1: \begin{abstract}

2: Astrophysical black holes could be nearly extremal (that is,

3: rotating nearly as fast as possible);

4: therefore, nearly extremal black holes could be

5: among the binaries

6: that current and future

7: gravitational-wave observatories will detect. Predicting the gravitational

8: waves emitted by merging black holes requires numerical-relativity

9: simulations, but these simulations are especially challenging 

10: when one or both holes have mass $m$ and spin $S$ 

11: exceeding the Bowen-York limit of 

12: $S/m^2=0.93$.

13: We present improved

14: methods that enable us to simulate merging, nearly extremal black holes more

15: robustly and more efficiently. We use these methods to

16: simulate an unequal-mass, precessing binary black hole coalescence,

17: where the larger black hole has

18: $S/m^2=0.99$.

19: We also use these methods to simulate

20: a non-precessing binary black hole coalescence, where both black holes have

21: $S/m^2=0.994$, nearly reaching

22: the Novikov-Thorne upper bound for

23: holes spun up by thin accretion disks. 

24: We demonstrate numerical convergence 

25: and estimate the numerical errors of 

26: the waveforms;

27: we compare numerical waveforms from

28: our

29: simulations with post-Newtonian 

30: and effective-one-body waveforms;

31: we compare the evolution of the black-hole

32: masses and spins with 

33: analytic predictions; and we explore 

34: the effect of increasing spin magnitude on 

35: the orbital dynamics (the so-called ``orbital hangup'' effect).

36: \end{abstract}

37: