Skip to content

parser.py doesn't work without Nvidia / CUDA #3

Description

@JuKu

Hi together,

I do a research project about MPI / OpenMPI 5.0.8 (with Peruse) and we try to trace things with ucTrace.
I used a slurm system without Nvidia / CUDA with 8 Nodes and ran LULESH 2.0 LULESH 2.0 with it:

#!/bin/bash
#SBATCH --nodes=1
#SBATCH --tasks-per-node=8
#SBATCH --cpus-per-task=1
#SBATCH --mem-per-cpu=2G
#SBATCH --output=lulesh-%j.out
#SBATCH --error=lulesh-%j.err
export LD_PRELOAD=/home/<USERNAME>/git/so-files-lib/preload.so
# This folder contains the built .so files from ucTrace folder
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/home/<USERNAME>/git/so-files-lib/
srun /home/<USERNAME>/git/LULESH/lulesh2.0

Everything worked and I got the 8 .asd.gz files.

After that, I tried to execute parse.py to parse the log files:

python3 parse.py my-outputs/ -p my_output.pkl -n 0

Badly, I got an error while the script was parsing the COMMS:

Traceback (most recent call last):
  File "/home/<USERNAME>/ucTrace/parser/parse.py", line 2017, in <module>
    main()
    ~~~~^^
  File "/home/<USERNAME>/ucTrace/parser/parse.py", line 1966, in main
    all_data, parsed_comms = parse_folder(args.folder_path, args)
                             ~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/<USERNAME>/ucTrace/parser/parse.py", line 1793, in parse_folder
    match_ucp_send_recv(parsed_comms, all_data)
    ~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/<USERNAME>/ucTrace/parser/parse.py", line 1253, in match_ucp_send_recv
    recvs = tmp_recvs[target_pid]
            ~~~~~~~~~^^^^^^^^^^^^
KeyError: "Error decoding hex string-- 'ascii' codec can't decode byte 0xf0 in position 8"

How can I fix this?
Thanks in advance!

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Fields

    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions