If this is purely JavaScript related you can use d8 (the v8 command line
profiler) to do the profiling for you.

I recently added source maps support to it so you can actually profile
obfuscated GWT code just fine.

Here is what you need to do:
1. Build v8 yourself
2. Compile your code using a custom linker like this one:
https://gwt-review.googlesource.com/#/c/4220 (only one permutation, turn on
source maps)
3. run that through d8 (e.g. d8 -prof mymodule.nocache.js)
4. use the tickprocessor to see the profile (e.g.
tools/linux-tick-processor --source-map=yoursourcemap.json v8.log

-Daniel


On Mon, Aug 26, 2013 at 8:45 PM, Goktug Gokdogan <[email protected]> wrote:

> The latest benchmarks (before the removal Benchmarks suite) was done by
> custom infra:
>
> https://gwt.googlesource.com/gwt/+/master/reference/Microbenchmarks/src/com/google/gwt/reference/microbenchmark
>
> If you would like to collect timing from some previous GWTTestCases you
> can do that by overriding some base methods on GWTTestcase (runBare
> includes gwtSetup and doRunTest does not) and log the timings.
>
> PS: Old benchmark suite was a hack over GWTTestCase infra that was adding
> a lot of complexity without much benefit.
>
>
>
> On Mon, Aug 26, 2013 at 11:16 AM, John A. Tamplin <[email protected]> wrote:
>
>> On Mon, Aug 26, 2013 at 1:47 PM, Colin Alworth <[email protected]>wrote:
>>
>>> Shortly after 2.5.1, the Benchmark classes were removed from GWT (
>>> https://gwt.googlesource.com/gwt/+/39eb6001a037fd8b6580a73a2540e6e9c04e54c2and
>>> https://gwt.googlesource.com/gwt/+/00c7ce43df3a629b7302ab902a07431db7224e2b)
>>> - what are folks using for low-level performance testing these days?
>>>
>>> I'm wrapping up a changeset to emit JS TypedArrays when int[], double[],
>>> etc are used in code, and while I can see that some projects clearly
>>> improve with this, others seem to take a hit (specifically in rapidly
>>> allocating arrays that have a short lifetime). I'd like to record some
>>> numbers to paint this picture without rolling my own benchmark/gwtPerf
>>> tool, but I'm not quite clear on the current state of this sort of metrics.
>>>
>>> From the comment that originally deprecated the Benchmark class:
>>>
>>>> This didn't get enough adaption externally or internally. The way it's
>>>> implemented requires special support from GWTTestCase infrastructure. Also,
>>>> we are not going have an environment to automate benchmarks written on this
>>>> any time soon.
>>>>
>>>> If there is enough demand later, parts of this code can be resurrected
>>>> to build a new benchmark system.
>>>>
>>>
>>> Thoughts?
>>>
>>
>> I argued against removing it at the time, since I found it useful for JRE
>> work in particular.  I'm not sure what sort of demand would be required to
>> get it reinstated, but it would probably mean someone taking ownership of
>> it.
>>
>> --
>> John A. Tamplin
>>
>> --
>> http://groups.google.com/group/Google-Web-Toolkit-Contributors
>> ---
>> You received this message because you are subscribed to the Google Groups
>> "GWT Contributors" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to [email protected].
>> For more options, visit https://groups.google.com/groups/opt_out.
>>
>
>  --
> http://groups.google.com/group/Google-Web-Toolkit-Contributors
> ---
> You received this message because you are subscribed to the Google Groups
> "GWT Contributors" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> For more options, visit https://groups.google.com/groups/opt_out.
>



-- 
Google Germany GmbH
*Dienerstr. 12*
*80331 München*

Registergericht und -nummer: Hamburg, HRB 86891
Sitz der Gesellschaft: Hamburg
Geschäftsführer: Graham Law, Katherine Stephens

-- 
http://groups.google.com/group/Google-Web-Toolkit-Contributors
--- 
You received this message because you are subscribed to the Google Groups "GWT 
Contributors" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to