I removed the stale data/shot_lists/**/*.txt files from this repository in #50 (specifically 035f3402b1048003aca5c7363cad6a710c36b186).
They were typically unused because most clusters and supercomputers that we use work best when the input/output data from compute nodes is located on a separate file system from the home directories (usually NFS) where we typically clone the FRNN source code, e.g. GPFS for /tigress/FRNN/ and Lustre /lus/theta-fs0/projects/fusiondl_aesp/ for ALCF machines.
These files are also generally read-only, similar to the raw shot data .txt files, so it seems inappropriate to keep them alongside the .py files.
However, we probably should version control the shot list .txt files for safety and bookkeeping the changes.
[x] Start a new repository under the @PPPLDeepLearning organization just for these files?
[ ] Replace /tigress/FRNN/shot_lists/ with a clone of the new repo to get the .git/ metadata.
I removed the stale
data/shot_lists/**/*.txt
files from this repository in #50 (specifically 035f3402b1048003aca5c7363cad6a710c36b186).They were typically unused because most clusters and supercomputers that we use work best when the input/output data from compute nodes is located on a separate file system from the home directories (usually NFS) where we typically clone the FRNN source code, e.g. GPFS for
/tigress/FRNN/
and Lustre/lus/theta-fs0/projects/fusiondl_aesp/
for ALCF machines.These files are also generally read-only, similar to the raw shot data
.txt
files, so it seems inappropriate to keep them alongside the.py
files.However, we probably should version control the shot list
.txt
files for safety and bookkeeping the changes./tigress/FRNN/shot_lists/
with a clone of the new repo to get the.git/
metadata.