Hi,
in fact I have not specied not used particular optimization arguments.
I have not disabled CastChecking and ClassMetadata, and I have not
explicitly set the optimization level (but the default is the maximum
value).
By using defaults my current configuration is:
*GWT Compiler Arguments -
For sure we release production code with the following flags
- disableCastCheckingtrue/disableCastChecking
- disableClassMetadatatrue/disableClassMetadata
Also you don't specify what version of IE you are testing with, I assume IE9
On Thursday, February 7, 2013 10:10:54 AM UTC-5,
Ouput is OBFUSCATED ;-)
Il giorno giovedì 7 febbraio 2013 16:10:54 UTC+1, Fabiano Tarlao ha scritto:
Hi,
in fact I have not specied not used particular optimization arguments.
I have not disabled CastChecking and ClassMetadata, and I have not
explicitly set the optimization level (but the
Il giorno giovedì 7 febbraio 2013 16:14:32 UTC+1, Paul Stockley ha scritto:
For sure we release production code with the following flags
- disableCastCheckingtrue/disableCastChecking
- disableClassMetadatatrue/disableClassMetadata
Also you don't specify what version of IE you are
JavaScript doesn't have a concept of type, so all type checks are really
just function calls. Both instanceof and casts need to be generated to have
JavaScript behave exactly like Java. Here's a quick demonstration:
public class Test implements EntryPoint {
public static class SomeObject {
One thing I will be interested to see is the effect
of enableClosureCompiler=true. It does significantly reduce the JS file
size in many cases. However, I have read people complaining that the
resulting code was slower.
On Thursday, February 7, 2013 10:53:21 AM UTC-5, Fabiano Tarlao wrote:
I can definitely confirm that with the closure compiler enabled the
compiled size drops on the few apps I've tried it on, on the order of 5-15%
(no hard and fast numbers yet, working on such a writeup now). I can
confirm both a performance and size improvement with turning off cast
checking, but I
I have released the code and executables,
it's not fully documented at the moment, and I'll definitively tune things
in the next days,
what to say.. I have used two standard benchmark codes sieve and GNU fft
(little modifications), and I have assured JIT to compile all the code
thanks to
Thanks Colin,
I think that make sense removing cast checks and metadata in production, I
have alway thought that getting cast errors in production is a good things,
but it looks marginal.
I think I'll add these optimizations in the next benchmarks.
Regards
Il giorno giovedì 7 febbraio 2013
I am not sure whether you turned on all GWT compiler arguments and some
turn off some Dev specific GWT features.
*GWT Compiler Arguments - *
- disableCastCheckingtrue/disableCastChecking
- disableClassMetadatatrue/disableClassMetadata
- enableClosureCompilertrue/enableClosureCompiler
Fabiano,
Well done! Those are some interesting insights.
Thanks.
Rick
On Saturday, February 2, 2013 6:50:40 PM UTC-6, Fabiano Tarlao wrote:
Hi,
I have wrote a simple benchmark suite in java and I have run with JavaSE
1.7.0 and, thanks to GWT, I have run the same code on
Hi,
I have wrote a simple benchmark suite in java and I have run with JavaSE
1.7.0 and, thanks to GWT, I have run the same code on Firefox,Chrome,MSIE
and Opera.
My results, with the experiment details are published
12 matches
Mail list logo