big.text.floating_point.md

Read and Write Time Comparison

Input filename	Size (MB)	Content
data/output/big.text.floating_point.json	794.72	Text,doubles

Environment	Windows, Intel
Computer	Lenovo Thinkpad X Extreme with 8 cores, 16 hardware threads, 32 GB memory, 2 TB solid state drive
Operating system	Windows 2011
Compiler	Visual Studio 2022 Version 17.14.16
Optimization	/O2,/Ot
Language	/std:c++23preview

Library	Version	Type to read into/write from
jsoncons	JSONCONS_VERSION_MAJOR.JSONCONS_VERSION_MINOR.JSONCONS_VERSION_PATCH	jsoncons::json
nlohmann	3.12.0	nlohmann::json
rapidjson	1.1.0	rapidjson::Document
jsoncpp	1.9.6	Json::Value
glaze	7.4.0	glz::generic
Boost.JSON	1_90	boost::json::value
yyjson	0.12.0	yyjson_doc
reflect-cpp	0.24.0	rfl::Generic

Library	Time to read (s)	Time to write (s)	Physical memory footprint of json value (MB)	Notes
jsoncons	8.879	3.879	423	Uses sorted std::vector of key/value pairs for objects, expect smaller memory footprint.Uses slightly modified grisu3_59_56 implementation by Florian Loitsch plus fallback for printing doubles, expect faster serializing.
nlohmann	15.997	7.165	494	Uses std::map for objects. Uses slightly modified Grisu2 implementation by Florian Loitsch for printing doubles, expect faster serializing.
rapidjson	3.443	3.061	457	Uses custom floating point parsing, expect faster parsing. Uses girsu3 for printing doubles, expect faster serializing. Uses custom allocation and flat map for objects, expect smaller memory footprint.
jsoncpp	52.944	30.007	2466	Uses std::map for both arrays and objects, expect larger memory footprint.
glaze	5.266	1.825	1087
Boost.JSON	3.579	2.722	621
yyjson	1.672	1.497	1205
reflect-cpp	19.924	5.482	1032	Uses yyjson for serializing and deserializing JSON.