Note

Go to the end to download the full example code.

is_all_true performance benchmarks#

This example benchmarks the is_all_true performance for the different algorithms available in Shamrock

 9 import random
10 import time
11
12 import matplotlib.colors as colors
13 import matplotlib.pyplot as plt
14 import numpy as np
15
16 import shamrock
17
18 # If we use the shamrock executable to run this script instead of the python interpreter,
19 # we should not initialize the system as the shamrock executable needs to handle specific MPI logic
20 if not shamrock.sys.is_initialized():
21     shamrock.change_loglevel(1)
22     shamrock.sys.init("0:0")

Main benchmark functions

30 def benchmark_is_all_true_random(N, nb_repeat=10):
31     times = []
32     for i in range(nb_repeat):
33         random.seed(111)
34         buf = shamrock.algs.mock_buffer_u8(random.randint(0, 1000000), N, 0, 1)
35         times.append(shamrock.algs.benchmark_is_all_true(buf, N))
36     return min(times), max(times), sum(times) / nb_repeat
37
38
39 def benchmark_is_all_true_ones(N, nb_repeat=10):
40     times = []
41     for i in range(nb_repeat):
42         buf = shamrock.backends.DeviceBuffer_u8()
43         buf.resize(N)
44         buf.fill(1)
45         times.append(shamrock.algs.benchmark_is_all_true(buf, N))
46     return min(times), max(times), sum(times) / nb_repeat
47
48
49 def benchmark_is_all_true_zeros(N, nb_repeat=10):
50     times = []
51     for i in range(nb_repeat):
52         buf = shamrock.backends.DeviceBuffer_u8()
53         buf.resize(N)
54         buf.fill(0)
55         times.append(shamrock.algs.benchmark_is_all_true(buf, N))
56     return min(times), max(times), sum(times) / nb_repeat

Run the performance test for all parameters

61 def run_performance_sweep():
62
63     # Define parameter ranges
64     # logspace as array
65     particle_counts = np.logspace(2, 7, 20).astype(int).tolist()
66
67     # Initialize results matrix
68     results_random = []
69     results_ones = []
70     results_zeros = []
71
72     print(f"Particle counts: {particle_counts}")
73
74     total_runs = len(particle_counts)
75     current_run = 0
76
77     for i, N in enumerate(particle_counts):
78         current_run += 1
79
80         print(
81             f"[{current_run:2d}/{total_runs}] Running N={N:5d}...",
82             end=" ",
83         )
84
85         start_time = time.time()
86         min_time, max_time, mean_time = benchmark_is_all_true_random(N)
87         results_random.append(mean_time)
88         min_time, max_time, mean_time = benchmark_is_all_true_ones(N)
89         results_ones.append(mean_time)
90         min_time, max_time, mean_time = benchmark_is_all_true_zeros(N)
91         results_zeros.append(mean_time)
92         elapsed = time.time() - start_time
93
94         print(f"mean={mean_time:.3f}s (took {elapsed:.1f}s)")
95
96     return particle_counts, results_random, results_ones, results_zeros

List current implementation

101 current_impl = shamrock.algs.get_current_impl_is_all_true()
102
103 print(current_impl)

impl_param(impl_name="host", params="")

List all implementations available

107 all_default_impls = shamrock.algs.get_default_impl_list_is_all_true()
108
109 print(all_default_impls)

[impl_param(impl_name="host", params=""), impl_param(impl_name="sum_reduction", params="")]

