TimeoutError when runing MLIR example

sofan110 commented 1 year ago

❓ Questions and Help

I use compiler_gym 0.2.5 installed with MLIR env and try to execute the example code

import gym from compiler_gym.wrappers.mlir import make_mlir_rl_wrapper_env env = gym.make("mlir-v0") wrapper = make_mlir_rl_wrapper_env(env) wrapper.reset() observation, reward, done, info = wrapper.step(wrapper.action_space.sample())

but I get the following error:

RuntimeError: {'error_type': 'TimeoutError', 'error_details': "Command '$CC -Ofast /dev/shm/compiler_gym_wufan/s/0404T150542-962363-5241/scratch-42e8-a740/out.mlir.ll -mllvm -enable-matrix -mllvm -matrix-allow-contract -mllvm -matrix-default-layout=row-major -c -o /dev/shm/compiler_gym_wufan/s/0404T150542-962363-5241/scratch-42e8-a740/benchmark_mlir.o /workspace/S/wufan/conda/anaconda3/envs/MLIR/lib/python3.8/site-packages/compiler_gym/envs/mlir/service/llvm/lib/libMLIRExecutionEngine.a -g3 && $CC -o /dev/shm/compiler_gym_wufan/s/0404T150542-962363-5241/scratch-42e8-a740/benchmark_main -I/workspace/S/wufan/conda/anaconda3/envs/MLIR/lib/python3.8/site-packages/compiler_gym/envs/mlir/service/llvm/include/ -I/workspace/S/wufan/conda/anaconda3/envs/MLIR/lib/python3.8/site-packages/compiler_gym/envs/mlir/service/google_benchmark/include/ -lstdc++ -lm -pthread -v -g3 /dev/shm/compiler_gym_wufan/s/0404T150542-962363-5241/scratch-42e8-a740/benchmark_main.cc /dev/shm/compiler_gym_wufan/s/0404T150542-962363-5241/scratch-42e8-a740/benchmark_mlir.o /workspace/S/wufan/conda/anaconda3/envs/MLIR/lib/python3.8/site-packages/compiler_gym/envs/mlir/service/google_benchmark/lib/libbenchmark.a' failed to complete within 120 seconds (300.0 seconds)"}

ChrisCummins commented 1 year ago

Hi @sofan110, let me CC the authors of the MLIR environment as they may be able to help you out: @KyleHerndon, @sogartar.

In the meantime, could you please add the extra system information needed to diagnose issues (click on "bug report" here for the template).

Cheers, Chris

sofan110 commented 1 year ago

🐛 Bug

TimeoutError when running mlir_rl_wrapper_env.step()

To Reproduce

Steps to reproduce the behavior:

install compilergym with cmake
run the demo code

Environment

CompilerGym: 0.2.5 Is debug build: N/A

Python version: 3.8.16 (default, Mar 2 2023, 03:21:46) [GCC 11.2.0] (64-bit runtime) Python platform: Linux-3.10.0-1127.13.1.el7.x86_64-x86_64-with-glibc2.17

OS: CentOS Linux 7 (Core) (x86_64) GCC version: (GCC) 9.3.0 Clang version: Could not collect CMake version: 3.21.7 Libc version: glibc-2.17

