# The nilpotency class and the derived length of a $p$-group

Let $G$ be a finite $p$-group of nilpotency class $c$ and of derived length $d$. As is well known, we have $d\leq \lfloor\log_2 c\rfloor+1$,

Does there exist any information or any classification of finite $p$-group $G$, where $d=\lfloor\log_2 c\rfloor+1$?

Any answer or comment will be greatly appreciated!

• I don't have any deep thoughts about it, but just an observation that if the class is very large, then the coclass is small and then the derived length is also small. So you won't have equality in this case. – Yiftach Barnea Mar 22 '18 at 23:40
• Another stupid observation any group of class 2 satisfies this condition. Thus, as observed by the answer below, a classification is unlikely. Maybe a more sensible question is to ask for a group of order $p^n$ what is the maximal class for which such equality can still be achieved? – Yiftach Barnea Mar 23 '18 at 8:02

My guess is "no", at least as regards a classification. A check with GAP shows that 30,591 of the 34,297 groups of order $5^7$ satisfy this equality; that 7,882 of the 9,310 groups of order $3^7$ satisfy the equality; and that 53,499 of the 56,092 groups of order $2^8$ satisfy the equality.
glist:=AllSmallGroups(Size,5^7,G->DerivedLength(G) = 1 + LogInt( NilpotencyClassOfGroup(G) , 2));; NrSmallGroups(5^7); Length(glist);