divappr tests

Torbjorn Granlund tg at gmplib.org
Mon Dec 14 22:57:05 CET 2009

nisse at lysator.liu.se (Niels Möller) writes:

  bodrato at mail.dm.unipi.it writes:
  > I tried your patch, but "make tune" aborted with:
  > #define DC_DIVAPPR_Q_THRESHOLD        sbpi1_divappr_q.c:181: GNU MP
  > assertion failed: np[1] == n1
  > *** Signal 6
  It's the patch of the final case which doesn't satisfy the assumptions
  of that assert (by storing a result directly in np[1] instead of going
  via the n1 variable). So I'll delete that assert and check in.
I think you slightly pessimises things by no longer keeping the top
remainder limb in a register.  I very carefully optimized that code,
including that aspect.


More information about the gmp-devel mailing list