-
Notifications
You must be signed in to change notification settings - Fork 3
/
main.py
123 lines (97 loc) · 5.36 KB
/
main.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
import utils.constants as constants
from utils.graph import plot, plot_multiple
from utils.system_operations.fio_runner import get_fio_result
from options_files.ops_options_file import parse_option_file_to_dict, get_initial_options_file
import rocksdb.subprocess_manager as spm
from utils.utils import log_update, store_best_option_file, path_of_db, store_diff_options_list
from utils.system_operations.get_sys_info import system_info
from gpt.prompts_generator import generate_option_file_with_gpt
import os
def main():
'''
Main function to run the project. This function will run the db_bench with the initial options file and then
generate new options files using GPT API and run db_bench with the new options file. This function will also
store the output of db_bench in a file. The output file will contain the benchmark results, the options file
used to generate the benchmark results and the reasoning behind the options file as provided by the GPT API.
There will be a separate file for each iteration.
Parameters:
- None
Returns:
- None
'''
# initialize variables
options_files = []
options_list = []
# Set up the path
output_folder_dir = constants.OUTPUT_PATH
os.makedirs(output_folder_dir, exist_ok=True)
db_path = path_of_db()
fio_result = get_fio_result(constants.FIO_RESULT_PATH)
log_update(f"[MFN] Starting the program with the case number: {constants.CASE_NUMBER}")
print(f"[MFN] Starting the program with the case number: {constants.CASE_NUMBER}")
# First run, Initial options file and see how the results are
options, reasoning = get_initial_options_file()
is_error, benchmark_results, average_cpu_usage, average_memory_usage, options = spm.benchmark(
db_path, options, output_folder_dir, reasoning, 0, None, options_files)
if is_error:
# If the initial options file fails, exit the program
log_update("[MFN] Failed to benchmark with the initial options file. Exiting.")
print("[MFN] Failed to benchmark with the initial options file. Exiting.")
exit(1)
else:
# If the initial options file succeeds, store the options file and benchmark results, pass it to the GPT API to generate a new options file
parsed_options = parse_option_file_to_dict(options)
options_list.append(parsed_options)
# Maintain a list of options files, benchmark results and why that option file was generated
options_files.append((options, benchmark_results, reasoning, ""))
iteration_count = 7
for i in range(1, iteration_count + 1):
log_update(f"[MFN] Starting iteration {i}")
log_update(f"[MFN] Querying ChatGPT for next options file")
print("-" * 50)
print(f"[MFN] Starting iteration {i}")
print("[MFN] Querying ChatGPT for next options file")
temperature = 0.4
retry_counter = 5
generated = False
for gpt_query_count in range(retry_counter, 0, -1):
# Generate new options file with retry limit of 5
new_options_file, reasoning, summary_of_changes = generate_option_file_with_gpt(
constants.CASE_NUMBER, options_files,
system_info(db_path, fio_result), temperature,
average_cpu_usage, average_memory_usage,
constants.TEST_NAME, constants.VERSION)
if new_options_file is None:
log_update(f"[MFN] Failed to generate options file. Retrying. Retries left: {gpt_query_count - 1}")
print("[MFN] Failed to generate options file. Retrying. Retries left: ", gpt_query_count - 1)
continue
# Parse output
is_error, benchmark_results, average_cpu_usage, average_memory_usage, new_options_file = spm.benchmark(
db_path, new_options_file, output_folder_dir, reasoning, iteration_count, benchmark_results, options_files)
if is_error:
log_update(f"[MFN] Benchmark failed. Retrying with new options file. Retries left: {gpt_query_count - 1}")
print("[MFN] Benchmark failed. Retrying with new options file. Retries left: ", gpt_query_count - 1)
temperature += 0.1
continue
else:
generated = True
break
if generated:
options = new_options_file
options_files.append((options, benchmark_results, reasoning,
summary_of_changes))
parsed_options = parse_option_file_to_dict(options)
options_list.append(parsed_options)
else:
log_update("[MFN] Failed to generate options file over 5 times. Exiting.")
print("[MFN] Failed to generate options file over 5 times. Exiting.")
exit(1)
store_best_option_file(options_files, output_folder_dir)
# Graph Ops/Sec
plot([e[1]["ops_per_sec"] for e in options_files], "OpsPerSec",
f"{output_folder_dir}/OpsPerSec.png")
plot_multiple(options_files, "Ops Per Second",
f"{output_folder_dir}/opsM_per_sec.png")
store_diff_options_list(options_list, output_folder_dir)
if __name__ == "__main__":
main()