Boost logo

Geometry :

Subject: Re: [geometry] SIMD optimizations
From: Vladimir Voznesensky (vvoznesensky_at_[hidden])
Date: 2016-01-29 01:39:51


Please, find the comments below.

29.01.2016, 02:23, "Adam Wulkiewicz" <adam.wulkiewicz_at_[hidden]>:
> Hi Vladimir,
>
>
> How would you like to zip/transpose it exactly? Instead of having a
> container of e.g. points, to have D containers of coordinates, where D
> is the dimension?
Exactly. The maximum number of points in each node/leaf would be determined by SIMD vector size.

> And what would you like to do with them? Would they still be used as
> coordinates?
For this topic, the answer is yes, still be coordinates.

> Would you like to change only the nodes contents, provide some
> lower-level algorithms e.g. testing for spatial overlap or would you
> like to use entirely different balancing and querying algorithms?
I'm not intended to change high-level balancing and querying math. Anyway, every algorithm here could benefit from SIMD optimizations.

Regards,
Vladimir


Geometry list run by mateusz at loskot.net