[hamradio-commits] [gnss-sdr] 114/149: add neon implementation
Carles Fernandez
carles_fernandez-guest at moszumanska.debian.org
Sat Feb 6 19:43:10 UTC 2016
This is an automated email from the git hooks/post-receive script.
carles_fernandez-guest pushed a commit to branch next
in repository gnss-sdr.
commit cd2f0b86f601054fa6f164486e771453fedc80fa
Author: Carles Fernandez <carles.fernandez at gmail.com>
Date: Sat Jan 23 21:22:30 2016 +0100
add neon implementation
---
.../volk_gnsssdr_16ic_x2_multiply_16ic.h | 48 ++++++++++++++++++++++
1 file changed, 48 insertions(+)
diff --git a/src/algorithms/libs/volk_gnsssdr_module/volk_gnsssdr/kernels/volk_gnsssdr/volk_gnsssdr_16ic_x2_multiply_16ic.h b/src/algorithms/libs/volk_gnsssdr_module/volk_gnsssdr/kernels/volk_gnsssdr/volk_gnsssdr_16ic_x2_multiply_16ic.h
index 0c6374a..83dfb8a 100644
--- a/src/algorithms/libs/volk_gnsssdr_module/volk_gnsssdr/kernels/volk_gnsssdr/volk_gnsssdr_16ic_x2_multiply_16ic.h
+++ b/src/algorithms/libs/volk_gnsssdr_module/volk_gnsssdr/kernels/volk_gnsssdr/volk_gnsssdr_16ic_x2_multiply_16ic.h
@@ -161,4 +161,52 @@ static inline void volk_gnsssdr_16ic_x2_multiply_16ic_u_sse2(lv_16sc_t* out, con
}
#endif /* LV_HAVE_SSE2 */
+#ifdef LV_HAVE_NEON
+#include <arm_neon.h>
+
+static inline void volk_gnsssdr_16ic_x2_multiply_16ic_neon(lv_16sc_t* out, const lv_16sc_t* in_a, const lv_16sc_t* in_b, unsigned int num_points)
+{
+ lv_16sc_t *a_ptr = (lv_16sc_t*) in_a;
+ lv_16sc_t *b_ptr = (lv_16sc_t*) in_b;
+ unsigned int quarter_points = num_points / 4;
+ int16x4x2_t a_val, b_val, c_val;
+ int16x4x2_t tmp_real, tmp_imag;
+ unsigned int number = 0;
+
+ for(number = 0; number < quarter_points; ++number)
+ {
+ a_val = vld2_s16((int16_t*)a_ptr); // a0r|a1r|a2r|a3r || a0i|a1i|a2i|a3i
+ b_val = vld2_s16((int16_t*)b_ptr); // b0r|b1r|b2r|b3r || b0i|b1i|b2i|b3i
+ __builtin_prefetch(a_ptr + 4);
+ __builtin_prefetch(b_ptr + 4);
+
+ // multiply the real*real and imag*imag to get real result
+ // a0r*b0r|a1r*b1r|a2r*b2r|a3r*b3r
+ tmp_real.val[0] = vmul_s16(a_val.val[0], b_val.val[0]);
+ // a0i*b0i|a1i*b1i|a2i*b2i|a3i*b3i
+ tmp_real.val[1] = vmul_s16(a_val.val[1], b_val.val[1]);
+
+ // Multiply cross terms to get the imaginary result
+ // a0r*b0i|a1r*b1i|a2r*b2i|a3r*b3i
+ tmp_imag.val[0] = vmul_s16(a_val.val[0], b_val.val[1]);
+ // a0i*b0r|a1i*b1r|a2i*b2r|a3i*b3r
+ tmp_imag.val[1] = vmul_s16(a_val.val[1], b_val.val[0]);
+
+ // store the results
+ c_val.val[0] = vsub_s16(tmp_real.val[0], tmp_real.val[1]);
+ c_val.val[1] = vadd_s16(tmp_imag.val[0], tmp_imag.val[1]);
+ vst2_s16((int16_t*)out, c_val);
+
+ a_ptr += 4;
+ b_ptr += 4;
+ out += 4;
+ }
+
+ for(number = quarter_points * 4; number < num_points; number++)
+ {
+ *out++ = (*a_ptr++) * (*b_ptr++);
+ }
+}
+#endif /* LV_HAVE_NEON */
+
#endif /*INCLUDED_volk_gnsssdr_16ic_x2_multiply_16ic_H*/
--
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/pkg-hamradio/gnss-sdr.git
More information about the pkg-hamradio-commits
mailing list