[TIMOB-13729] Create JavaScript benchmark suite specific to Titanium use cases
| GitHub Issue | n/a | 
|---|---|
| Type | Epic | 
| Priority | Low | 
| Status | Open | 
| Resolution | Unresolved | 
| Affected Version/s | Release 3.1.0 | 
| Fix Version/s | n/a | 
| Components | Android, Core, iOS | 
| Labels | performance | 
| Reporter | Matt Langston | 
| Assignee | Pedro Enrique | 
| Created | 2013-04-30T20:45:28.000+0000 | 
| Updated | 2018-05-17T17:18:42.000+0000 | 
Description
	From the meeting minutes from "How to measure V8 performance" held on April 30, 2013:
Consensus that Appcelerator should iteratively build their own JavaScript benchmark suite that measures Titanium use cases. This is a perquisite to experiment #4 "Titanium apps are slow because JavaScript is slow and ASM JS will make them run faster" and experiment #8 "The deployed JavaScript engines are slow and should be upgraded". See this page for the list of experiments :https://wiki.appcelerator.org/display/spe/Platform+Performance+Improvements
The Titanium JavaScript benchmark should not run on Titanium initially. Instead it should be a native app on each platform that exercises Titanium JavaScript use cases. The allows a rapid iteration on measuring Titanium use cases without requiring a port of the Kroll layer to each engine on each platform. A separate effort to identify these Titanium JavaScript use cases is requried.
Three engines should initially be included: JavaScriptCore, V8 and SpiderMonkey. The Code Processor was identified as a fourth engine possibility. It was roughly estimated that the Code Processor could be self-hosted in 4 weeks, so this it should be included in longer range plans.
No comments