For servers no, but we’re talking about endpoints here. Also this isn’t only about reducing the existing vector bandwidth but also about not increasing it outsi

For servers no, but we’re talking about endpoints here. Also this isn’t only about reducing the existing vector bandwidth but also about not increasing it outside of dedicated co-processors.