I build a decompiler to convert bytecode generated by dynamo into readable source code!

compiler
1

Hi, folks, if you are also suffering from reading bytecode generated by dynamo, you can try this out!

Simple usage with dynamo:

First, run a pytorch program with torch.compile:

from typing import List
import torch
from torch import _dynamo as torchdynamo
def my_compiler(gm: torch.fx.GraphModule, example_inputs: List[torch.Tensor]):
    print("my_compiler() called with FX graph:")
    gm.graph.print_tabular()
    return gm.forward  # return a python callable

@torchdynamo.optimize(my_compiler)
def toy_example(a, b):
    x = a / (torch.abs(a) + 1)
    if b.sum() < 0:
        b = b * -1
    return x * b
for _ in range(100):
    toy_example(torch.randn(10), torch.randn(10))

Second, get compiled code and guard code from pytorch:

from torch._dynamo.eval_frame import _debug_get_cache_entry_list
cache_entries = _debug_get_cache_entry_list(toy_example._torchdynamo_orig_callable.__code__)
guard, code = cache_entries[0]

Third, decompile the code to see how the code works:

from depyf import decompile

print("guard code:")
print(decompile(guard))

print("compiled code:")
print(decompile(code))

Output on my computer:

guard code:
def guard(L):
    if not getattr(___guarded_code, 'valid'):
        return False
    else:
        _var0 = L['a']
        __temp_1 = hasattr(_var0, '_dynamo_dynamic_indices')
        if not (__temp_1 == False):
            return False
        else:
            _var1 = L['b']
            __temp_2 = hasattr(_var1, '_dynamo_dynamic_indices')
            if not (__temp_2 == False):
                return False
            else:
                __temp_3 = ___is_grad_enabled()
                if not __temp_3:
                    return False
                else:
                    __temp_4 = ___are_deterministic_algorithms_enabled()
                    if __temp_4:
                        return False
                    else:
                        __temp_5 = ___is_torch_function_enabled()
                        if not __temp_5:
                            return False
                        else:
                            if not (getattr(utils_device, 'CURRENT_DEVICE') == None):
                                return False
                            else:
                                __temp_6 = ___check_tensors(_var0, _var1, tensor_check_names=tensor_check_names)
                                if not __temp_6:
                                    return False
                                else:
                                    return True

compiled code:
def toy_example(a, b):
    __temp_1 = __compiled_fn_0(a, b)
    x = __temp_1[0]
    if __temp_1[1]:
        __temp_2 = __resume_at_30_1(b, x)
        return __temp_2
    else:
        __temp_3 = __resume_at_38_2(b, x)
        return __temp_3

Hopefully, by using this package, you can understand python bytecode now!

2 Likes
2

@jansel I build this mainly for the doc we write :smile: Is it good to add this package to the doc?

3

Neat! Yeah, feel free to update the docs.

4

The code is better and more readable now, I eliminated some temp variables:

guard code:
def guard(L):
    if not getattr(___guarded_code, 'valid'):
        return False
    else:
        _var0 = L['a']
        if not hasattr(_var0, '_dynamo_dynamic_indices') == False:
            return False
        else:
            _var1 = L['b']
            if not hasattr(_var1, '_dynamo_dynamic_indices') == False:
                return False
            elif not ___is_grad_enabled():
                return False
            elif ___are_deterministic_algorithms_enabled():
                return False
            elif not ___is_torch_function_enabled():
                return False
            elif not getattr(utils_device, 'CURRENT_DEVICE') == None:
                return False
            elif not ___check_tensors(_var0, _var1, tensor_check_names=
                tensor_check_names):
                return False
            else:
                return True

compiled code:
def toy_example(a, b):
    __temp_1 = __compiled_fn_0(a, b)
    x = __temp_1[0]
    if __temp_1[1]:
        return __resume_at_30_1(b, x)
    else:
        return __resume_at_38_2(b, x)
1 Like
5

@jansel PR created at [Doc] Update the dynamo deepdive doc by youkaichao · Pull Request #108147 · pytorch/pytorch · GitHub . And I add support for python 3.7 – 3.11 today. The depyf package should be good to use for understanding dynamo bytecode now. It has almost everything except while loops and for loops, which I suppose rarely occur in dynamo.

1 Like