You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Describe the bug
Previously generating (0.14) code is now failing to parse triggering a
networkx.exception.NetworkXError: nbunch is not a node or a sequence of nodes.
This is with dace:gpu backend.
To Reproduce
I couldn't recreate the error in a pure DaCe setting, so the above uses a regression test from the Pace respository. It is the smallest test that triggers the above bug
# Repo is to run the FiniteVolumeTransport regression test
# Original code: fv3core/pace/fv3core/stencils/fvtp2d.py
# DaCe is applied on the __call__ function
# Get Pace repository
git clone -b update/gt4py_dace git@github.com:GEOS-ESM/pace
cd pace
git submodule init
git submodule update
# Setup the venv, including Cupy for GPU
python -m venv .venv
source .venv/bin/activate
pip install --upgrade pip
pip install external/gt4py/
pip install -r requirements_dev.txt -c constraints.txt
pip install cupy-cuda12x # REPLACED with relevant CUDA
# Download data
mkdir -p test_data/8.1.3/c12_6ranks_standard/dycore
cd test_data/8.1.3/c12_6ranks_standard/dycore
pip install gdown
gdown https://drive.google.com/uc?id=1Fz2KSWwmy4cpIjYzXKJbkdzsVwHgZuok
gdown https://drive.google.com/uc?id=1GJRYQxRAkQcHbBPvF5KVngrYjrTt1Pdw
gdown https://drive.google.com/uc?id=1SOO97ncz-fCGVoPD7pUuYY9uYowUjato
gdown https://drive.google.com/uc?id=1Wcb1l7GXE5C_82oItGo7RkKloWlJCKR4
cd -
# Run test of FvTp2d
export FV3_DACEMODE=BuildAndRun
export PACE_CONSTANTS=GFS
pytest -v -s --data_path=./test_data/8.1.3/c12_6ranks_standard/dycore \
--backend=dace:gpu --which_modules=FvTp2d --which_rank=0 \
--threshold_overrides_file=./fv3core/tests/savepoint/translate/overrides/standard.yaml \
./fv3core/tests/savepoint
tested with python 3.8.10
Expected behavior
The above code runs a full regression test, expected to pass.
The text was updated successfully, but these errors were encountered:
#1314 has a fix for this, although another underlying issue might exist. The problem can be seen in SDFG subgraph below. In state assign_1059_4,q_x_flux writes to another array, which is implied to be "return" data (name ends in __ret_N) and, therefore, should not be transient. Could you verify if this is OK (the test seems to pass with the fix), i.e., the actual return data is q_x_flux?
All 75 smaller regressions tests pass with #1314, that's a coverage of all non-mpi elements of the model. Previously ~10% where seeing this bugs - looks like it's fixed nicely!
Describe the bug
Previously generating (0.14) code is now failing to parse triggering a
This is with
dace:gpu
backend.To Reproduce
I couldn't recreate the error in a pure DaCe setting, so the above uses a regression test from the Pace respository. It is the smallest test that triggers the above bug
tested with python 3.8.10
Expected behavior
The above code runs a full regression test, expected to pass.
The text was updated successfully, but these errors were encountered: