gem5-users@gem5.org

The gem5 Users mailing list

View all threads

MemoryError: std::bad_alloc error with BFS benchmark

MM
Mirco Mannino
Wed, Sep 20, 2023 2:59 PM

Hi all,

I am trying to run BFS benchmark from the GAP benchmark suite
(https://github.com/sbeamer/gapbs) using GEM5 SE.
If I use an input graph with up to 16 million vertices everything works
fine.
However, when I use an input graph with more than 16 million vertices I
obtain the following error:

**** REAL SIMULATION ****
src/sim/simulate.cc:194: info: Entering event queue @ 0.  Starting
simulation...
src/sim/syscall_emul.cc:85: warn: ignoring syscall set_robust_list(...)
      (further warnings will be suppressed)
src/sim/mem_state.cc:443: info: Increasing stack size by one page.
src/sim/syscall_emul.cc:74: warn: ignoring syscall mprotect(...)
src/sim/mem_state.cc:443: info: Increasing stack size by one page.
src/sim/mem_state.cc:443: info: Increasing stack size by one page.
MemoryError: std::bad_alloc

At:
  src/python/m5/simulate.py(201): simulate
  /workspace/gem5/configs/common/Simulation.py(279): benchCheckpoints
  /workspace/gem5/configs/common/Simulation.py(807): run
  ./configs/example/se.py(300): <module>
  src/python/m5/main.py(630): main

I have tried invoking gem5.debug with GDB, but I don’t get a backtrace.
On my machine the benchmark runs correctly.

Has anyone ever had similar problems using relatively large workloads?

PS. I'm aware that I'm using the old "se.py," but I believe this is not
the cause of the problem.

Thank you in advance,
Mirco Mannino

Hi all, I am trying to run BFS benchmark from the GAP benchmark suite (https://github.com/sbeamer/gapbs) using GEM5 SE. If I use an input graph with up to 16 million vertices everything works fine. However, when I use an input graph with more than 16 million vertices I obtain the following error: **** REAL SIMULATION **** src/sim/simulate.cc:194: info: Entering event queue @ 0.  Starting simulation... src/sim/syscall_emul.cc:85: warn: ignoring syscall set_robust_list(...)       (further warnings will be suppressed) src/sim/mem_state.cc:443: info: Increasing stack size by one page. src/sim/syscall_emul.cc:74: warn: ignoring syscall mprotect(...) src/sim/mem_state.cc:443: info: Increasing stack size by one page. src/sim/mem_state.cc:443: info: Increasing stack size by one page. MemoryError: std::bad_alloc At:   src/python/m5/simulate.py(201): simulate   /workspace/gem5/configs/common/Simulation.py(279): benchCheckpoints   /workspace/gem5/configs/common/Simulation.py(807): run   ./configs/example/se.py(300): <module>   src/python/m5/main.py(630): main I have tried invoking gem5.debug with GDB, but I don’t get a backtrace. On my machine the benchmark runs correctly. Has anyone ever had similar problems using relatively large workloads? PS. I'm aware that I'm using the old "se.py," but I believe this is not the cause of the problem. Thank you in advance, Mirco Mannino