[1/4] Add get_device_stats to accelerator interface (#9586)

CHANGELOG.md

@@ -147,6 +147,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 - Added `PL_RECONCILE_PROCESS` environment variable to enable process reconciliation regardless of cluster environment settings ([#9389](https://github.com/PyTorchLightning/pytorch-lightning/pull/9389))
 
 
+- Added `get_device_stats` to the Accelerator Interface and added its implementation for GPU and TPU ([#9586](https://github.com/PyTorchLightning/pytorch-lightning/pull/9586))
+
+
 - Added `multifile` option to `LightningCLI` to enable/disable config save to preserve multiple files structure ([#9073](https://github.com/PyTorchLightning/pytorch-lightning/pull/9073))
 
 

pytorch_lightning/accelerators/accelerator.py

@@ -41,6 +41,7 @@ class Accelerator:
     - CPU
     - GPU
     - TPU
+    - IPU
 
     Each Accelerator gets two plugins upon initialization:
     One to handle differences from the training routine and one to handle different precisions.
@@ -422,6 +423,17 @@ class Accelerator:
         """
         return self.training_type_plugin.restore_checkpoint_after_pre_dispatch
 
+    def get_device_stats(self, device: Union[str, torch.device]) -> Dict[str, Any]:
+        """Gets stats for a given device.
+
+        Args:
+            device: device for which to get stats
+
+        Returns:
+            Dictionary of device stats
+        """
+        raise NotImplementedError
+
     def on_train_start(self) -> None:
         """Called when train begins."""
         return self.training_type_plugin.on_train_start()

pytorch_lightning/accelerators/cpu.py

@@ -11,6 +11,10 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+from typing import Any, Dict, Union
+
+import torch
+
 import pytorch_lightning as pl
 from pytorch_lightning.accelerators.accelerator import Accelerator
 from pytorch_lightning.utilities.exceptions import MisconfigurationException
@@ -29,3 +33,7 @@ class CPUAccelerator(Accelerator):
             raise MisconfigurationException(f"Device should be CPU, got {self.root_device} instead.")
 
         return super().setup(trainer)
+
+    def get_device_stats(self, device: Union[str, torch.device]) -> Dict[str, Any]:
+        """Returns dummy implementation for now."""
+        return {}

pytorch_lightning/accelerators/gpu.py

@@ -13,12 +13,16 @@
 # limitations under the License.
 import logging
 import os
+import shutil
+import subprocess
+from typing import Any, Dict, List, Union
 
 import torch
 
 import pytorch_lightning as pl
 from pytorch_lightning.accelerators.accelerator import Accelerator
 from pytorch_lightning.utilities.exceptions import MisconfigurationException
+from pytorch_lightning.utilities.imports import _TORCH_GREATER_EQUAL_1_8
 
 _log = logging.getLogger(__name__)
 
@@ -53,6 +57,83 @@ class GPUAccelerator(Accelerator):
         devices = os.getenv("CUDA_VISIBLE_DEVICES", all_gpu_ids)
         _log.info(f"LOCAL_RANK: {local_rank} - CUDA_VISIBLE_DEVICES: [{devices}]")
 
+    def get_device_stats(self, device: Union[str, torch.device]) -> Dict[str, Any]:
+        """Gets stats for the given GPU device.
+
+        Args:
+            device: GPU device for which to get stats
+
+        Returns:
+            A dictionary mapping the metrics to their values.
+
+        Raises:
+            FileNotFoundError:
+                If nvidia-smi installation not found
+        """
+        if _TORCH_GREATER_EQUAL_1_8:
+            return torch.cuda.memory_stats(device)
+        return _get_nvidia_gpu_stats(device)
+
     def teardown(self) -> None:
         super().teardown()
         self._move_optimizer_state(torch.device("cpu"))
+
+
+def _get_nvidia_gpu_stats(device: torch.device) -> Dict[str, float]:
+    """Get GPU stats including memory, fan speed, and temperature from nvidia-smi.
+
+    Args:
+        device: GPU device for which to get stats
+
+    Returns:
+        A dictionary mapping the metrics to their values.
+
+    Raises:
+        FileNotFoundError:
+            If nvidia-smi installation not found
+    """
+    gpu_stat_metrics = [
+        ("utilization.gpu", "%"),
+        ("memory.used", "MB"),
+        ("memory.free", "MB"),
+        ("utilization.memory", "%"),
+        ("fan.speed", "%"),
+        ("temperature.gpu", "°C"),
+        ("temperature.memory", "°C"),
+    ]
+    gpu_stat_keys = [k for k, _ in gpu_stat_metrics]
+    gpu_query = ",".join(gpu_stat_keys)
+
+    gpu_id = _get_gpu_id(device.index)
+    nvidia_smi_path = shutil.which("nvidia-smi")
+    if nvidia_smi_path is None:
+        raise FileNotFoundError("nvidia-smi: command not found")
+    result = subprocess.run(
+        [nvidia_smi_path, f"--query-gpu={gpu_query}", "--format=csv,nounits,noheader", f"--id={gpu_id}"],
+        encoding="utf-8",
+        stdout=subprocess.PIPE,
+        stderr=subprocess.PIPE,  # for backward compatibility with python version 3.6
+        check=True,
+    )
+
+    def _to_float(x: str) -> float:
+        try:
+            return float(x)
+        except ValueError:
+            return 0.0
+
+    s = result.stdout.strip()
+    stats = [_to_float(x) for x in s.split(", ")]
+
+    gpu_stats = {}
+    for i, (x, unit) in enumerate(gpu_stat_metrics):
+        gpu_stats[f"{x} ({unit})"] = stats[i]
+    return gpu_stats
+
+
+def _get_gpu_id(device_id: int) -> str:
+    """Get the unmasked real GPU IDs."""
+    # All devices if `CUDA_VISIBLE_DEVICES` unset
+    default = ",".join(str(i) for i in range(torch.cuda.device_count()))
+    cuda_visible_devices: List[str] = os.getenv("CUDA_VISIBLE_DEVICES", default=default).split(",")
+    return cuda_visible_devices[device_id].strip()

