ssrjson-benchmark 0.0.4__cp310-cp310-win_amd64.whl → 0.0.6__cp310-cp310-win_amd64.whl

ssrjson_benchmark/__main__.py

@@ -37,10 +37,17 @@ def main():
         action="store_true",
     )
     parser.add_argument(
-        "--process-bytes",
-        help="Total process bytes per test, default 1e8",
+        "--process-gigabytes",
+        help="Total process gigabytes per test, default 0.1 (float)",
         required=False,
-        default=1e8,
+        default=0.1,
+        type=float,
+    )
+    parser.add_argument(
+        "--bin-process-megabytes",
+        help="Maximum bytes to process per read for binary formats, default 32 (int)",
+        required=False,
+        default=32,
         type=int,
     )
     parser.add_argument(
@@ -70,7 +77,14 @@ def main():
         result = parse_file_result(result_)
         file = args.file.split("/")[-1]
     else:
-        result, file = run_benchmark(benchmark_files_dir, args.process_bytes)
+        process_bytes = int(args.process_gigabytes * 1024 * 1024 * 1024)
+        bin_process_bytes = args.bin_process_megabytes * 1024 * 1024
+        if process_bytes <= 0 or bin_process_bytes <= 0:
+            print("process-gigabytes and bin-process-megabytes must be positive.")
+            sys.exit(1)
+        result, file = run_benchmark(
+            benchmark_files_dir, process_bytes, bin_process_bytes
+        )
         file = file.split("/")[-1]
 
     if args.markdown:

ssrjson_benchmark/benchmark_impl.py

@@ -2,6 +2,7 @@ import gc
 import io
 import json
 import math
+import multiprocessing
 import os
 import pathlib
 import platform
@@ -43,6 +44,8 @@ LIBRARIES_COLORS = {
     "ssrjson": "#fd8d3c",
 }
 
+MAX_BIN_BYTES_SIZE = 512 * 1024 * 1024  # 512MiB
+
 
 class BenchmarkFunction:
     def __init__(self, func: Callable, library_name: str) -> None:
@@ -65,11 +68,13 @@ class BenchmarkGroup:
 
 
 # benchmarkers
-def _benchmark(repeat_time: int, func, *args):
+def _benchmark(repeat_time: int, times_per_bin: int, func, *args):
     """
     Run repeat benchmark, with utf-8 cache.
     returns time used (ns).
     """
+    # times_per_bin not used
+    # disable automatic GC
     gc_was_enabled = _gc_prepare()
     try:
         # warm up
@@ -82,53 +87,62 @@ def _benchmark(repeat_time: int, func, *args):
             gc.enable()
 
 
-def _benchmark_unicode_arg(repeat_time: int, func, unicode: str):
+def _benchmark_unicode_arg(repeat_time: int, times_per_bin: int, func, unicode: str):
     """
     Run repeat benchmark, disabling utf-8 cache.
     returns time used (ns).
     """
+    # disable automatic GC
     gc_was_enabled = _gc_prepare()
     try:
-        warmup_size = 100
-        # prepare identical data, without sharing objects
-        warmup_data = _ssrjson_benchmark.copy_unicode_list_invalidate_cache(
-            unicode, warmup_size
-        )
-        benchmark_data = _ssrjson_benchmark.copy_unicode_list_invalidate_cache(
-            unicode, repeat_time
-        )
-        # warm up
-        for i in range(warmup_size):
-            _ssrjson_benchmark.run_object_benchmark(func, (warmup_data[i],))
+        times_left = repeat_time
         total = 0
-        for i in range(repeat_time):
-            total += _ssrjson_benchmark.run_object_benchmark(func, (benchmark_data[i],))
+        while times_left != 0:
+            cur_bin_size = min(times_left, times_per_bin)
+            times_left -= cur_bin_size
+            # prepare identical data, without sharing objects
+            benchmark_data = _ssrjson_benchmark.copy_unicode_list_invalidate_cache(
+                unicode, cur_bin_size + 1
+            )
+            # warm up
+            _ssrjson_benchmark.run_object_benchmark(func, (benchmark_data[0],))
+            #
+            for i in range(1, cur_bin_size + 1):
+                total += _ssrjson_benchmark.run_object_benchmark(
+                    func, (benchmark_data[i],)
+                )
+            del benchmark_data
         return total
     finally:
         if gc_was_enabled:
             gc.enable()
 
