Lucas Mongrain

What I'm trying to say is maybe there's error with the way I launch sde, because function that's suppose to work doesn't

Hi @ihhub I could jump on this after the avx512 implementation

oh this project sounds fun! But do you have an idea how can we do that? We have to make a different alignment size for SSE (16-bit) and AVX (32-bit)....

I think the easiest way to go is to make or use a specific allocator to std::vector

we cannot assume that std::vector is aligned, but according to the template prototype: `template < class T, class Alloc = allocator > class vector;` we can pass an allocator to...

some people on stack overflow have tried this: https://stackoverflow.com/questions/12942548/making-stdvector-allocate-aligned-memory

yeah that's true...

maybe we can give to the end user a "special" kind of vector, by doing a `using image_vector = vector` and we change the function call to projectionProfile and Accumulate

but the image also use an std::vector, no?

oh I understand now... ROI make everything more difficult... the only way I can think is copying the ROI in an aligned memory space... but I think, that will hurt...