pytorch_lightning/accelerators/tpu.py

@@ -11,7 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-from typing import Any, Callable, Optional
+from typing import Any, Callable, Dict, Optional, Union
 
 import torch
 from torch.optim import Optimizer
@@ -61,3 +61,21 @@ class TPUAccelerator(Accelerator):
         for opt in self.optimizers:
             for p, v in opt.state.items():
                 opt.state[p] = apply_to_collection(v, torch.Tensor, move_data_to_device, self.root_device)
+
+    def get_device_stats(self, device: Union[str, torch.device]) -> Dict[str, Any]:
+        """Gets stats for the given TPU device.
+
+        Args:
+            device: TPU device for which to get stats
+
+        Returns:
+            A dictionary mapping the metrics (free memory and peak memory) to their values.
+        """
+        memory_info = xm.get_memory_info(device)
+        free_memory = memory_info["kb_free"]
+        peak_memory = memory_info["kb_total"] - free_memory
+        device_stats = {
+            "avg. free memory (MB)": free_memory,
+            "avg. peak memory (MB)": peak_memory,
+        }
+        return device_stats

tests/accelerators/test_gpu.py

@@ -0,0 +1,36 @@
+import torch
+
+from pytorch_lightning.accelerators import GPUAccelerator
+from pytorch_lightning.plugins.precision.precision_plugin import PrecisionPlugin
+from pytorch_lightning.plugins.training_type.dp import DataParallelPlugin
+from tests.helpers.runif import RunIf
+
+
+@RunIf(min_torch="1.8")
+@RunIf(min_gpus=1)
+def test_get_torch_gpu_stats(tmpdir):
+    """Test GPU get_device_stats with Pytorch >= 1.8.0."""
+    current_device = torch.device(f"cuda:{torch.cuda.current_device()}")
+    GPUAccel = GPUAccelerator(
+        training_type_plugin=DataParallelPlugin(parallel_devices=[current_device]), precision_plugin=PrecisionPlugin()
+    )
+    gpu_stats = GPUAccel.get_device_stats(current_device)
+    fields = ["allocated_bytes.all.freed", "inactive_split.all.peak", "reserved_bytes.large_pool.peak"]
+
+    for f in fields:
+        assert any(f in h for h in gpu_stats.keys())
+
+
+@RunIf(max_torch="1.7")
+@RunIf(min_gpus=1)
+def test_get_nvidia_gpu_stats(tmpdir):
+    """Test GPU get_device_stats with Pytorch < 1.8.0."""
+    current_device = torch.device(f"cuda:{torch.cuda.current_device()}")
+    GPUAccel = GPUAccelerator(
+        training_type_plugin=DataParallelPlugin(parallel_devices=[current_device]), precision_plugin=PrecisionPlugin()
+    )
+    gpu_stats = GPUAccel.get_device_stats(current_device)
+    fields = ["utilization.gpu", "memory.used", "memory.free", "utilization.memory"]
+
+    for f in fields:
+        assert any(f in h for h in gpu_stats.keys())

tests/accelerators/test_tpu.py

@@ -0,0 +1,16 @@
+from pytorch_lightning.accelerators import TPUAccelerator
+from pytorch_lightning.plugins import SingleTPUPlugin
+from pytorch_lightning.plugins.training_type import TPUSpawnPlugin
+from tests.helpers.runif import RunIf
+
+
+@RunIf(tpu=True)
+def test_device_stats_tpu(tmpdir):
+    """Test TPU get_device_stats."""
+    plugin = SingleTPUPlugin(1)
+    TPUAccel = TPUAccelerator(training_type_plugin=TPUSpawnPlugin(), precision_plugin=plugin)
+    tpu_stats = TPUAccel.get_device_stats("1")
+    fields = ["avg. free memory (MB)", "avg. peak memory (MB)"]
+
+    for f in fields:
+        assert any(f in h for h in tpu_stats.keys())