Versions of all installed libraries: [pip3] absl-py==1.4.0 [pip3] asttokens==2.2.1 [pip3] backcall==0.2.0 [pip3] certifi==2022.12.7 [pip3] charset-normalizer==3.1.0 [pip3] cloudpickle==2.2.1 [pip3] comm==0.1.2 [pip3] compiler-gym==0.2.5 [pip3] debugpy==1.6.6 [pip3] decorator==5.1.1 [pip3] Deprecated==1.2.13 [pip3] docker==6.0.1 [pip3] executing==1.2.0 [pip3] fasteners==0.18 [pip3] grpcio==1.43.0 [pip3] grpcio-tools==1.43.0 [pip3] gym==0.21.0 [pip3] humanize==4.6.0 [pip3] idna==3.4 [pip3] importlib-metadata==6.1.0 [pip3] ipykernel==6.21.3 [pip3] ipython==8.11.0 [pip3] jedi==0.18.2 [pip3] jupyter_client==8.0.3 [pip3] jupyter_core==5.3.0 [pip3] loop-tool-py==0.0.6 [pip3] matplotlib-inline==0.1.6 [pip3] nest-asyncio==1.5.6 [pip3] networkx==3.0 [pip3] numpy==1.24.2 [pip3] packaging==23.0 [pip3] parso==0.8.3 [pip3] pexpect==4.8.0 [pip3] pickleshare==0.7.5 [pip3] pip==23.0.1 [pip3] platformdirs==3.1.1 [pip3] prompt-toolkit==3.0.38 [pip3] protobuf==3.20.3 [pip3] psutil==5.9.4 [pip3] ptyprocess==0.7.0 [pip3] pure-eval==0.2.2 [pip3] pybind11==2.10.4 [pip3] pydantic==1.10.6 [pip3] Pygments==2.14.0 [pip3] python-dateutil==2.8.2 [pip3] pyzmq==25.0.1 [pip3] requests==2.28.2 [pip3] setuptools==65.6.3 [pip3] six==1.16.0 [pip3] stack-data==0.6.2 [pip3] tabulate==0.9.0 [pip3] tornado==6.2 [pip3] traitlets==5.9.0 [pip3] typing_extensions==4.5.0 [pip3] urllib3==1.26.15 [pip3] wcwidth==0.2.6 [pip3] websocket-client==1.5.1 [pip3] wheel==0.38.4 [pip3] wrapt==1.15.0 [pip3] zipp==3.15.0 [conda] # [conda] # Name Version Build Channel [conda] _libgcc_mutex 0.1 main defaults [conda] _openmp_mutex 5.1 1_gnu defaults [conda] absl-py 1.4.0 pypi_0 pypi [conda] asttokens 2.2.1 pypi_0 pypi [conda] backcall 0.2.0 pypi_0 pypi [conda] ca-certificates 2023.01.10 h06a4308_0 defaults [conda] certifi 2022.12.7 py38h06a4308_0 defaults [conda] charset-normalizer 3.1.0 pypi_0 pypi [conda] cloudpickle 2.2.1 pypi_0 pypi [conda] comm 0.1.2 pypi_0 pypi [conda] compiler-gym 0.2.5 pypi_0 pypi [conda] debugpy 1.6.6 pypi_0 pypi [conda] decorator 5.1.1 pypi_0 pypi [conda] deprecated 1.2.13 pypi_0 pypi [conda] docker 6.0.1 pypi_0 pypi [conda] executing 1.2.0 pypi_0 pypi [conda] fasteners 0.18 pypi_0 pypi [conda] grpcio 1.43.0 pypi_0 pypi [conda] grpcio-tools 1.43.0 pypi_0 pypi [conda] gym 0.21.0 pypi_0 pypi [conda] humanize 4.6.0 pypi_0 pypi [conda] idna 3.4 pypi_0 pypi [conda] importlib-metadata 6.1.0 pypi_0 pypi [conda] ipykernel 6.21.3 pypi_0 pypi [conda] ipython 8.11.0 pypi_0 pypi [conda] jedi 0.18.2 pypi_0 pypi [conda] jupyter-client 8.0.3 pypi_0 pypi [conda] jupyter-core 5.3.0 pypi_0 pypi [conda] ld_impl_linux-64 2.38 h1181459_1 defaults [conda] libffi 3.4.2 h6a678d5_6 defaults [conda] libgcc-ng 11.2.0 h1234567_1 defaults [conda] libgomp 11.2.0 h1234567_1 defaults [conda] libstdcxx-ng 11.2.0 h1234567_1 defaults [conda] loop-tool-py 0.0.6 pypi_0 pypi [conda] matplotlib-inline 0.1.6 pypi_0 pypi [conda] ncurses 6.4 h6a678d5_0 defaults [conda] nest-asyncio 1.5.6 pypi_0 pypi [conda] networkx 3.0 pypi_0 pypi [conda] numpy 1.24.2 pypi_0 pypi [conda] openssl 1.1.1t h7f8727e_0 defaults [conda] packaging 23.0 pypi_0 pypi [conda] parso 0.8.3 pypi_0 pypi [conda] pexpect 4.8.0 pypi_0 pypi [conda] pickleshare 0.7.5 pypi_0 pypi [conda] pip 23.0.1 py38h06a4308_0 defaults [conda] platformdirs 3.1.1 pypi_0 pypi [conda] prompt-toolkit 3.0.38 pypi_0 pypi [conda] protobuf 3.20.3 pypi_0 pypi [conda] psutil 5.9.4 pypi_0 pypi [conda] ptyprocess 0.7.0 pypi_0 pypi [conda] pure-eval 0.2.2 pypi_0 pypi [conda] pybind11 2.10.4 pypi_0 pypi [conda] pydantic 1.10.6 pypi_0 pypi [conda] pygments 2.14.0 pypi_0 pypi [conda] python 3.8.16 h7a1cb2a_3 defaults [conda] python-dateutil 2.8.2 pypi_0 pypi [conda] pyzmq 25.0.1 pypi_0 pypi [conda] readline 8.2 h5eee18b_0 defaults [conda] requests 2.28.2 pypi_0 pypi [conda] setuptools 65.6.3 py38h06a4308_0 defaults [conda] six 1.16.0 pypi_0 pypi [conda] sqlite 3.41.1 h5eee18b_0 defaults [conda] stack-data 0.6.2 pypi_0 pypi [conda] tabulate 0.9.0 pypi_0 pypi [conda] tk 8.6.12 h1ccaba5_0 defaults [conda] tornado 6.2 pypi_0 pypi [conda] traitlets 5.9.0 pypi_0 pypi [conda] typing-extensions 4.5.0 pypi_0 pypi [conda] urllib3 1.26.15 pypi_0 pypi [conda] wcwidth 0.2.6 pypi_0 pypi [conda] websocket-client 1.5.1 pypi_0 pypi [conda] wheel 0.38.4 py38h06a4308_0 defaults [conda] wrapt 1.15.0 pypi_0 pypi [conda] xz 5.2.10 h5eee18b_1 defaults [conda] zipp 3.15.0 pypi_0 pypi

Additional context

I just remember that I use CMAKE_C_COMPILER=gcc & CMAKE_CXX_COMPILER=g++, not clang & clang++ mentioned in the documentation. Is it the key point?

facebookresearch / CompilerGym