Thread overview
CPU cores & threads & fibers
Jun 14, 2015
Robert M. Münch
Jun 14, 2015
John Colvin
Jun 14, 2015
Etienne Cimon
Jun 15, 2015
Robert M. Münch
Jun 16, 2015
Rob T
Jun 18, 2015
Robert M. Münch
June 14, 2015
Hi, just to x-check if I have the correct understanding:

fibers 	= look parallel, are sequential 	=> use 1 CPU core
threads 	= look parallel, are parallel 	=> use several CPU cores

Is that right?

-- 
Robert M. Münch
http://www.saphirion.com
smarter | better | faster

June 14, 2015
On Sunday, 14 June 2015 at 12:35:44 UTC, Robert M. Münch wrote:
> Hi, just to x-check if I have the correct understanding:
>
> fibers 	= look parallel, are sequential 	=> use 1 CPU core
> threads 	= look parallel, are parallel 	=> use several CPU cores
>
> Is that right?

Pretty much.
June 14, 2015
On 2015-06-14 08:35, Robert M. Münch wrote:
> Hi, just to x-check if I have the correct understanding:
>
> fibers     = look parallel, are sequential     => use 1 CPU core
> threads     = look parallel, are parallel     => use several CPU cores
>
> Is that right?
>

Yes, however nothing really guarantees multi-threading = multi-core. The kernel reserves the right and will most likely do everything possible to keep your process core-local to use caching efficiently.

There's a few ways around that though

https://msdn.microsoft.com/en-us/library/windows/desktop/ms686247%28v=vs.85%29.aspx
http://man7.org/linux/man-pages/man2/sched_setaffinity.2.html
June 14, 2015
On Sunday, 14 June 2015 at 12:35:44 UTC, Robert M. Münch wrote:
> Hi, just to x-check if I have the correct understanding:
>
> fibers 	= look parallel, are sequential 	=> use 1 CPU core
> threads 	= look parallel, are parallel 	=> use several CPU cores
>
> Is that right?

Fibers/co-routines run on a thread and is conceptually the same as a functor/object-with-method that can suspend itself and hold onto the state until it is restarted. Like yield in Python generators.

Fibers have their own stack, but that is an implementation detail. It is possible to do the same thing with object-method-calls if you have stackless code-generation (D does not support stackless runtimes, but you'll find this in other languages).

June 15, 2015
On 2015-06-14 15:54:30 +0000, Etienne Cimon said:

> Yes, however nothing really guarantees multi-threading = multi-core. The kernel reserves the right and will most likely do everything possible to keep your process core-local to use caching efficiently.

Hi, sure. It's more about raising the chance to use the cores ;-)

-- 
Robert M. Münch
http://www.saphirion.com
smarter | better | faster

June 16, 2015
On Sunday, 14 June 2015 at 15:54:30 UTC, Etienne Cimon wrote:
> On 2015-06-14 08:35, Robert M. Münch wrote:
>> Hi, just to x-check if I have the correct understanding:
>>
>> fibers     = look parallel, are sequential     => use 1 CPU core
>> threads     = look parallel, are parallel     => use several CPU cores
>>
>> Is that right?
>>
>
> Yes, however nothing really guarantees multi-threading = multi-core. The kernel reserves the right and will most likely do everything possible to keep your process core-local to use caching efficiently.
>
> There's a few ways around that though
>
> https://msdn.microsoft.com/en-us/library/windows/desktop/ms686247%28v=vs.85%29.aspx
> http://man7.org/linux/man-pages/man2/sched_setaffinity.2.html

FYI:

https://issues.dlang.org/show_bug.cgi?id=11686
https://issues.dlang.org/show_bug.cgi?id=11687


June 18, 2015
On 2015-06-16 18:36:09 +0000, Rob T said:

> FYI:
> 
> https://issues.dlang.org/show_bug.cgi?id=11686
> https://issues.dlang.org/show_bug.cgi?id=11687

Thanks. We are currently experimenting to see how want to use the threads and what code to refactor. If we are going to bite the bullet I keep this in mind, maybe we can contribute something along.

-- 
Robert M. Münch
http://www.saphirion.com
smarter | better | faster