 
-def _benchmark_invalidate_dump_cache(repeat_time: int, func, raw_bytes: bytes):
+def _benchmark_invalidate_dump_cache(
+    repeat_time: int, times_per_bin: int, func, raw_bytes: bytes
+):
     """
     Invalidate utf-8 cache for the same input.
     returns time used (ns).
     """
-    # prepare identical data, without sharing objects
-    data_warmup = [json.loads(raw_bytes) for _ in range(10)]
-    data = [json.loads(raw_bytes) for _ in range(repeat_time)]
-    # disable GC
+    # disable automatic GC
     gc_was_enabled = _gc_prepare()
     try:
-        # warm up
-        for i in range(10):
-            new_args = (data_warmup[i],)
-            _ssrjson_benchmark.run_object_benchmark(func, new_args)
-        #
+        times_left = repeat_time
         total = 0
-        for i in range(repeat_time):
-            new_args = (data[i],)
-            total += _ssrjson_benchmark.run_object_benchmark(func, new_args)
+        while times_left != 0:
+            cur_bin_size = min(times_left, times_per_bin)
+            times_left -= cur_bin_size
+            # prepare identical data, without sharing objects
+            benchmark_data = [json.loads(raw_bytes) for _ in range(cur_bin_size + 1)]
+            # warm up
+            _ssrjson_benchmark.run_object_benchmark(func, (benchmark_data[0],))
+            #
+            for i in range(1, cur_bin_size + 1):
+                total += _ssrjson_benchmark.run_object_benchmark(
+                    func, (benchmark_data[i],)
+                )
+            del benchmark_data
         return total
     finally:
         if gc_was_enabled:
@@ -269,6 +283,7 @@ def _get_processed_size(func: Callable, input_data, is_dumps):
 def _run_benchmark(
     cur_result_file: BenchmarkResultPerFile,
     repeat_times: int,
+    times_per_bin: int,
     input_data: str | bytes,
     benchmark_group: BenchmarkGroup,
 ):
@@ -278,8 +293,14 @@ def _run_benchmark(
     input_data = benchmark_group.input_preprocessor(input_data)
 
     for benchmark_target in benchmark_group.functions:
+        prefix = f"[{benchmark_target.library_name}][{benchmark_group.group_name}]"
+        print(
+            prefix
+            + (" " * max(0, 40 - len(prefix)))
+            + f"repeat_times={repeat_times} times_per_bin={times_per_bin}"
+        )
         speed = benchmark_group.benchmarker(
-            repeat_times, benchmark_target.func, input_data
+            repeat_times, times_per_bin, benchmark_target.func, input_data
         )
         cur_lib = cur_target[benchmark_target.library_name]
         cur_lib.speed = speed
@@ -308,6 +329,7 @@ def _run_file_benchmark(
     benchmark_libraries: dict[str, BenchmarkGroup],
     file: pathlib.Path,
     process_bytes: int,
+    bin_process_bytes: int,
 ):
     print(f"Running benchmark for {file.name}")
     with open(file, "rb") as f:
@@ -316,14 +338,19 @@ def _run_file_benchmark(
     base_file_name = os.path.basename(file)
     cur_result_file = BenchmarkResultPerFile()
     cur_result_file.byte_size = bytes_size = len(raw_bytes)
+    if bytes_size == 0:
+        raise RuntimeError(f"File {file} is empty.")
     kind, str_size, is_ascii, _ = _ssrjson_benchmark.inspect_pyunicode(raw)
     cur_result_file.pyunicode_size = str_size
     cur_result_file.pyunicode_kind = kind
     cur_result_file.pyunicode_is_ascii = is_ascii
     repeat_times = int((process_bytes + bytes_size - 1) // bytes_size)
+    times_per_bin = max(1, bin_process_bytes // bytes_size)
 
     for benchmark_group in benchmark_libraries.values():
-        _run_benchmark(cur_result_file, repeat_times, raw_bytes, benchmark_group)
+        _run_benchmark(
+            cur_result_file, repeat_times, times_per_bin, raw_bytes, benchmark_group
+        )
     return base_file_name, cur_result_file
 
