Launch many tasks on the same GPU
If you want to use a powerful GPU efficiently, you can run many tasks on same GPU
using a combination of sbatch arguments. In your sbatch script:
Specify only 1 GPU to use, e.g. with
--gres=gpu:rtx8000:1Specify number of tasks to run on the selected GPU with
--ntasks-per-gpu=NLaunch your job using
srun main.pyinstead of justmain.py.
srun will then launch main.py script N times.
Each task will receive specific environment variables, such as SLURM_PROCID,
which you can then use to parameterize the script execution.
Prerequisites Make sure to read the following sections of the documentation before using this example:
The full source code for this example is available on the mila-docs GitHub repository.
job.sh
# distributed/single_gpu/job.sh -> good_practices/many_tasks_per_gpu/job.sh
#!/bin/bash
-#SBATCH --ntasks=1
-#SBATCH --ntasks-per-node=1
+#SBATCH --ntasks=2
+#SBATCH --ntasks-per-gpu=2
#SBATCH --cpus-per-task=4
-#SBATCH --gpus-per-task=l40s:1
+#SBATCH --gres=gpu:l40s:1
#SBATCH --mem-per-gpu=16G
#SBATCH --time=00:15:00
# Exit on error
set -e
# Echo time and hostname into log
echo "Date: $(date)"
echo "Hostname: $(hostname)"
# To make your code as much reproducible as possible with
# `torch.use_deterministic_algorithms(True)`, uncomment the following block:
## === Reproducibility ===
## Be warned that this can make your code slower. See
## https://pytorch.org/docs/stable/notes/randomness.html#cublas-and-cudnn-deterministic-operations
## for more details.
# export CUBLAS_WORKSPACE_CONFIG=:4096:8
## === Reproducibility (END) ===
# Stage dataset into $SLURM_TMPDIR
mkdir -p $SLURM_TMPDIR/data
cp /network/datasets/cifar10/cifar-10-python.tar.gz $SLURM_TMPDIR/data/
# General-purpose alternatives combining copy and unpack:
# unzip /network/datasets/some/file.zip -d $SLURM_TMPDIR/data/
# tar -xf /network/datasets/some/file.tar -C $SLURM_TMPDIR/data/
# Execute Python script
# Use the `--offline` option of `uv run` on clusters without internet access on compute nodes.
# Using the `--locked` option can help make your experiments easier to reproduce (it forces
# your uv.lock file to be up to date with the dependencies declared in pyproject.toml).
srun uv run python main.py
main.py
# distributed/single_gpu/main.py -> good_practices/many_tasks_per_gpu/main.py
-"""Single-GPU training example."""
+"""Many tasks per GPU (job packing) example."""
import argparse
import logging
import os
import random
import sys
from pathlib import Path
import numpy as np
import rich.logging
import torch
from torch import Tensor, nn
from torch.nn import functional as F
from torch.utils.data import DataLoader, random_split
from torchvision import transforms
from torchvision.datasets import CIFAR10
from torchvision.models import resnet18
from tqdm import tqdm
# To make your code as much reproducible as possible, uncomment the following
# block:
## === Reproducibility ===
## Be warned that this can make your code slower. See
## https://pytorch.org/docs/stable/notes/randomness.html#cublas-and-cudnn-deterministic-operations
## for more details.
# torch.use_deterministic_algorithms(True)
## === Reproducibility (END) ===
def main():
# Use an argument parser so we can pass hyperparameters from the command line.
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("--epochs", type=int, default=10)
parser.add_argument("--learning-rate", type=float, default=5e-4)
parser.add_argument("--weight-decay", type=float, default=1e-4)
parser.add_argument("--batch-size", type=int, default=128)
- parser.add_argument("--seed", type=int, default=42)
+ # Get SLURM_PROCID and use it as a random seed for the script.
+ # This makes it so each task within a job uses a different initialization with the same
+ # hyper-parameters.
+ parser.add_argument(
+ "--seed",
+ type=int,
+ default=int(os.environ.get("SLURM_PROCID", 0)),
+ help="Random seed used for network initialization and the training loop.",
+ )
args = parser.parse_args()
epochs: int = args.epochs
learning_rate: float = args.learning_rate
weight_decay: float = args.weight_decay
batch_size: int = args.batch_size
seed: int = args.seed
# Seed the random number generators as early as possible for reproducibility
random.seed(seed)
np.random.seed(seed)
torch.random.manual_seed(seed)
torch.cuda.manual_seed_all(seed)
# Check that the GPU is available
assert torch.cuda.is_available() and torch.cuda.device_count() > 0
device = torch.device("cuda", 0)
