[Interest] SIMD accelerated wrappers for relevant Qt container classes (QVector, ...) or QByteArray?

René J.V. Bertin rjvbertin at gmail.com
Tue May 9 19:01:11 CEST 2017


On Tuesday May 09 2017 18:59:17 Konstantin Shegunov wrote:

Hi,

>I don't know what exactly you're going to do/need, but for math you

Me neither, for now it's just a general question. The answers to which will probably hint at what I could do with them :)

>should consider using a specialized library (e.g. Eigen -
>http://eigen.tuxfamily.org/index.php?title=Main_Page); I don't believe
>Qt features such optimizations.

Exactly, hence the question! The specialised libraries (Eigen, Vc) have their own APIs, AFAIK only the old MacSTL aims to be a drop-in replacement to STL classes. That drop-in replacement aspect is very convenient: it would probably not be hard to find applications for an accelerated QVector (or so) that speeds up things like calculating a CRC with minimal effort (and most of the time without adding a library dependency).

R.



More information about the Interest mailing list