Some fixes for the memory_profiler module in benchmark

Author: FrancescAlted

bench/ndarray/slice-expr.py

@@ -1,8 +1,17 @@
-# Imports
+#######################################################################
+# Copyright (c) 2019-present, Blosc Development Team <[email protected]>
+# All rights reserved.
+#
+# This source code is licensed under a BSD-style license (found in the
+# LICENSE file in the root directory of this source tree)
+#######################################################################
+
+# Benchmark for computing a slice of a expression in a 4D array.
+
 import numpy as np
 import blosc2
 import time
-from memory_profiler import memory_usage
+from memory_profiler import memory_usage, profile
 import matplotlib.pyplot as plt
 
 file = "dset-ones.b2nd"
@@ -26,46 +35,52 @@
 print(f"Time to create expression: {time.time() - t:.6f} seconds")
 
 # dim0
+@profile
 def slice_dim0():
     t = time.time()
     res = expr[1]
     t0 = time.time() - t
     print(f"Time to access dim0: {t0:.6f} seconds")
+    print(f"dim0 slice size: {np.prod(res.shape) * res.dtype.itemsize / 2**30:.6f} GB")
 
 # dim1
+@profile
 def slice_dim1():
     t = time.time()
     res = expr[:,1]
     t1 = time.time() - t
     print(f"Time to access dim1: {t1:.6f} seconds")
+    print(f"dim1 slice size: {np.prod(res.shape) * res.dtype.itemsize / 2**30:.6f} GB")
 
 # dim2
+@profile
 def slice_dim2():
     t = time.time()
     res = expr[:,:,1]
     t2 = time.time() - t
     print(f"Time to access dim2: {t2:.6f} seconds")
+    print(f"dim2 slice size: {np.prod(res.shape) * res.dtype.itemsize / 2**30:.6f} GB")
 
 # dim3
+@profile
 def slice_dim3():
     t = time.time()
     res = expr[:,:,:,1]
-    #res = expr[1]
     t3 = time.time() - t
-
     print(f"Time to access dim3: {t3:.6f} seconds")
+    print(f"dim3 slice size: {np.prod(res.shape) * res.dtype.itemsize / 2**30:.6f} GB")
 
-fig = plt.figure()
-interval = 0.001
-offset = 0
-for f in [slice_dim0, slice_dim1, slice_dim2, slice_dim3]:
-    mem = memory_usage((f,), interval=interval)
-    times = offset + interval * np.arange(len(mem))
-    offset = times[-1]
-    plt.plot(times, mem)
+if __name__ == '__main__':
+    interval = 0.001
+    offset = 0
+    for f in [slice_dim0, slice_dim1, slice_dim2, slice_dim3]:
+        mem = memory_usage((f,), interval=interval)
+        times = offset + interval * np.arange(len(mem))
+        offset = times[-1]
+        plt.plot(times, mem)
 
-plt.xlabel('Time (s)')
-plt.ylabel('Memory usage (MiB)')
-plt.title('Memory usage over time for slicing operations, slice-expr.py')
-plt.legend(['dim0', 'dim1', 'dim2', 'dim3'])
-plt.savefig('plots/slice-expr.png', format="png")
+    plt.xlabel('Time (s)')
+    plt.ylabel('Memory usage (MiB)')
+    plt.title('Memory usage lazyexpr slice (fast path), Linux Blosc2 3.5.1')
+    plt.legend(['expr[1]', 'expr[:,1]', 'expr[:,:,1]', 'expr[:,:,:,1]'])
+    plt.savefig('Linux_Blosc3_5_1_fast.png', format="png")