[Caml-list] muti-core programming

[Caml-list] muti-core programming

[刘坚]

[-- Attachment #1: Type: text/plain, Size: 459 bytes --]

Hi,

         I’m recently writing a formal verification tool in OCaml, and it
works really well, but I’m considering writing a concurrent version.
However, until now, there seems to be no way to write programs that take
advantage of multi-cores. So, I’m wondering when will OCaml support
multi-core programming? Or else, do I have other choices by using some
external extensions of OCaml instead of the standard library?

 

Thanks,

Jian


[-- Attachment #2: Type: text/html, Size: 2610 bytes --]

Re: [Caml-list] muti-core programming

[Francois Berenger]

On 03/09/2016 10:50 AM, 刘坚 wrote:
> Hi,
>
>           I’m recently writing a formal verification tool in OCaml, and
> it works really well, but I’m considering writing a concurrent version.
> However, until now, there seems to be no way to write programs that take
> advantage of multi-cores.

To accelerate something, probably you want paralellism, not concurrency.

I recomend parmap, but there are some other libraries out there
too for that purpose (in opam: forkwork and probably others I don't know).

But be careful that too fine granularity calculations don't parallelize 
well.
For example, if you are analyzing source code, maybe analyzing
distinct files in parallel would be a coarse enough granularity.

 > So, I’m wondering when will OCaml support
> multi-core programming? Or else, do I have other choices by using some
> external extensions of OCaml instead of the standard library?
>
> Thanks,
>
> Jian

-- 
Regards,
Francois.
"When in doubt, use more types"

Re: [Caml-list] muti-core programming

[Mohamed Iguernlala]

Hi,

Functory may be suitable as well 
(http://opam.ocaml.org/packages/functory/functory.0.5/)

--
Mohamed Iguernlala.
Senior R&D Engineer, OCamlPro SAS
Research Associate, VALS team, LRI

Le 09/03/2016 11:07, Francois Berenger a écrit :
> On 03/09/2016 10:50 AM, 刘坚 wrote:
>> Hi,
>>
>>           I’m recently writing a formal verification tool in OCaml, and
>> it works really well, but I’m considering writing a concurrent version.
>> However, until now, there seems to be no way to write programs that take
>> advantage of multi-cores.
>
> To accelerate something, probably you want paralellism, not concurrency.
>
> I recomend parmap, but there are some other libraries out there
> too for that purpose (in opam: forkwork and probably others I don't 
> know).
>
> But be careful that too fine granularity calculations don't 
> parallelize well.
> For example, if you are analyzing source code, maybe analyzing
> distinct files in parallel would be a coarse enough granularity.
>
> > So, I’m wondering when will OCaml support
>> multi-core programming? Or else, do I have other choices by using some
>> external extensions of OCaml instead of the standard library?
>>
>> Thanks,
>>
>> Jian
>

Re: [Caml-list] muti-core programming

[Francois Berenger]

On 03/09/2016 11:23 AM, Mohamed Iguernlala wrote:
> Hi,
>
> Functory may be suitable as well
> (http://opam.ocaml.org/packages/functory/functory.0.5/)

and maybe lwt-parallel by Ivan Gotovchits or
procord by Cryptosense; all available libraries in opam.

> --
> Mohamed Iguernlala.
> Senior R&D Engineer, OCamlPro SAS
> Research Associate, VALS team, LRI
>
> Le 09/03/2016 11:07, Francois Berenger a écrit :
>> On 03/09/2016 10:50 AM, 刘坚 wrote:
>>> Hi,
>>>
>>>           I’m recently writing a formal verification tool in OCaml, and
>>> it works really well, but I’m considering writing a concurrent version.
>>> However, until now, there seems to be no way to write programs that take
>>> advantage of multi-cores.
>>
>> To accelerate something, probably you want paralellism, not concurrency.
>>
>> I recomend parmap, but there are some other libraries out there
>> too for that purpose (in opam: forkwork and probably others I don't
>> know).
>>
>> But be careful that too fine granularity calculations don't
>> parallelize well.
>> For example, if you are analyzing source code, maybe analyzing
>> distinct files in parallel would be a coarse enough granularity.
>>
>> > So, I’m wondering when will OCaml support
>>> multi-core programming? Or else, do I have other choices by using some
>>> external extensions of OCaml instead of the standard library?
>>>
>>> Thanks,
>>>
>>> Jian
>>
>
>

-- 
Regards,
Francois.
"When in doubt, use more types"

Re: [Caml-list] muti-core programming

[Yaron Minsky]

[-- Attachment #1: Type: text/plain, Size: 1972 bytes --]

rpc_parallel might also be of use.

On Wed, Mar 9, 2016 at 6:04 AM, Francois Berenger <
francois.berenger@inria.fr> wrote:

> On 03/09/2016 11:23 AM, Mohamed Iguernlala wrote:
>
>> Hi,
>>
>> Functory may be suitable as well
>> (http://opam.ocaml.org/packages/functory/functory.0.5/)
>>
>
> and maybe lwt-parallel by Ivan Gotovchits or
> procord by Cryptosense; all available libraries in opam.
>
>
> --
>> Mohamed Iguernlala.
>> Senior R&D Engineer, OCamlPro SAS
>> Research Associate, VALS team, LRI
>>
>> Le 09/03/2016 11:07, Francois Berenger a écrit :
>>
>>> On 03/09/2016 10:50 AM, 刘坚 wrote:
>>>
>>>> Hi,
>>>>
>>>>           I’m recently writing a formal verification tool in OCaml, and
>>>> it works really well, but I’m considering writing a concurrent version.
>>>> However, until now, there seems to be no way to write programs that take
>>>> advantage of multi-cores.
>>>>
>>>
>>> To accelerate something, probably you want paralellism, not concurrency.
>>>
>>> I recomend parmap, but there are some other libraries out there
>>> too for that purpose (in opam: forkwork and probably others I don't
>>> know).
>>>
>>> But be careful that too fine granularity calculations don't
>>> parallelize well.
>>> For example, if you are analyzing source code, maybe analyzing
>>> distinct files in parallel would be a coarse enough granularity.
>>>
>>> > So, I’m wondering when will OCaml support
>>>
>>>> multi-core programming? Or else, do I have other choices by using some
>>>> external extensions of OCaml instead of the standard library?
>>>>
>>>> Thanks,
>>>>
>>>> Jian
>>>>
>>>
>>>
>>
>>
> --
> Regards,
> Francois.
> "When in doubt, use more types"
>
> --
> Caml-list mailing list.  Subscription management and archives:
> https://sympa.inria.fr/sympa/arc/caml-list
> Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
> Bug reports: http://caml.inria.fr/bin/caml-bugs
>

[-- Attachment #2: Type: text/html, Size: 3422 bytes --]