From 2b1258a28e074ce4ad5db7d00d62b1ea6b152d4d Mon Sep 17 00:00:00 2001 From: Luke Wagner Date: Wed, 10 Jun 2015 00:06:26 -0500 Subject: Add Epic demo to table --- FAQ.md | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'FAQ.md') diff --git a/FAQ.md b/FAQ.md index b791676..f0d747f 100644 --- a/FAQ.md +++ b/FAQ.md @@ -4,7 +4,8 @@ Yes, this is a [high-level goal](HighLevelGoals.md) and there is a [prototype](https://github.com/WebAssembly/polyfill) with demos -[[1](http://lukewagner.github.io/AngryBotsPacked), [2](TODO)]. Although the +[[1](http://lukewagner.github.io/AngryBotsPacked), +[2](http://lukewagner.github.io/PlatformerGamePacked)]. Although the [binary format](BinaryEncoding.md) is not yet specified in any detail, the format used by the prototype has promising initial experimental results. To allow direct comparison with asm.js, the prototype has a tool to [pack asm.js](https://github.com/WebAssembly/polyfill/blob/master/src/pack-asmjs.cpp#L3117) @@ -14,7 +15,7 @@ after gzip compression: | Demo | asm.js | binary | `gzip` asm.js | `gzip` binary | |------|--------|--------|---------------|---------------| | [AngryBots](http://lukewagner.github.io/AngryBotsPacked) | 19MiB | 6.3MiB | 4.1MiB | 3.0MiB | -| TODO | +| [PlatformerGame](http://lukewagner.github.io/PlatformerGamePacked) | 49MiB | 18MiB | 11MiB | 7.3MiB | By writing the [decoder prototype in C++](https://github.com/WebAssembly/polyfill/blob/611ec5c8c41b08b112cf064ec49b13bf87e400cd/src/unpack.cpp#L2306) and Emscripten-compiling to asm.js, the polyfill is able to perform the translation to asm.js @@ -23,8 +24,8 @@ on an Intel® Xeon® E5-2665 @ 2.40GHz): | Demo | binary | time to decode into asm.js | |-------|--------|----------------------------| -| Unity | 6.3MiB | 240ms | -| TODO | +| [AngryBots](http://lukewagner.github.io/AngryBotsPacked) | 6.3MiB | 240ms | +| [PlatformerGame](http://lukewagner.github.io/PlatformerGamePacked) | 18MiB | 550ms | Since the polyfill algorithm (at least in the prototype) is simple and single-pass, memory usage is basically the size of the input plus the size of the decoded text. -- cgit v1.2.3