My favorites | Sign in
Project Home Downloads Wiki Issues Source
Search
for
Benchmark  
JSON-Smart Vs all other java json parser.
Featured
Updated Oct 10, 2011 by uriel.chemouni

Json BecnhMark

Bench Protole

Those test are generate by parrsing 50.000 pre-build json string.

to minimise gc effect on bench, a new VM is lanched for each test.

New Tests

The bench run on an intel P9700 CPU, using sun JVM 1.6.22 64bit + SSD each bench had been lunch 10 times, the best time had been keeped.

BenchMark run directly after the JVM start

update on 30 May

  • JsonIJ version is now 0.2.7
  • Json-smart is now 1.0.6.3
  • argo is now 2.8

BenchMark : bench

Result display as % of the fastest API. Lower is better, for example 400% means that JSon-smart is 4 times faster.

Test Sojo Argo json-lib JsonMe org.json json-Simple Jackson StringTree JsonIJ Json-fast Json-Smart
mixte 1 129 %822 %834 %395 %437 %324 %269 %287 %205 %134 %100 %
text 1 628 %965 %879 %363 %610 %434 %419 %369 %271 %174 %100 %
int 1 300 %966 %718 %359 %521 %281 %286 %293 %224 %118 %100 %
unicode 1 177 %674 %549 %218 %445 %351 %315 %216 %256 %147 %100 %
boolean 1 285 %724 %809 %391 %498 %332 %291 %210 %228 %138 %100 %
float 1 145 %794 %711 %1 152 %450 %243 %265 %272 %201 %126 %100 %
TOTAL 1 251 %818 %739 %511 %484 %319 %300 %272 %228 %137 %100 %

BenchMark with preload classes

Not updated yet

preloading class look to have no significant effect, due to SSD usage.

Test JsonIJ Sojo Argo json-lib org.json JsonMe json-Simple Jackson StringTree Json-fast Json-Smart
mixte 1 302 %1 150 %804 %858 %452 %365 %323 %258 %275 %133 %100 %
text 1 772 %1 492 %914 %782 %582 %332 %409 %355 %332 %159 %100 %
int 1 266 %1 345 %907 %723 %447 %290 %279 %250 %304 %118 %100 %
unicode N/A1 185 %674 %538 %448 %216 %358 %280 %216 %135 %100 %
boolean 1 401 %1 168 %680 %717 %459 %283 %300 %254 %192 %129 %100 %
float 1 024 %1 129 %740 %707 %384 %1 043 %247 %242 %280 %125 %100 %
TOTAL N/A1 230 %781 %717 %454 %452 %314 %269 %266 %132 %100 %

Comment by tsaloranta@gmail.com, May 16, 2011

Where is the source? Are you warming up GC (if not, this test is totally meaningless)? What kind of JSON are you parsing?

Comment by project member uriel.chemouni, May 16, 2011

Sources are available now: http://json-smart.googlecode.com/svn/trunk/bench/

Json test string generator: http://json-smart.googlecode.com/svn/trunk/bench/src/net/minidev/bench/json/TestData.java

a new JVM is started for each test, GC is never explicitely called.

Comment by tsaloranta@gmail.com, May 19, 2011

Whops, meant JVM warmed up, not GC. :-)

It is necessary to run code to be tested for couple of seconds first on new JVM, before taking measurements; otherwise test is about how fast JVM can load (and possibly inline) bytecode.

Comment by project member uriel.chemouni, May 23, 2011

Warmed up test is now available.

Comment by Programm...@gmail.com, Jun 22, 2011

Using the jvm-serializers (https://github.com/eishay/jvm-serializers/wiki) test suite, I get significantly different results. When testing with about 736 bytes of JSON input (in java.lang.String format), my results show the different APIs performing in the following order, ranked from fastest to slowest.

                                ser     deser   total
json/jackson-databind-strings   6034    8085    14209
json/fastjson-databind          5837    8924    14826
json/json-smart/manual/tree     13104   9644    23182
json/org.json/manual/tree       16404   20039   36526
json/json.simple/manual/tree    15971   21954   38170
json/jsonij-manual/tree         75124   16888   92275
json/argo-manual/tree-builders  176566  51439   228847
json/json-lib-databind          73202   169966  243474

Using the json-smart performance as the baseline, here are the same numbers as percentages (rounded to the nearest whole number).

                                ser     deser   total
json/jackson-databind-strings   46%     84%     61%
json/fastjson-databind          45%     93%     64%
json/json-smart/manual/tree     100%    100%    100%
json/org.json/manual/tree       125%    208%    158%
json/json.simple/manual/tree    122%    228%    165%
json/jsonij-manual/tree         573%    175%    398%
json/argo-manual/tree-builders  1347%   533%    987%
json/json-lib-databind          559%    1762%   1050%

These results place the performance of some of the different APIs in significantly different orders, and some with significantly different percentages.

As time permits, I'll try to figure out what the cause of the different results is.

(I have no affiliation with any of these projects, or with any other JSON processing project.)

Comment by matthias...@gmail.com, Oct 24, 2011

I've tested jackson, simple-json and json-smart with some 100k of tweets and in our case smart-json is very fast parsing those quite long and nested json lines.

Comment by bakker.mark, Feb 21, 2012

If you lanch 10 times and only take the best result in account, the measurement has statistically no meaning.

See for more info

http://stats.org/in_depth/faq/statistical_significance.htm

The current measurements only state that that json-smart can give faster result then others. It could be the case that json-smart is using caching and other do not. A better presentation would be a normal distribution see;

http://en.wikipedia.org/wiki/Normal_distribution

Comment by inpo...@gmail.com, Mar 6, 2012

programmerbruce, what do you mean by "significantly"? I see the different results, but no mention of confidence or other quantification of significance.

Comment by tsaloranta@gmail.com, Apr 11, 2012

Wrt programmer bruce's results, much more info is available on the linked-to web site.


Sign in to add a comment
Powered by Google Project Hosting