 
@@ -732,11 +759,32 @@ def parse_file_result(j):
     return BenchmarkFinalResult.parse(j)
 
 
-def run_benchmark(files: list[pathlib.Path], process_bytes: int = int(1e8)):
+def is_unix_except_macos():
+    system = platform.system()
+    return system in ("Linux", "AIX", "FreeBSD")
+
+
+def _set_multiprocessing_start_method():
+    try:
+        multiprocessing.set_start_method("fork")
+    except RuntimeError as e:
+        if "context has already been set" not in str(e):
+            raise
+
+
+def run_benchmark(
+    files: list[pathlib.Path],
+    process_bytes: int,
+    bin_process_bytes: int,
+):
     """
     Generate a JSON result of benchmark.
     Also returns a result object.
     """
+    # Set multiprocessing start method to fork, if Python version is 3.14+ on Unix
+    if sys.version_info >= (3, 14) and is_unix_except_macos():
+        _set_multiprocessing_start_method()
+
     file = _get_real_output_file_name()
 
     result = BenchmarkFinalResult()
@@ -747,7 +795,9 @@ def run_benchmark(files: list[pathlib.Path], process_bytes: int = int(1e8)):
     result.catagories = sorted(list(benchmark_libraries.keys()))
 
     for bench_file in files:
-        k, v = _run_file_benchmark(benchmark_libraries, bench_file, process_bytes)
+        k, v = _run_file_benchmark(
+            benchmark_libraries, bench_file, process_bytes, bin_process_bytes
+        )
         result.results[k] = v
     output_result = result.dumps()
 

{ssrjson_benchmark-0.0.4.dist-info → ssrjson_benchmark-0.0.6.dist-info}/METADATA

@@ -1,30 +1,9 @@
 Metadata-Version: 2.4
 Name: ssrjson-benchmark
-Version: 0.0.4
+Version: 0.0.6
 Summary: benchmark of ssrJSON
 Author-email: Eritque Arcus <eritque-arcus@ikuyo.dev>, Antares <antares0982@gmail.com>
-License: MIT License
-        
-        Copyright (c) 2025 Eritque arcus
-        
-        Permission is hereby granted, free of charge, to any person obtaining a copy
-        of this software and associated documentation files (the "Software"), to deal
-        in the Software without restriction, including without limitation the rights
-        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-        copies of the Software, and to permit persons to whom the Software is
-        furnished to do so, subject to the following conditions:
-        
-        The above copyright notice and this permission notice shall be included in all
-        copies or substantial portions of the Software.
-        
-        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-        SOFTWARE.
-        
+License-Expression: MIT
 Project-URL: Homepage, https://github.com/Nambers/ssrJSON-benchmark
 Project-URL: Issues, https://github.com/Nambers/ssrJSON-benchmark/issues
 Project-URL: Repository, https://github.com/Nambers/ssrJSON-benchmark.git
@@ -35,7 +14,7 @@ Classifier: Intended Audience :: Developers
 Classifier: Topic :: System :: Benchmark
 Classifier: Programming Language :: Python :: 3
 Classifier: Operating System :: OS Independent
-Requires-Python: >=3.9
+Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: ssrjson

