Subject: Re: [boost] [compute] kernels as strings impairs readability and maintainability
From: Kyle Lutz (kyle.r.lutz_at_[hidden])
Date: 2014-12-23 17:26:55
On Tue, Dec 23, 2014 at 1:54 PM, AsbjÃ¸rn <lordcrc_at_[hidden]> wrote:
> On 23.12.2014 20:21, Kyle Lutz wrote:
>> Another concern is that Boost.Compute is a header-only library and
>> doesn't control the build system or how it the library will be loaded.
>> This limits our ability to pre-compile certain programs and "install"
>> them for later use by the library.
> Given that Boost.Compute relies on Boost.Chrono, and Boost.Chrono is not a
> head-only library, I don't see a strong reason why Boost.Compute should stay
> header-only if that blocks some interesting and/or important features.
> Just my 2 cents as a library user :)
Boost.Compute only relies on the header-only porition of Boost.Chrono
(specifically <boost/chrono/duration.hpp>) so it remains header-only.
And I agree that being non-header-only would allow some features to be
developed more easily. I have considered having a
"BOOST_COMPUTE_NO_HEADER_ONLY" flag which would enable a compiled
version of Boost.Compute which could support more features like this
and also potentially improve compile times (though I would still want
to keep a large majority of Boost.Compute's functionality available in
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk