Thank you... avoiding growth and setting 2gb - 64kb ended up being the most 
reliable approach.

On Monday, March 18, 2019 at 9:01:52 PM UTC-5, Alon Zakai wrote:
>
> Memory never shrinks, once grown.
>
> There's no power of 2 requirement - asm.js requires multiples of 16MB, and 
> wasm 64Kb.
>
> On Mon, Mar 18, 2019 at 5:42 PM John Muehlhausen <[email protected] 
> <javascript:>> wrote:
>
>> I tried doing a malloc/memset/free of 1800000000ull at the beginning of 
>> my program, which seemed to fix the problem.  Once the heap grows this 
>> large, is it ever downsized?  If it is downsized then I’m not sure why this 
>> worked, because it seems like there would be no effect from the temporary 
>> usage.
>>
>> I don’t think I can set 2gb - 64k because I think the initial memory was 
>> limited to powers of two for a security reason?  So 1gb is the largest 
>> initial....
>>
>> On Mon, Mar 18, 2019 at 7:16 PM Alon Zakai <[email protected] 
>> <javascript:>> wrote:
>>
>>> Yes, when memory grows in asm.js we create a new one, and the old one 
>>> will be GC'd eventually, but when depends on the browser. So both will be 
>>> in memory at once for a while, which can be a problem.
>>>
>>> Setting 2GB on Chrome is an issue currently, but I think you can set 2GB 
>>> - 64K, so you can get close.
>>>
>>> On Mon, Mar 18, 2019 at 4:56 PM John Muehlhausen <[email protected] 
>>> <javascript:>> wrote:
>>>
>>>> It seems as if my entire heap is being duplicated (due to 
>>>> ALLOW_MEMORY_GROWTH) and then the old one is garbage collected?  Is that 
>>>> how it works?  Seems like I cannot set an initial size more than 1GB and 
>>>> the "growth" increment is less than 1GB.  What I'd really like is to set 
>>>> 2GB (the max?) initially, but it seems as if this doesn't work?
>>>>
>>>> On Monday, March 18, 2019 at 4:39:18 PM UTC-5, John Muehlhausen wrote:
>>>>>
>>>>> I have a memory-intensive process that runs fine on Firefox 
>>>>> (everywhere) and Chrome (Max+Linux) and Safari.  However, Chrome on 
>>>>> Windows 
>>>>> "aw, snap"'s me most-- but not all-- of the time.  I was finally able to 
>>>>> resolve this with `--max_old_space_size=4096` Chrome command line 
>>>>> parameter, but I don't want to place this requirement on my users.
>>>>>
>>>>> My guess is that gc is not keeping up with me on Windows in particular 
>>>>> (fewer points where gc wants to run? ) or else that the defaults differ 
>>>>> per 
>>>>> platform.  Any ideas about how to nudge garbage collection to happen more 
>>>>> often?  Any other thoughts?
>>>>>
>>>>> Thanks!
>>>>>
>>>> -- 
>>>> 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] 
>>>> <javascript:>.
>>>
>>>
>>>> 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/1V1aoe7Pxlg/unsubscribe
>>> .
>>> To unsubscribe from this group and all its topics, send an email to 
>>> [email protected] <javascript:>.
>>> 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] <javascript:>.
>> 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