Always use the correct numeric type for all kernels

[tom91136]

There are several instances where the double constant 0.0 is used in a way that promotes everything it touches.
For example:

BabelStream/src/std-data/STDDataStream.cpp

Line 85 in 1d423fc

return std::transform_reduce(exe_policy, a.begin(), a.end(), b.begin(), 0.0);

In this case, the value is used on a templated function where the deducted type is based on that constant, so the kernel essentially gets fixed to double precision and not T.

There are several other models that does T sum = 0.0; which is probably OK.
But for consistency I think we should probably be using aggregate initialisation for everything:

std::transform_reduce(..., T{}); 
T sum{};  

[tomdeakin]

LGTM, apply this (with a comment in the code)