agozillon

Thanks Dave, I'll have a look over the changes!

You are correct Ronan, alignas appears to have something to do with the problem. A cl:sycl::vec is not actually the size of NumElements * sizeof(DataType) a few extra bytes are...

It appears that the number of boost classes being inherited from by small_array causes the additional overhead.. if it inherits from just 1 e.g: template struct small_array : std::array, boost::euclidean_ring_operators...

I believe the current pull request fixes the problem, it doesn't bring in any breaking changes to the test suite and the size of the cl::sycl::vec is what it should...

The problem appears to be a segmentation fault caused by the following section of code (FinalType conversion operator inside small_array) during swizzles: ``` /** Since the boost::operator work on the...

A type trait along the lines of: ``` template struct SYCLTypeAlignment { static constexpr int Alignment() { return 0; } }; template struct SYCLTypeAlignment { static constexpr int Alignment() {...

I don't think @keryell's original point was that namespaces would fix it all, this is intended as only part of the larger puzzle I believe. Although, as far as I'm...

With the addition of: https://github.com/intel/llvm/pull/345/files this task is likely to be less simple as it may require altering the Clang CodeGen a little in SYCLLowerIR/LowerWGScope to CodeGen SPIR instead of...

interesting, not too sure why as it wasn't any knowing fix on my end and it still isn't working on my machine, but I'll take a free win. But i'll...

Ah unfortunate the mystery remains open for now, all good I know the feeling!