Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

I couldn't run code "env.fork()" #826

Open
Rain0802 opened this issue Jan 26, 2024 · 0 comments
Open

I couldn't run code "env.fork()" #826

Rain0802 opened this issue Jan 26, 2024 · 0 comments
Labels
Question Further information is requested

Comments

@Rain0802
Copy link

When I run below code in the tutorial Chapter “Autotuning”,

# Import the compiler_gym namespace
import compiler_gym

# Create an LLVM CompilerGym environment
env = compiler_gym.make("llvm-v0")

# Set the benchmark. We'll use cBench/susan
env.reset(benchmark = "benchmark://cbench-v1/susan")

# We'll use IrInstructionCount as reward and observation
env.observation_space = "IrInstructionCount"
env.reward_space = "IrInstructionCount"
print(f"Total instructions at start = {env.observation['IrInstructionCount']}")

# Here we do our hill climbing. We'll do one hundred steps.
num_steps = 100

for i in range(num_steps):
  # We will need to back track if the action isn't good for us.
  # In CompilerGym, you can `fork` an environment and kill it later if you 
  # don't like it. That will allow us tohave the same effect as having `back`.
  candidate = env.fork()

  # We will choose a rancdom action from the action space
  action = candidate.action_space.sample()

  # And now apply it to the program
  observation, reward, done, info = candidate.step(action)

  print(f"{i}: Action {action} = {candidate.action_space.names[action]}", end="")
  print(f" gave reward {reward} for total instructions {observation}", end="")
  print(f". Accept = {reward > 0}")

  if reward <= 0:
    # If we don't like it we can throw it away.
    candidate.close()
  else:
    # If we like it, we can replace `env` with it.
    env.close()
    env = candidate

I met such reported errors in the console,

/home/rain/anaconda3/lib/python3.11/site-packages/gym/utils/passive_env_checker.py:195: UserWarning: WARN: The result returned by `env.reset()` was not a tuple of the form `(obs, info)`, where `obs` is a observation and `info` is a dictionary containing additional information. Actual type: `<class 'str'>`
  logger.warn(
Total instructions at start = 13282
Traceback (most recent call last):
  File "/home/rain/PycharmProjects/CompilerGymPractice/main.py", line 61, in <module>
    candidate = env.fork()
                ^^^^^^^^^^
  File "/home/rain/anaconda3/lib/python3.11/site-packages/compiler_gym/envs/compiler_env.py", line 599, in fork
    new_env.reward.spaces = deepcopy(self.reward.spaces)
                            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rain/anaconda3/lib/python3.11/copy.py", line 146, in deepcopy
    y = copier(x, memo)
        ^^^^^^^^^^^^^^^
  File "/home/rain/anaconda3/lib/python3.11/copy.py", line 231, in _deepcopy_dict
    y[deepcopy(key, memo)] = deepcopy(value, memo)
                             ^^^^^^^^^^^^^^^^^^^^^
  File "/home/rain/anaconda3/lib/python3.11/copy.py", line 172, in deepcopy
    y = _reconstruct(x, memo, *rv)
        ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rain/anaconda3/lib/python3.11/copy.py", line 273, in _reconstruct
    y.__setstate__(state)
  File "/home/rain/anaconda3/lib/python3.11/site-packages/gym/spaces/space.py", line 125, in __setstate__
    state = dict(state)
            ^^^^^^^^^^^
TypeError: cannot convert dictionary update sequence element #0 to a sequence
ERROR: Killing a service with 2 active sessions!
Exception ignored in: <function CompilerEnv.__del__ at 0x7f8bebe89a80>
Traceback (most recent call last):
  File "/home/rain/anaconda3/lib/python3.11/site-packages/compiler_gym/envs/compiler_env.py", line 670, in __del__
  File "/home/rain/anaconda3/lib/python3.11/site-packages/compiler_gym/envs/compiler_env.py", line 661, in close
  File "/home/rain/anaconda3/lib/python3.11/site-packages/compiler_gym/service/connection.py", line 681, in close
  File "/home/rain/anaconda3/lib/python3.11/site-packages/compiler_gym/service/connection.py", line 457, in close
compiler_gym.service.connection.ServiceError: Service exited with returncode 6

It seems that something goes wrong with "env.fork()", could anybody tell me why, thank you!

@Rain0802 Rain0802 added the Question Further information is requested label Jan 26, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Question Further information is requested
Projects
None yet
Development

No branches or pull requests

1 participant