{ssrjson_benchmark-0.0.4.dist-info → ssrjson_benchmark-0.0.6.dist-info}/RECORD

@@ -1,7 +1,7 @@
 ssrjson_benchmark/__init__.py,sha256=i4lVJ3CRKTxfxGyXREUi5OUesvmwWm8fEkcME7usGpk,380
-ssrjson_benchmark/__main__.py,sha256=PQgkJRKuQT9NPQ3JZD7qyAbHhRKbh9UvkxOlXsJnITA,2312
-ssrjson_benchmark/_ssrjson_benchmark.pyd,sha256=Z15oFGhdAWs8RKyYngxiAFRcvUQvtbUZ-1FYsTHcQw0,13312
-ssrjson_benchmark/benchmark_impl.py,sha256=5yRsbhcB5jz1VC4wuNgI4BZndZVxogW3_obz9jC2OBA,23828
+ssrjson_benchmark/__main__.py,sha256=kCW5S5avgZcb2W4I4Mlb0ytXCF9epg4P441JEJywxac,2901
+ssrjson_benchmark/_ssrjson_benchmark.pyd,sha256=ceJ3cc178Jv3rlTiP2yRB0OJMfqRkMizzeAmlcqmuCc,13312
+ssrjson_benchmark/benchmark_impl.py,sha256=YgXaMpYkpApPvthuDQ8TGJoKDruSH46Lf6L_1QLWovU,25427
 ssrjson_benchmark/result_types.py,sha256=_rrM4Lz4-8fJhtHQ9tOF2tr6dRc-FU4CmabodUO_hk8,2379
 ssrjson_benchmark/template.md,sha256=ZRM36XR-EB2Up0BkRiE1IL7GmlaozwNQ-S1TFw_hnwM,272
 ssrjson_benchmark/_files/MotionsQuestionsAnswersQuestions2016.json,sha256=EqcdpROZfEPUSuOq1PXmDY3WXQqPvNYmvqAYF_30ICQ,10323246
@@ -16,8 +16,8 @@ ssrjson_benchmark/_files/simple_object_zh.json,sha256=uPOPBRqqFwxxu1Sh0DMoe1q9o3
 ssrjson_benchmark/_files/truenull.json,sha256=enl_cy6qWa8b7bdVpx3-e0k6X9BFe_aCPot1rEquGiU,12000
 ssrjson_benchmark/_files/tweet.json,sha256=nd_aAjLwHMcvzh_4ntacdK6S1Dlz65bcsPEzS-9MEtc,5128
 ssrjson_benchmark/_files/twitter.json,sha256=apbJI9QLVdnFuler9fbDV8AyK9Tz2AuygSujmeMKn2o,770627
-ssrjson_benchmark-0.0.4.dist-info/licenses/LICENSE,sha256=grt4_GrNwicUFuNLSSbvvdGq2fFW6s81CFBO9_mtQbg,1091
-ssrjson_benchmark-0.0.4.dist-info/METADATA,sha256=9LlYrNP1-cBf3hP8WakHkcFLNVBNzwVNCT8JqFAji08,5885
-ssrjson_benchmark-0.0.4.dist-info/WHEEL,sha256=KUuBC6lxAbHCKilKua8R9W_TM71_-9Sg5uEP3uDWcoU,101
-ssrjson_benchmark-0.0.4.dist-info/top_level.txt,sha256=l1O9IjI1lR5DczhKv9O-GeItgq5HZySDOI5KjfFfvq8,37
-ssrjson_benchmark-0.0.4.dist-info/RECORD,,
+ssrjson_benchmark-0.0.6.dist-info/licenses/LICENSE,sha256=grt4_GrNwicUFuNLSSbvvdGq2fFW6s81CFBO9_mtQbg,1091
+ssrjson_benchmark-0.0.6.dist-info/METADATA,sha256=U20IW7AekyqgYTHaNIgFN9jgq6m7nm6KxLZCiIdwKHI,4641
+ssrjson_benchmark-0.0.6.dist-info/WHEEL,sha256=KUuBC6lxAbHCKilKua8R9W_TM71_-9Sg5uEP3uDWcoU,101
+ssrjson_benchmark-0.0.6.dist-info/top_level.txt,sha256=l1O9IjI1lR5DczhKv9O-GeItgq5HZySDOI5KjfFfvq8,37
+ssrjson_benchmark-0.0.6.dist-info/RECORD,,