Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

ENH: runtime HEATMAP/DXT cross-check log #22

Merged
merged 1 commit into from
Feb 18, 2022
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
Original file line number Diff line number Diff line change
@@ -0,0 +1,58 @@
This is an example darshan log that contains both runtime
heatmap module data and DXT tracing, to facilitate both
visual and robust numerical comparison of the two approaches
of generating heatmaps in the new Python summary report.

The log follows the following "graphical" IO pattern
of writing only a single byte of data per rank, for each
of 32 ranks, on staggered/exaggerated delays:

Time Bins -->
Rank
|
|
V
31 x
30 x
29 x
28 x
..
3 x
2 x
1 x
0 x

The example application code used to generate the log
is available here:
https://github.com/tylerjereddy/heatmap_diagonal

In particular, commit hash `4712f54` from that project
was used, and the precise timing and related details
may be obtained from that project. In short, 32 bytes (1 per rank)
were written at minor timing offsets in a Python mpi4py-based
module on a single node of the LANL snow supercomputer
(Intel skylake gold). Only, POSIX, DXT_POSIX, and
(runtime) HEATMAP modules should be present. There is
no *read* activity in the log (just 1 byte written
per rank).

The commit hash of the darshan main branch used to build
the darshan runtime was `b4932a02a`.

Here is the command line that was used on a
snow backend node:
`mpirun -x DXT_ENABLE_IO_TRACE=1 -x DARSHAN_LOGPATH=/yellow/users/treddy/generated_darshan_logs -x DARSHAN_DISABLE_SHARED_REDUCTION=1 -x LD_PRELOAD=/yellow/users/treddy/darshan_install/lib/libdarshan.so -np 32 python heatmap_diagonal.py`

Here are the steps there were followed to build
the `darshan-runtime` on a snow backend:

```
module load gcc/9.3.0
module load openmpi/3.1.6
module load hdf5-parallel/1.10.7
./prepare.sh
cd darshan-runtime/
./configure --prefix=/yellow/users/treddy/darshan_install --with-log-path-by-env=DARSHAN_LOGPATH --with-jobid-env=SLURM_JOBID --enable-hdf5-mod --with-hdf5=/usr/projects/hpcsoft/toss3/snow/hdf5/1.10.7_gcc-9.3.0_openmpi-3.1.6 CC=mpicc
make
make install
```
Empty file.
Binary file not shown.