For linear algebra, there is a significant overhead when using the packed format. This is partly because the indexing operations are slightly more complicated but more importantly, it is because the compact storage doesn’t allow blocked (BLAS-3) linear algebra operations. This was the main reason I decided not to use the compact storage scheme when I implemented XTriangular and Symmetric since I thought that 2x memory saving was not sufficient to justify slower linear algebra. However, for other applications, this might, of course, be completely different.
An interesting alternative is the Rectangular Full Packaged storage scheme which maintains BLAS-3 performance without wasting half the memory. However, I never really needed the memory saving so the development stalled.