# Setup logging (optional, but much better than using print statements)
# Uses the `rich` package to make logs pretty.
logging.basicConfig(
level=logging.INFO,
format="%(message)s",
handlers=[
rich.logging.RichHandler(
markup=True,
console=rich.console.Console(
# Allower wider log lines in sbatch output files than on the terminal.
width=120 if not sys.stdout.isatty() else None
),
)
],
)
logger = logging.getLogger(__name__)
# Create a model and move it to the GPU.
model = resnet18(num_classes=10)
model.to(device=device)
optimizer = torch.optim.AdamW(
model.parameters(), lr=learning_rate, weight_decay=weight_decay
)
# Setup CIFAR10
num_workers = get_num_workers()
dataset_path = Path(os.environ.get("SLURM_TMPDIR", ".")) / "data"
train_dataset, valid_dataset, test_dataset = make_datasets(str(dataset_path))
train_dataloader = DataLoader(
train_dataset,
batch_size=batch_size,
num_workers=num_workers,
shuffle=True,
)
valid_dataloader = DataLoader(
valid_dataset,
batch_size=batch_size,
num_workers=num_workers,
shuffle=False,
)
_test_dataloader = DataLoader( # NOTE: Not used in this example.
test_dataset,
batch_size=batch_size,
num_workers=num_workers,
shuffle=False,
)
# Checkout the "checkpointing and preemption" example for more info!
logger.debug("Starting training from scratch.")
for epoch in range(epochs):
logger.debug(f"Starting epoch {epoch}/{epochs}")
# Set the model in training mode (important for e.g. BatchNorm and Dropout layers)
model.train()
# NOTE: using a progress bar from tqdm because it's nicer than using `print`.
progress_bar = tqdm(
total=len(train_dataloader),
desc=f"Train epoch {epoch}",
disable=not sys.stdout.isatty(), # Disable progress bar in non-interactive environments.
)
# Training loop
for batch in train_dataloader:
# Move the batch to the GPU before we pass it to the model
batch = tuple(item.to(device) for item in batch)
x, y = batch
# Forward pass
logits: Tensor = model(x)
loss = F.cross_entropy(logits, y)
optimizer.zero_grad()
loss.backward()
optimizer.step()
# Calculate some metrics:
n_correct_predictions = logits.detach().argmax(-1).eq(y).sum()
n_samples = y.shape[0]
accuracy = n_correct_predictions / n_samples
logger.debug(f"Accuracy: {accuracy.item():.2%}")
logger.debug(f"Average Loss: {loss.item()}")
# Advance the progress bar one step and update the progress bar text.
progress_bar.update(1)
progress_bar.set_postfix(loss=loss.item(), accuracy=accuracy.item())
progress_bar.close()
val_loss, val_accuracy = validation_loop(model, valid_dataloader, device)
logger.info(
f"Epoch {epoch}: Val loss: {val_loss:.3f} accuracy: {val_accuracy:.2%}"
)
print("Done!")
@torch.no_grad()
def validation_loop(model: nn.Module, dataloader: DataLoader, device: torch.device):
model.eval()
total_loss = 0.0
n_samples = 0
correct_predictions = 0
for batch in dataloader:
batch = tuple(item.to(device) for item in batch)
x, y = batch
logits: Tensor = model(x)
loss = F.cross_entropy(logits, y)
batch_n_samples = x.shape[0]
batch_correct_predictions = logits.argmax(-1).eq(y).sum()
total_loss += loss.item()
n_samples += batch_n_samples
correct_predictions += batch_correct_predictions
accuracy = correct_predictions / n_samples
return total_loss, accuracy
def make_datasets(
dataset_path: str,
val_split: float = 0.1,
val_split_seed: int = 42,
):
"""Returns the training, validation, and test splits for CIFAR10.
NOTE: We don't use image transforms here for simplicity.
Having different transformations for train and validation would complicate things a bit.
Later examples will show how to do the train/val/test split properly when using transforms.
"""
train_dataset = CIFAR10(
root=dataset_path, transform=transforms.ToTensor(), download=True, train=True
)
test_dataset = CIFAR10(
root=dataset_path, transform=transforms.ToTensor(), download=True, train=False
)
# Split the training dataset into a training and validation set.
n_samples = len(train_dataset)
n_valid = int(val_split * n_samples)
n_train = n_samples - n_valid
train_dataset, valid_dataset = random_split(
train_dataset, (n_train, n_valid), torch.Generator().manual_seed(val_split_seed)
)
return train_dataset, valid_dataset, test_dataset
def get_num_workers() -> int:
"""Gets the optimal number of DatLoader workers to use in the current job."""
if "SLURM_CPUS_PER_TASK" in os.environ:
return int(os.environ["SLURM_CPUS_PER_TASK"])
if hasattr(os, "sched_getaffinity"):
return len(os.sched_getaffinity(0))
return torch.multiprocessing.cpu_count()
if __name__ == "__main__":
main()
Running this example
You can launch this example with sbatch:
$ sbatch job.sh