[profiler] primary memory tracer

colossalai/engine/_base_engine.py

@@ -26,6 +26,8 @@ class Engine:
     :type gradient_handlers: list
     :param clip_grad_norm: The norm of gradient clipping
     :type clip_grad_norm: float, optional
+    :param ophook_list: List of ophook
+    :type ophook_list: list
     :param verbose: whether to display log info
     :type verbose: bool
     """

colossalai/engine/ophooks/_memtracer_ophook.py

+from re import S
+from colossalai.context.parallel_mode import ParallelMode
 import torch
 from . import BaseOpHook
 from concurrent.futures import ThreadPoolExecutor
@@ -5,18 +7,20 @@ from colossalai.registry import OPHOOKS
 from colossalai.logging import get_dist_logger
 from time import sleep, time
 import pickle
+from typing import Union, Optional
+from colossalai.core import global_context as gpc
 
 
-def get_cuda_memory_used(device):
+def get_cuda_memory_used(device: Optional[torch.device]) -> int:
     """
     Get the free memory info of device.
     Notice that for CPU, this function will return 1/N of the total free memory,
     where N is the world size.
     """
-    ret = torch.cuda.memory_allocated()
+    ret: int = torch.cuda.memory_allocated(device)
     # get the peak memory to report correct data, so reset the counter for the next call
     if hasattr(torch.cuda, "reset_peak_memory_stats"):  # pytorch 1.4+
-        torch.cuda.reset_peak_memory_stats()
+        torch.cuda.reset_peak_memory_stats(device)
     return ret
 
 
@@ -34,6 +38,9 @@ class AsyncMemoryMonitor:
         self.time_stamps = []
         self.mem_stats = []
 
+    def __len__(self):
+        return len(self.mem_stats)
+
     def set_interval(self, power: int):
         self.interval = 1 / (10**power)
 
@@ -75,21 +82,64 @@ class AsyncMemoryMonitor:
         with open(filename, "wb") as f:
             pickle.dump(self.state_dict(), f)
     
+    def clear(self):
+        self.mem_stats.clear()
+        self.time_stamps.clear()
+
 
 @OPHOOKS.register_module
 class MemTracerOpHook(BaseOpHook):
-    def __init__(self, niter=5):
+    '''
+    Collect GPU memory usage information
+
+    Args: 
+        warmup (int): This parameter indicates how many iterations to truncate
+        before profiling, e.g. set to 5 and the data will start from 6-th iteration
+        refreshrate (int): This parameter decides the frequency of write file.
+        datafile(string): the name of the stats data file
+    Attributes:
+        _warmup (int): warmup iterations
+        _refreshrate(int): how many iterations we shall refresh the file
+        _logger (colossalai.logging.logger): output log file
+        _curiter (int): current iteration number
+        _count (int): the number of times the data file was written
+        _data_prefix (string): the prefix of the stats data file
+        _rank (int): the rank of current node
+    '''
+    def __init__(self, warmup: int = 50, refreshrate: int = 10, data_prefix: str = "memstats"):
         super().__init__()
         self.async_mem_monitor = AsyncMemoryMonitor()
-        self._niter = niter
         self._curiter = 0
         self._logger = get_dist_logger()
+        self._count = 0
+        self._warmup = warmup
+        self._refreshrate = refreshrate
+        self._data_prefix = data_prefix
+        # in distributed environment
+        if gpc.is_initialized(ParallelMode.GLOBAL):
+            self._rank = gpc.get_global_rank()
+        else:
+            self._rank = 0
+
+    def _isvalid(self, module) -> bool:
+        assert isinstance(module, torch.nn.Module)
+        return module.training
+
+    @property
+    def refreshrate(self) -> int:
+        return self._refreshrate
     
-    def _isvalid(self, module):
-        return module.training and self._curiter < self._niter
+    @property
+    def warmup(self) -> int:
+        return self._warmup
 
-    def niter(self):
-        return self._niter
+    @property
+    def curiter(self) -> int:
+        return self._curiter
+
+    @property
+    def valid_iter(self) -> int:
+        return self.curiter - self.warmup
 
     def pre_fwd_exec(self, module: torch.nn.Module, *args):
         if self._isvalid(module):
@@ -103,14 +153,12 @@ class MemTracerOpHook(BaseOpHook):
             self._logger.debug(f'FWD POST {module.__class__.__name__}')
 
     def pre_bwd_exec(self, module: torch.nn.Module, input, output):
-        assert isinstance(module, torch.nn.Module)
         if self._isvalid(module):
             self.async_mem_monitor.finish()
             self.async_mem_monitor.start()
             self._logger.debug(f'BWD PRE {module.__class__.__name__}')
 
     def post_bwd_exec(self, module: torch.nn.Module, input):
-        assert isinstance(module, torch.nn.Module)
         if self._isvalid(module):
             self.async_mem_monitor.finish()
             self._logger.debug(f'BWD POST {module.__class__.__name__}')
@@ -120,11 +168,24 @@ class MemTracerOpHook(BaseOpHook):
 
     def post_iter(self):
         self.async_mem_monitor.finish()
-        if self._curiter == self._niter:
+        # in the warmup stage
+        if self._curiter < self.warmup:
+            # TODO: record time and adaptively change sampling rate
+            pass
+        elif self._curiter == self._warmup:
+            self.async_mem_monitor.clear()
+        else:
+            # every `refreshrate` times, refresh the file
+            if self.valid_iter != 0 and self.valid_iter % self.refreshrate == 0:
+                # output file info
                 self._logger.info(
-                f'dump a memory statistics as pickle to ./memstats.pkl')
-            self.save_results("memstats.pkl")
+                    f'dump a memory statistics as pickle to {self._dataprefix}-{self._rank}.pkl')
+                self.save_results()
+                self._count += 1
+                self._logger.debug(f'data file has been refreshed {self._count} times')
+        # finish a iteration
         self._curiter += 1
 
-    def save_results(self, filename):
-        self.async_mem_monitor.save(filename)
+    def save_results(self):
+        datafile = f"{self._data_prefix}-{self._rank}.pkl"
+        self.async_mem_monitor.save(datafile)

colossalai/utils/timer.py

@@ -19,6 +19,11 @@ class Timer:
     def has_history(self):
         return len(self._history) != 0
 
+    @property
+    def current_time(self) -> float:
+        synchronize()
+        return time.time()
+
     def start(self):
         """Fisrtly synchronize cuda, reset the clock and then start the timer.
         """
@@ -27,6 +32,11 @@ class Timer:
         self._start_time = time.time()
         self._started = True
 
+    def lap(self):
+        """lap time and return elapsed time
+        """
+        return self.current_time - self._start_time
+
     def stop(self, keep_in_history: bool = False):
         """Stop the timer and record the start-stop time interval.
 

tests/test_config/test_load_config.py

@@ -22,6 +22,7 @@ def test_load_config():
 
 @pytest.mark.cpu
 def test_load_ophooks():
-    dict = {'type': 'MemTracerOpHook', 'niter': 2}
+    dict = {'type': 'MemTracerOpHook', 'warmup': 10, 'refreshrate': 20}
     ophook = build_ophooks(dict)
-    assert ophook.niter() == 2
+    assert ophook.refreshrate == 20
+    assert ophook.warmup == 10