The output of EMTERPRETER_ADVISE is "67% out of 2554 functions", which
agrees with our previous conclusion with Asyncify.


regards,
- Lu

On Thu, Mar 5, 2015 at 1:50 PM, Lu Wang <[email protected]> wrote:

> I remembered that we have pessimistic analysis for function pointers, and
> we left that on for safety.
> I'll do a static analysis later. I wonder if it can be integrated, i.e. I
> trust emscripten such that it will decide which functions to be interpreted
> and which are not, instead of copying the arguments and paste to emscripten
> every time.
>
>
> regards,
> - Lu
>
>
> On Thu, Mar 5, 2015 at 11:51 AM, Alon Zakai <[email protected]> wrote:
>
>> Wow, that is faster than I would expect (both startup and later)! Yes,
>> without a blacklist or whitelist, it is 100% in the interpreter.
>>
>> With blacklist it could probably be even faster. Actually, I'd be curious
>> how the emterpreter static analysis does, if you feel like trying it,
>>
>> https://github.com/kripken/emscripten/wiki/Emterpreter#static-analysis
>>
>> I wonder if the size of the list will be most of the program, which I
>> think was the case with asyncify? The difference might that the emterpreter
>> analysis uses function pointer types, which I think asyncify didn't (but
>> could add)?
>>
>> - Alon
>>
>>
>> On Tue, Mar 3, 2015 at 8:32 PM, Lu Wang <[email protected]> wrote:
>>
>>> Hi all,
>>>
>>>    Recently I tried EMTERPRETER on Vim.js, and I got this:
>>> http://coolwanglu.github.io/vim.js/emterpreter/vim.html
>>>
>>>    I was compiling using -s EMTERPRETIFY=1 -s EMTERPRETIFY_ASYNC=1,
>>> without any WHITELIST or BLACKLIST, which means all the functions are
>>> interpreted? The file size is reduced significantly, comparing with the 26M
>>> monster-sized ASM.js version, this version with EMTERPRETER is only 2.5M,
>>> which improves the startup time a lot.
>>>
>>>   While I was planning to optimize after making it work, I found the
>>> performance is good enough, I don't feel any delay for normal navigation or
>>> editing operations, but it does become slightly slower for commands like
>>> `%s/^/hello/g`, which inserts 'hello' at the beginning of each line.
>>>
>>>   This interpreter is surprisingly faster than I had expected, probably
>>> there were already lots of async operations in Vim maybe? Anyway I'm quite
>>> happy with this new technique!
>>>
>>>   Thanks!
>>>
>>>
>>>    regards,
>>>    - Lu
>>>
>>>
>>> --
>>> You received this message because you are subscribed to the Google
>>> Groups "emscripten-discuss" 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/d/optout.
>>>
>>
>>  --
>> You received this message because you are subscribed to a topic in the
>> Google Groups "emscripten-discuss" group.
>> To unsubscribe from this topic, visit
>> https://groups.google.com/d/topic/emscripten-discuss/yfvDabeq8TM/unsubscribe
>> .
>> To unsubscribe from this group and all its topics, send an email to
>> [email protected].
>> For more options, visit https://groups.google.com/d/optout.
>>
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"emscripten-discuss" 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/d/optout.

Reply via email to