https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82989
--- Comment #16 from Wilco <wilco at gcc dot gnu.org> --- (In reply to Matthijs van Duin from comment #15) > (In reply to Wilco from comment #14) > > Yes on older cores it can be a bad idea to allow accidental use of Neon > > instructions. The simplest workaround is to switch off Neon, just use > > -mfpu=vfp. > > Sure, though that's not exactly ideal since it is very desirable for the > Neon unit to be used for single-precision floats. (9-10 cycles for add/sub > in VFP, 10-12 cycles for mul in VFP, 1 cycle for two add/sub/mul ops in Neon) That's why it's a workaround until the bug is fixed - assuming the performance critical functions use both 64-bit integer and 32-bit fp operations, you could benchmark both options and choose the one that works best for now.
