{"id":307,"date":"2012-10-04T17:58:56","date_gmt":"2012-10-04T22:58:56","guid":{"rendered":"http:\/\/www.compdigitec.com\/labs\/?p=307"},"modified":"2012-10-04T18:02:04","modified_gmt":"2012-10-04T23:02:04","slug":"libavcodec-native-neon-compile-vs-hybrid-mode","status":"publish","type":"post","link":"http:\/\/www.compdigitec.com\/labs\/2012\/10\/04\/libavcodec-native-neon-compile-vs-hybrid-mode\/","title":{"rendered":"Libavcodec &#8211; native NEON compile vs hybrid mode"},"content":{"rendered":"<p><strong>Background<\/strong><\/p>\n<p>Benchmarks using the latest VLC beta and benching whether a native build solely for NEON is faster or a hybrid build. The expectation is that the native build should be very slightly faster than the hybrid build. This is done on a TI OMAP 4460 (Galaxy Nexus). Method is <a href=\"http:\/\/www.compdigitec.com\/labs\/2012\/02\/01\/android-codec-benchmarks\/\">same as previous benchmarks<\/a>.<\/p>\n<p><strong>Results<\/strong><\/p>\n<p>Ironically, it appears that the hybrid build is slightly faster(?) than the native NEON build, although in practice there is no real difference for the user. So the proper result should be <strong>inconclusive<\/strong>.<\/p>\n<table>\n<tbody>\n<tr>\n<td>Hybrid mean<\/td>\n<td><strong>80.25%<\/strong><\/td>\n<td>Native NEON mean<\/td>\n<td>81.56%<\/td>\n<\/tr>\n<tr>\n<td>Hybrid median<\/td>\n<td><strong>82.10%<\/strong><\/td>\n<td>Native NEON median<\/td>\n<td>82.80%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p><a href=\"http:\/\/www.compdigitec.com\/labs\/wp-content\/uploads\/2012\/10\/libavcodec-neon-hybrid-bench.png\"><img decoding=\"async\" loading=\"lazy\" class=\"size-full wp-image-308\" title=\"Libavcodec Native NEON mode vs hybrid mode\" src=\"http:\/\/www.compdigitec.com\/labs\/wp-content\/uploads\/2012\/10\/libavcodec-neon-hybrid-bench.png\" alt=\"Libavcodec Native NEON mode vs hybrid mode\" width=\"560\" height=\"375\" \/><\/a><\/p>\n<p><strong>Raw data<\/strong><\/p>\n<pre class=\"brush: plain; title: ; notranslate\" title=\"\">\r\nHybrid build:\r\n19453   126 10079    R     575m 82.6   1 72.0 org.videolan.vlc\r\n19453   126 10079    S     593m 85.3   1 64.4 org.videolan.vlc\r\n19453   126 10079    S     594m 85.4   1 85.0 org.videolan.vlc\r\n19453   126 10079    S     594m 85.5   0 83.1 org.videolan.vlc\r\n19453   126 10079    S     594m 85.5   0 81.8 org.videolan.vlc\r\n19453   126 10079    S     597m 85.9   0 77.9 org.videolan.vlc\r\n19453   126 10079    S     597m 85.9   0 85.6 org.videolan.vlc\r\n19453   126 10079    S     597m 85.9   0 85.5 org.videolan.vlc\r\n19453   126 10079    S     598m 85.9   0 81.6 org.videolan.vlc\r\n19453   126 10079    S     598m 86.0   0 82.1 org.videolan.vlc\r\n19453   126 10079    S     598m 86.0   1 83.8 org.videolan.vlc\r\n\r\nNative NEON build:\r\n19703   126 10079    S     590m 84.8   0 68.4 org.videolan.vlc\r\n19703   126 10079    S     589m 84.7   1 82.3 org.videolan.vlc\r\n19703   126 10079    S     592m 85.1   1 81.1 org.videolan.vlc\r\n19703   126 10079    S     592m 85.1   1 83.0 org.videolan.vlc\r\n19703   126 10079    S     593m 85.2   1 81.0 org.videolan.vlc\r\n19703   126 10079    S     593m 85.3   1 83.2 org.videolan.vlc\r\n19703   126 10079    S     593m 85.3   1 82.8 org.videolan.vlc\r\n19703   126 10079    S     594m 85.3   1 79.8 org.videolan.vlc\r\n19703   126 10079    S     594m 85.4   1 83.8 org.videolan.vlc\r\n19703   126 10079    S     594m 85.4   1 85.6 org.videolan.vlc\r\n19703   126 10079    S     594m 85.4   1 86.2 org.videolan.vlc\r\n<\/pre>","protected":false},"excerpt":{"rendered":"<p>Background Benchmarks using the latest VLC beta and benching whether a native build solely for NEON is faster or a hybrid build. The expectation is that the native build should be very slightly faster than the hybrid build. This is done on a TI OMAP 4460 (Galaxy Nexus). Method is same as previous benchmarks. Results [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[759],"tags":[811,760,813,812,814,776],"_links":{"self":[{"href":"http:\/\/www.compdigitec.com\/labs\/wp-json\/wp\/v2\/posts\/307"}],"collection":[{"href":"http:\/\/www.compdigitec.com\/labs\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.compdigitec.com\/labs\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.compdigitec.com\/labs\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.compdigitec.com\/labs\/wp-json\/wp\/v2\/comments?post=307"}],"version-history":[{"count":0,"href":"http:\/\/www.compdigitec.com\/labs\/wp-json\/wp\/v2\/posts\/307\/revisions"}],"wp:attachment":[{"href":"http:\/\/www.compdigitec.com\/labs\/wp-json\/wp\/v2\/media?parent=307"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.compdigitec.com\/labs\/wp-json\/wp\/v2\/categories?post=307"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.compdigitec.com\/labs\/wp-json\/wp\/v2\/tags?post=307"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}