Subject: Re: [boost] interest in structure of arrays container?
From: degski (degski_at_[hidden])
Date: 2016-10-16 02:59:02
On 16 October 2016 at 08:36, Michael Marcin <mike.marcin_at_[hidden]> wrote:
> It doesn't have the wasted duplicated information that a tuple of
> std::vectors has.
In your example, which I think is flawed as an example, you create in
memory a 20+ GB data structure (hope I got the maths right). I realised
this after a std::bad_alloc was thrown. Some duplicate info in the
std::vector(s) seems rather irrelevant.
This data structure optimizes one operation (calculating the average) at
the cost of pessimising almost any other operation (relocation if the
vector needs to grow, insert, delete, push_back etc are all done 4 times in
your example), while iterating over the records will land you right into
cache-miss-heaven. Keeping a running average while you read/input/delete
the data gives you an O(1) average. Consider using a pector
<https://github.com/aguinet/pector> instead of a vector.
You state that the example is a toy example. To me the example shows that
iterating over a vector of smaller objects (pods) is faster than iterating
over a vector of larger objects, duh. The real use case might be more
interesting, maybe you can describe it.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk