Created
October 21, 2015 17:48
-
-
Save mhroth/2b494cc18b1c6eabf4e4 to your computer and use it in GitHub Desktop.
A basic NEON implementation of SSE _mm_movemask_ps
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
uint32_t _mm_movemask_ps(float32x4_t x) { | |
uint32x4_t mmA = vandq_u32( | |
vreinterpretq_u32_f32(x), (uint32x4_t) {0x1, 0x2, 0x4, 0x8}); // [0 1 2 3] | |
uint32x4_t mmB = vextq_u32(mmA, mmA, 2); // [2 3 0 1] | |
uint32x4_t mmC = vorrq_u32(mmA, mmB); // [0+2 1+3 0+2 1+3] | |
uint32x4_t mmD = vextq_u32(mmC, mmC, 3); // [1+3 0+2 1+3 0+2] | |
uint32x4_t mmE = vorrq_u32(mmC, mmD); // [0+1+2+3 ...] | |
return vgetq_lane_u32(mmE, 0); | |
} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Unfortunately a NEON analog to SSE's
_mm_movemask_ps
does not exist. There are some reimplementations (e.g. here), this one is mine. The input is assumed to be the result of a comparison (e.g. consisting of either0x0
or~0x0
). This implementation is "reasonably" fast.