Comment by samaparicio

Liked this post. Would love more detail on GPU query offloading and on how different the engine looks compared to one that would run queries on the CPU, as that seems to be the primary innovation