JOQULAR Performance

Platform

Processor: Intel(R) Core(TM) i7-4510U CPU @2.00Ghz

RAM: 8.00GB

Operating System: 64-bit Windows 8

Browser: Chrome Version 42.0.2311.90 m

Disk Drive: Solid State

Test Sample & Parameters

function TestObject(id) {
   this.id = id;
}

10,000 instances with id values dispersed randomly between 1 and 100.

A search value is randomly generated so the result set is usually around 100 records.

10 test runs with the average time computed for each metric (insert, index, save, find, load) throwing out the highs and lows.

Results

The times are in milliseconds. For some databases it is not possible to measure index speed independent of insert or even independent of disk save, only total persistence time can be measured.

For data insert associated metrics the value is for the entire sample set, e.g. 10,000 records are saved in 514.92 milliseconds by JOQULAR.

For find, the number represents the time to find a set of records associated with the search value from the test specification.

JOQULAR has a 0.00 insertion time because it uses the JavaScript objects in memory and only saves on command.

We are not experts with PouchDB, IndexedDB, or Forerunner. We welcome feedback on how we can make the benchmarking code for those databases run faster.

	Insert	Index	Insert + Index	Save	Total Persist Time	Find	Load
Linear						3.98
JOQULAR	0.00	264.32	264.32.87	514.92	779.24	0.15	615.26
Forerunner	129.11		129.11	3.75	132.86	0.33	*
PouchDB					3723.22	18826.74
IndexedDB					1825.34	194.06

*Forerunner load time not tested because there is no way to flush its cache without reloading the page.