From 97ed31fd8ed7241dade2e579488146e8fa3473f7 Mon Sep 17 00:00:00 2001
From: wm4 <wm4@nowhere>
Date: Fri, 21 Dec 2012 18:13:16 +0100
Subject: audio: make de-planarization faster

Uses the same trick as the planarization code to turn per-sample memcpy
calls into mov instructions. Makes decoding a ~25min 48000Hz 2ch floatle
audio file faster from 3.8s to 2.7s.
---
 audio/reorder_ch.c | 37 +++++++++++++++++++++++++++++++++++--
 1 file changed, 35 insertions(+), 2 deletions(-)

(limited to 'audio/reorder_ch.c')

diff --git a/audio/reorder_ch.c b/audio/reorder_ch.c
index 330a807eae..72849c3d13 100644
--- a/audio/reorder_ch.c
+++ b/audio/reorder_ch.c
@@ -1399,7 +1399,8 @@ int main(int argc, char *argv[]) {
 #endif
 
 static inline void reorder_to_planar_(void *restrict out, const void *restrict in,
-        size_t size, size_t nchan, size_t nmemb) {
+                                      size_t size, size_t nchan, size_t nmemb)
+{
     size_t i, c;
     char *outptr = (char *) out;
     size_t instep = nchan * size;
@@ -1413,7 +1414,7 @@ static inline void reorder_to_planar_(void *restrict out, const void *restrict i
 }
 
 void reorder_to_planar(void *restrict out, const void *restrict in,
-        size_t size, size_t nchan, size_t nmemb)
+                       size_t size, size_t nchan, size_t nmemb)
 {
     // special case for mono (nothing to do...)
     if (nchan == 1)
@@ -1432,3 +1433,35 @@ void reorder_to_planar(void *restrict out, const void *restrict in,
     else
         reorder_to_planar_(out, in, size, nchan, nmemb);
 }
+
+static inline void reorder_to_packed_(uint8_t *out, uint8_t **in,
+                                      size_t size, size_t nchan, size_t nmemb)
+{
+    size_t outstep = nchan * size;
+
+    for (size_t c = 0; c < nchan; ++c) {
+        char *outptr = out + c * size;
+        char *inptr = in[c];
+        for (size_t i = 0; i < nmemb; ++i, outptr += outstep, inptr += size) {
+            memcpy(outptr, inptr, size);
+        }
+    }
+}
+
+// out = destination array of packed samples of given size, nmemb frames
+// in[channel] = source array of samples for the given channel
+void reorder_to_packed(uint8_t *out, uint8_t **in,
+                       size_t size, size_t nchan, size_t nmemb)
+{
+    if (nchan == 1)
+        memcpy(out, in, size * nchan * nmemb);
+    // See reorder_to_planar() why this is done this way
+    else if (size == 1)
+        reorder_to_packed_(out, in, 1, nchan, nmemb);
+    else if (size == 2)
+        reorder_to_packed_(out, in, 2, nchan, nmemb);
+    else if (size == 4)
+        reorder_to_packed_(out, in, 4, nchan, nmemb);
+    else
+        reorder_to_packed_(out, in, size, nchan, nmemb);
+}
-- 
cgit v1.2.3