Lines Matching full:shifts
3324 // Lower vector shifts on NEON to use VSHL.
3327 // Left shifts translate directly to the vshiftu intrinsic.
3336 // NEON uses the same intrinsics for both left and right shifts. For
3337 // right shifts, the shift amounts are negative, so negate the vector of
3377 // First, build a SRA_FLAG/SRL_FLAG op, which shifts the top part by one and
3382 // The low part is an ARMISD::RRX operand, which shifts the carry in.
6647 // Vector shifts: check for immediate versions and lower them.
6718 // Narrowing shifts require an immediate right shift.
6800 /// PerformShiftCombine - Checks for immediate versions of vector shifts and
6809 // Nothing to be done for scalar shifts.
