Edit page memo/benchmark/compression

[wiki.git] / memo / benchmark / compression

diff --git a/memo/benchmark/compression b/memo/benchmark/compression
index 2b633315dacc39400b12149ad1c811a230ceecf4..c73702c24aa07ccc2f48b53bc55f143f238cdead 100644 (file)
--- a/memo/benchmark/compression
+++ b/memo/benchmark/compression
@@ -1,14 +1,18 @@
-== Crude CPU benchmarks ==
+== A crude CPU benchmark using lossless compression algorithms ==
 
-=== Lossless compression algorithms ===
+=== How to test ===
 
 {{{
-time lzop  <file >/dev/null
-time gzip  <file >/dev/null
-time bzip2 <file >/dev/null
-time xz    <file >/dev/null
+time lzop  &gt;/dev/null &lt;testfile
+time gzip  &gt;/dev/null &lt;testfile
+time bzip2 &gt;/dev/null &lt;testfile
+time xz    &gt;/dev/null &lt;testfile
 }}}
 
+"testfile" is a 175MB avi (southpark_12x12_about_last_night.avi). It's not really compressible, which helps level the ground.
+
+=== Results ===
+
 || **Hostname** || **System**     || ** CPU Brand**              || **CPU** ||**Clock**||**Arch**||
 || giskard-x201 || Lenovo x201    || Intel(R) Core(TM) i7 CPU    ||   M 620 || 2.67GHz || x86_64 ||
 || istek        || Lenovo x201    || Intel(R) Core(TM) i5 CPU    ||   M 540 || 2.53GHz || x86_64 ||
@@ -33,6 +37,7 @@ time xz    <file >/dev/null
 ==== Time (seconds) ====
 
 || **hostname** ||**lzop**||**gzip**||**bzip2**||**lzma** || **xz**  ||**index**||
+|| gdesk        ||        ||        ||         ||         ||  61.46  ||        ||
 || giskard-x201 ||  2.960 ||  6.952 ||  37.157 ||  97.508 ||  95.945 ||  2.213 ||
 || istek        ||        ||  7.550 ||  35.480 ||         || 103.210 ||        ||
 || r2d2         ||  3.896 ||  9.029 ||  37.652 ||  84.727 ||  84.554 ||        ||
@@ -58,10 +63,8 @@ time xz    <file >/dev/null
 || giskard      ||        ||        ||         ||         ||   4.442 ||        ||
 
 ==== Notes ====
- * test file is a 175MB avi (southpark_12x12_about_last_night.avi)
  * all targets using 1 core only!
  * The MB/s figure refers to trinity
  * "index" is xz compression performance relative to trinity
  * bender and trinity run same lzma version
- * is xz faster or slower than lzma?
-
+ * is xz faster or slower than lzma?
\ No newline at end of file