commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Gilles <gil...@harfang.homelinux.org>
Subject Re: [Math] MATH-1371: Provide accelerated kmeans++ implementation
Date Tue, 31 May 2016 00:31:22 GMT
On Tue, 31 May 2016 03:10:20 +0300, Artem Barger wrote:
>>
>>>>
>>>> ​Yes, you can parallelize it, though it will cancel several 
>>>> optimizations
>>> I've added.
>>> In fact you can partition the input according to number of threads 
>>> you'd
>>> like to use
>>> and make each thread to take care of relevant data chunk.
>>>
>>> I guess it will increase performance, not sure why current 
>>> implementation
>>> wasn't
>>> parallelized.​
>>>
>>>
>> You are most welcome to enhance it. ;-)
>>
>
> Well, already did it ;-)
> And if you mean to parallelize my current implementation for Elkan
> algorithm,
> I think that it should be handled as a separate task.​
>

What I mean is that algorithms that can perform their computation
on multiple threads should be implemented in such a way that the
feature can be used.

Of course, that means a more complex code, to ensure thread-safety.
So, for sure, your code could be added now (modulo the remarks)
if you don't want to handle that now.v And as you pointed out the
other implementations are not multi-thread either.

But it's a direction worth exploring I think.

Gilles



---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org


Mime
View raw message