Hi,<div>I send pixman patch for NEON optimizations for several bilinear scaled scanline functions.</div><div>Following functions are optimized.</div><div><br></div><div>over_8888_n_8888</div><div>add_8888_n_8888</div><div> src_8888_8_8888</div><div>over_8888_8_8888</div><div>add_8888_8_8888</div><div><br></div><div>This patch is based on pixman master branch with latest commit id = a2153222677327be43251012f462d19a7e98ce14. (Soeren's commit on April 3)</div> <meta http-equiv="content-type" content="text/html; charset=utf-8"><div>Because there can be some conflicts with latest commit of siarhei's bilinear optimizations.</div><div><br></div><div>It sill have lots of places to optimize, for example, preloading mask and destination pixels, better handling of two, one pixel case.</div> <div>However it gives us reasonable performance than before.</div><div>Review would be appreciated.</div><div><br></div><div>Performance Data on S5PC110:</div><div><p>== bilinear scaled SRC for comparison ==</p><p>transl: op=1, src=20028888, mask=- dst=20028888, speed=72.87 MPix/s</p> <p>== bilinear scaled OVER ==<br></p> <p>(before) transl: op=3, src=20028888, mask=- dst=20028888, speed=6.73 MPix/s<br>(after) transl: op=3, src=20028888, mask=- dst=20028888, speed=35.47 MPix/s<br><br>== bilinear scaled ADD ==<br></p> <p>(before) transl: op=12, src=20028888, mask=- dst=20028888, speed=6.91 MPix/s<br>(after) transl: op=12, src=20028888, mask=- dst=20028888, speed=43.18 MPix/s<br><br>== bilinear scaled mask SRC ==<br></p> <p>(before) transl: op=1, src=20028888, mask=8 dst=20028888, speed=5.74 MPix/s<br>(after) transl: op=1, src=20028888, mask=8 dst=20028888, speed=52.01 MPix/s<br><br>== bilinear scaled mask OVER ==<br></p> <p>(before) transl: op=3, src=20028888, mask=8 dst=20028888, speed=6.33 MPix/s<br>(after) transl: op=3, src=20028888, mask=8 dst=20028888, speed=30.00 MPix/s<br><br>== bilinear scaled mask ADD ==<br></p> <p>(before) transl: op=12, src=20028888, mask=8 dst=20028888, speed=6.41 MPix/s<br>(after) transl: op=12, src=20028888, mask=8 dst=20028888, speed=33.47 MPix/s</p>-- <br>Best Regards,<div>Taekyun Kim</div><br> </div>