Run the performance benchmarks for all implementations

114 for impl in all_default_impls:
115     shamrock.algs.set_impl_is_all_true(impl.impl_name, impl.params)
116
117     print(f"Running is_all_true performance benchmarks for {impl}...")
118
119     # Run the performance sweep
120     particle_counts, results_random, results_ones, results_zeros = run_performance_sweep()
121
122     plt.plot(particle_counts, results_random, "--", label=impl.impl_name + " (random set)")
123     plt.plot(particle_counts, results_ones, "--+", label=impl.impl_name + " (all ones)")
124     plt.plot(particle_counts, results_zeros, "--o", label=impl.impl_name + " (all zeros)")
125
126
127 Nobj = np.array(particle_counts)
128 Time100M = Nobj / 1e8
129 plt.plot(particle_counts, Time100M, color="grey", linestyle="-", alpha=0.7, label="100M obj/sec")
130
131
132 plt.xlabel("Number of elements")
133 plt.ylabel("Time (s)")
134 plt.title("is_all_true performance benchmarks")
135
136 plt.xscale("log")
137 plt.yscale("log")
138
139 plt.grid(True)
140
141 plt.legend()
142 plt.show()
is_all_true performance benchmarks
Info: setting is_all_true implementation to impl : host                              [tree][rank=0]
Running is_all_true performance benchmarks for impl_param(impl_name="host", params="")...
Particle counts: [100, 183, 335, 615, 1128, 2069, 3792, 6951, 12742, 23357, 42813, 78475, 143844, 263665, 483293, 885866, 1623776, 2976351, 5455594, 10000000]
[ 1/20] Running N=  100... mean=0.000s (took 0.0s)
[ 2/20] Running N=  183... mean=0.000s (took 0.0s)
[ 3/20] Running N=  335... mean=0.000s (took 0.0s)
[ 4/20] Running N=  615... mean=0.000s (took 0.0s)
[ 5/20] Running N= 1128... mean=0.000s (took 0.0s)
[ 6/20] Running N= 2069... mean=0.000s (took 0.0s)
[ 7/20] Running N= 3792... mean=0.000s (took 0.0s)
[ 8/20] Running N= 6951... mean=0.000s (took 0.0s)
[ 9/20] Running N=12742... mean=0.000s (took 0.0s)
[10/20] Running N=23357... mean=0.000s (took 0.0s)
[11/20] Running N=42813... mean=0.000s (took 0.0s)
[12/20] Running N=78475... mean=0.000s (took 0.0s)
[13/20] Running N=143844... mean=0.000s (took 0.0s)
[14/20] Running N=263665... mean=0.000s (took 0.0s)
[15/20] Running N=483293... mean=0.000s (took 0.1s)
[16/20] Running N=885866... mean=0.000s (took 0.1s)
[17/20] Running N=1623776... mean=0.000s (took 0.2s)
[18/20] Running N=2976351... mean=0.000s (took 0.4s)
[19/20] Running N=5455594... mean=0.001s (took 0.6s)
[20/20] Running N=10000000... mean=0.001s (took 1.1s)
Info: setting is_all_true implementation to impl : sum_reduction                     [tree][rank=0]
Running is_all_true performance benchmarks for impl_param(impl_name="sum_reduction", params="")...
Particle counts: [100, 183, 335, 615, 1128, 2069, 3792, 6951, 12742, 23357, 42813, 78475, 143844, 263665, 483293, 885866, 1623776, 2976351, 5455594, 10000000]
[ 1/20] Running N=  100... mean=0.000s (took 0.0s)
[ 2/20] Running N=  183... mean=0.000s (took 0.0s)
[ 3/20] Running N=  335... mean=0.000s (took 0.0s)
[ 4/20] Running N=  615... mean=0.000s (took 0.0s)
[ 5/20] Running N= 1128... mean=0.000s (took 0.0s)
[ 6/20] Running N= 2069... mean=0.000s (took 0.0s)
[ 7/20] Running N= 3792... mean=0.000s (took 0.0s)
[ 8/20] Running N= 6951... mean=0.000s (took 0.0s)
[ 9/20] Running N=12742... mean=0.000s (took 0.0s)
[10/20] Running N=23357... mean=0.000s (took 0.0s)
[11/20] Running N=42813... mean=0.000s (took 0.0s)
[12/20] Running N=78475... mean=0.000s (took 0.0s)
[13/20] Running N=143844... mean=0.000s (took 0.0s)
[14/20] Running N=263665... mean=0.000s (took 0.0s)
[15/20] Running N=483293... mean=0.000s (took 0.1s)
[16/20] Running N=885866... mean=0.001s (took 0.1s)
[17/20] Running N=1623776... mean=0.001s (took 0.3s)
[18/20] Running N=2976351... mean=0.002s (took 0.5s)
[19/20] Running N=5455594... mean=0.004s (took 0.9s)
[20/20] Running N=10000000... mean=0.010s (took 1.8s)

Total running time of the script: (0 minutes 7.502 seconds)

Estimated memory usage: 171 MB

Gallery generated by Sphinx-Gallery