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

separate CPU/GPU target longrun configs #852

Merged
merged 1 commit into from
Jun 13, 2024
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.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
28 changes: 14 additions & 14 deletions .buildkite/longruns/pipeline.yml
Original file line number Diff line number Diff line change
Expand Up @@ -254,9 +254,9 @@ steps:
soft_fail: true

- label: "MPI AMIP FINE: new target amip: topo + diagedmf"
key: "amip_target_topo_diagedmf"
command: "srun julia --color=yes --project=experiments/ClimaEarth/ experiments/ClimaEarth/run_amip.jl --config_file $CONFIG_PATH/amip_target_topo_diagedmf.yml --job_id amip_target_topo_diagedmf"
artifact_paths: "experiments/ClimaEarth/output/amip/amip_target_topo_diagedmf_artifacts/*"
key: "amip_target_topo_diagedmf_cpu"
command: "srun julia --color=yes --project=experiments/ClimaEarth/ experiments/ClimaEarth/run_amip.jl --config_file $CONFIG_PATH/amip_target_topo_diagedmf_cpu.yml --job_id amip_target_topo_diagedmf_cpu"
artifact_paths: "experiments/ClimaEarth/output/amip/amip_target_topo_diagedmf_cpu_artifacts/*"
env:
CLIMACORE_DISTRIBUTED: "MPI"
BUILD_HISTORY_HANDLE: ""
Expand All @@ -271,20 +271,20 @@ steps:
steps:

- label: "GPU AMIP FINE: new target amip: topo"
key: "gpu_amip_target_topo"
command: "srun julia --color=yes --project=experiments/ClimaEarth/ experiments/ClimaEarth/run_amip.jl --config_file $CONFIG_PATH/amip_target_topo.yml --job_id gpu_amip_target_topo"
artifact_paths: "experiments/ClimaEarth/output/amip/gpu_amip_target_topo_artifacts/*"
key: "amip_target_topo_gpu"
command: "srun julia --color=yes --project=experiments/ClimaEarth/ experiments/ClimaEarth/run_amip.jl --config_file $CONFIG_PATH/amip_target_topo.yml --job_id amip_target_topo_gpu"
artifact_paths: "experiments/ClimaEarth/output/amip/amip_target_topo_gpu_artifacts/*"
agents:
slurm_gpus: 1
slurm_mem: 16GB
soft_fail: true

- label: "GPU AMIP FINE: new target amip: topo + diagedmf"
key: "gpu_amip_target_topo_diagedmf"
key: "amip_target_topo_diagedmf_gpu"
command:
- echo "--- Run simulation"
- "srun --cpu-bind=threads --cpus-per-task=4 julia --threads=3 --color=yes --project=experiments/ClimaEarth/ experiments/ClimaEarth/run_amip.jl --config_file $CONFIG_PATH/amip_target_topo_diagedmf.yml --job_id gpu_amip_target_topo_diagedmf"
artifact_paths: "experiments/ClimaEarth/output/amip/gpu_amip_target_topo_diagedmf_artifacts/*"
- "srun --cpu-bind=threads --cpus-per-task=4 julia --threads=3 --color=yes --project=experiments/ClimaEarth/ experiments/ClimaEarth/run_amip.jl --config_file $CONFIG_PATH/amip_target_topo_diagedmf.yml --job_id amip_target_topo_diagedmf_gpu"
artifact_paths: "experiments/ClimaEarth/output/amip/amip_target_topo_diagedmf_gpu_artifacts/*"
timeout_in_minutes: 1440
env:
CLIMACOMMS_CONTEXT: "MPI"
Expand All @@ -299,11 +299,11 @@ steps:

# DYAMOND AMIP: 1 day (convection resolving)
- label: "GPU AMIP SUPERFINE: dyamond_target"
key: "gpu_longrun_amip_dyamond"
key: "longrun_amip_dyamond_gpu"
command:
- echo "--- Run simulation"
- "julia --color=yes --project=experiments/ClimaEarth/ experiments/ClimaEarth/run_amip.jl --config_file $CONFIG_PATH/longrun_amip_dyamond.yml --job_id gpu_longrun_amip_dyamond"
artifact_paths: "experiments/ClimaEarth/output/amip/gpu_longrun_amip_dyamond_artifacts/*"
- "julia --color=yes --project=experiments/ClimaEarth/ experiments/ClimaEarth/run_amip.jl --config_file $CONFIG_PATH/longrun_amip_dyamond.yml --job_id longrun_amip_dyamond_gpu"
artifact_paths: "experiments/ClimaEarth/output/amip/longrun_amip_dyamond_gpu_artifacts/*"
agents:
queue: clima
slurm_mem: 20GB
Expand Down Expand Up @@ -338,9 +338,9 @@ steps:

- label: ":envelope: Slack report: target AMIP"
command:
- slack-upload -c "#coupler-report" -f experiments/ClimaEarth/output/amip/gpu_amip_target_topo_diagedmf_artifacts/amip_ncep.png -m png -n amip_fine -x "300d Target AMIP v NCEP Last Month Mean"
- slack-upload -c "#coupler-report" -f experiments/ClimaEarth/output/amip/amip_target_topo_diagedmf_gpu_artifacts/amip_ncep.png -m png -n amip_fine -x "300d Target AMIP v NCEP Last Month Mean"
- |
find experiments/ClimaEarth/output/amip/gpu_amip_target_topo_diagedmf_artifacts/ -type f -name 'bias*.png' -print0 | while IFS= read -r -d '' file; do
find experiments/ClimaEarth/output/amip/amip_target_topo_diagedmf_gpu_artifacts/ -type f -name 'bias*.png' -print0 | while IFS= read -r -d '' file; do
slack-upload -c "#coupler-report" -f "$$file" -m png -n "$$(basename "$$file" .png)" -x "$$(basename "$$file" .png)"
done
agents:
Expand Down
23 changes: 23 additions & 0 deletions config/longrun_configs/amip_target_topo_diagedmf_gpu.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,23 @@
FLOAT_TYPE: "Float32"
albedo_model: "CouplerAlbedo"
anim: false
atmos_config_file: "config/longrun_configs/amip_target_diagedmf.yml"
dt: "120secs"
dt_cpl: 120
dt_save_state_to_disk: "20days"
dt_save_to_sol: "10days"
energy_check: false
hourly_checkpoint: false
land_albedo_type: "map_temporal"
mode_name: "amip"
mono_surface: false
netcdf_output_at_levels: true
output_default_diagnostics: true
start_date: "20100101"
surface_setup: "PrescribedSurface"
t_end: "300days"
topo_smoothing: true
topography: "Earth"
coupler_toml_file: "toml/amip_target_topo_diagedmf.toml"
turb_flux_partition: "CombinedStateFluxesMOST"
use_reference_state: false
Loading