Error while running cc benchmark from GAPBS

[akshaybaviskar]

Hello @idanyani and @magbarya,

I ran cc benchmark with Twitter dataset as input from gapbs benchmark suite.

I got following error on running it :

Traceback (most recent call last):
  File "/home/akshaybavisk/mosmodel/cc_twit_working/scripts/runBenchmark.py", line 103, in <module>
    benchmark_run.wait()
  File "/home/akshaybavisk/mosmodel/cc_twit_working/scripts/runBenchmark.py", line 51, in wait
    raise subprocess.CalledProcessError(self._run_process.returncode, ' '.join(self._run_process.args))
subprocess.CalledProcessError: Command '/home/akshaybavisk/mosmodel/cc_twit_working/scripts/measureGeneralMetrics.sh /home/akshaybavisk/mosmodel/cc_twit_working/scripts/setCpuMemoryAffinity.sh 1 /home/akshaybavisk/mosmodel/cc_twit_working/mosalloc/runMosalloc.py --library /home/akshaybavisk/mosmodel/cc_twit_working/mosalloc/src/libmosalloc.so -fps 1GB -aps 107374182400 -as1 0 -ae1 0 -as2 0 -ae2 107374182400 -bps 4194304 -bs1 0 -be1 0 -bs2 0 -be2 4194304 --analyze ./run.sh' returned non-zero exit status 127.
make: *** [/home/akshaybavisk/mosmodel/cc_twit_working/experiments/template.mk:35: experiments/single_page_size/layout2mb/repeat1/perf.out] Error 1

I have attached the complete log here

What could be the reason behind this?

Thanks,
Akshay

[magbarya]

Hi Akshay,
I have tried to reproduce this issue but unfortunately could not. The cc-twitter benchmark runs, for me, without any failures.
Actually, I noticed that the benchmark memory footprint, from your log, is incorrect for some reason.
Can you please:

1. attach the following two files:
   experiments/without_mosalloc/repeat1/time.out
   experiments/without_mosalloc/repeat1/max_mem.txt
2. try to run "make clean" and then "make" again

[akshaybaviskar]

Hello Mohammad,

Sorry for the late reply.
Please find the files attached.
max_mem.txt
time.out.txt

I get the same error if I run make clean and make again.

Thanks,
Akshay

[idanyani]

@magbarya : maybe this is the same bug of issue #16 ?
@akshaybaviskar : can you please pull the latest commit and try again?

[magbarya]

@idanyani: it could be related to #16, but I see another problem from the logs so I want to isolate the issue that Akshay observes.

@akshaybaviskar: I can see two issues by looking at the attached files.

1. I can see in the log that the benchmark footprint (the anonymous pool size when running the 2MB layout) is 100 GB, which does not make sense.
2. I can see too that the "without_mosalloc" experiment was failed (you can see the exit code in the time.out file).
   Can you please try to run the run.sh script, which runs the cc-twitter, just to make sure that it runs with no errors.