaveisi]

I would like to work on this!

[ngimel]

Sure @nima10khodaveisi, replacing divmod line with C//group, C % group (and adding a test) should be enough. Also, can you take a look if divmod is used in any other decompositions in _refs/_prims and similarly replace it, that would be more general.

[nima10khodaveisi]

Sure @nima10khodaveisi, replacing divmod line with C//group, C % group (and adding a test) should be enough. Also, can you take a look if divmod is used in any other decompositions in _refs/_prims and similarly replace it, that would be more general.

Of course, thanks for the hint!

[ezyang]

We could also add __divmod__ on SymInt, that sounds good to me too

[ngimel]

Yeah adding divmod is even better (would handle arbitrary user code using divmod), you can look in symbolic_shapes.py how other python ops are handled with symints.

[nima10khodaveisi]

I think implementing divmod for SymInt requires lots of changes since the output of divmod is tuple which is not handled by symbolic_shapes.py functionality.
@ngimel @ezyang

[ngimel]

Ok then easy fix will do for now

[ezyang]

to be clear, I would have suggested a shallow integration, where __divmod__ internally just calls div and mod under the hood.

[ngimel]

But it would need to return 2 SymNodes?

[nima10khodaveisi]

But it would need to return 2 SymNodes?

I think so and then you have to make other functions like wrap_node working for tuple and it requires lots of changes in this case.

[tugsbayasgalan]

@nima10khodaveisi any updates?