1 |
Hey there |
2 |
|
3 |
As I mentioned in an earlier thread, I switched from 32 to 64 bit after some |
4 |
convinction work done by the ML and a friend. In order to justify the switch |
5 |
for myself, I made some performance comparisons. |
6 |
|
7 |
So, in case anyone is interested, here are my results. |
8 |
|
9 |
The only thing I don't really like is of course the increased RAM usage. |
10 |
While the old installation took 400 MB of RAM after Login to KDE (Akonadi is |
11 |
a hog), it now takes 500. The memory meter now stands always at least at 50% |
12 |
(3 GB available). I will have to tune down multitasking a bit. |
13 |
|
14 |
|
15 |
|
16 |
The following items first display the command excuted (denoted by $), and then |
17 |
the output of time for the command; first for 32 bit and then 64 bit. |
18 |
|
19 |
All tests were done on my Core 2 Duo laptop (T7200, max. 2GHz) fixed at 1 GHz |
20 |
and with 3 GB of RAM. This is not a theoretical benchmark, but rather about |
21 |
stuff I usually to do in my every-day computing. I excluded compiling, |
22 |
because it involves more than just crunching. |
23 |
|
24 |
Resulting observation: there seems to be an inherent increase of about 10% |
25 |
in memory throughput. I was most surprised by the performance of lilypond |
26 |
and blender, two computing-intensive applications I tend to use regularly. |
27 |
I wanted to do a framerate comparison of the Java-based CPU hog Minecraft, |
28 |
but didn't get around to it. |
29 |
|
30 |
|
31 |
All the following tasks were done in ramdisk to rule out HDD hindrance. |
32 |
|
33 |
|
34 |
|
35 |
$ 7z b (7zip's own benchmark function, abridged output) |
36 |
|
37 |
32 bit | 64 bit |
38 |
=========================================================================== |
39 |
RAM size: 3037 MB | RAM size: 3013 MB |
40 |
RAM usage: 425 MB | RAM usage: 425 MB |
41 |
| |
42 |
Dict Compressing | Decomp. | Dict Compressing | Decomp. |
43 |
Speed Rating | Speed Rating | Speed Rating | Speed Rating |
44 |
KB/s MIPS | KB/s MIPS | KB/s MIPS | KB/s MIPS |
45 |
| |
46 |
22: 1487 1446 | 19039 1719 | 22: 1612 1568 | 20974 1893 |
47 |
23: 1443 1470 | 19049 1744 | 23: 1612 1642 | 20758 1900 |
48 |
24: 1499 1612 | 18854 1749 | 24: 1591 1711 | 20292 1883 |
49 |
25: 1489 1700 | 18611 1750 | 25: 1584 1809 | 20030 1884 |
50 |
-------------------------------------------------------------------------- |
51 |
Avr: 1557 1740 | 1682 1890 |
52 |
Tot: 1649 | 1786 |
53 |
|
54 |
|
55 |
|
56 |
Various compressions of the High Voltage SID collection version 56 |
57 |
(41356 files, 1416 folders, total dir size 307.676k according to du -s). |
58 |
|
59 |
Extract: |
60 |
$ unrar x hvsc.rar |
61 |
real 0m38.582s 0m38.763s |
62 |
user 0m36.031s 0m36.190s |
63 |
sys 0m2.523s 0m2.496s |
64 |
---> neglibible |
65 |
|
66 |
Repack witz p7zip, resulting archive size 54.8 MB: |
67 |
$ 7za a -t7z -m0=lzma -mx=9 -mfb=64 -md=32m -ms=on hvsc.7z C64Music/ > /dev/null |
68 |
real 3m0.530s 2m41.780s |
69 |
user 5m22.359s 4m55.810s |
70 |
sys 2m2.973s 0m3.144s |
71 |
---> 1/9 faster |
72 |
|
73 |
Extract from 7z: |
74 |
$ 7z x hvsc.7z |
75 |
real 0m24.541s 0m21.437s |
76 |
user 0m19.302s 0m16.929s |
77 |
sys 0m4.403s 0m4.472s |
78 |
---> 1/10 faster |
79 |
|
80 |
Simple taring of the directory: |
81 |
$ tar cf hvsc.tar C64Music/ |
82 |
real 0m1.334s 0m1.226s |
83 |
user 0m0.297s 0m0.304s |
84 |
sys 0m1.020s 0m0.872s |
85 |
---> ~1/10 faster |
86 |
|
87 |
XZing the tar, resulting archive size 54.2 MB: |
88 |
$ xz -k -z hvsc.tar |
89 |
real 6m26.383s 4m31.747s |
90 |
user 6m23.375s 4m30.969s |
91 |
sys 0m2.733s 0m0.728s |
92 |
---> ~1/3 faster |
93 |
|
94 |
XZing with --extreme option (about 4% smaller archive): |
95 |
$ xz -e -k -z hvsc.tar |
96 |
real 15m37.732s 10m39.348s |
97 |
user 15m36.592s 10m38.900s |
98 |
sys 0m0.977s 0m0.456s |
99 |
---> ~1/3 faster |
100 |
|
101 |
Packing in squashfs: |
102 |
$ mksquashfs C64Music/ hvsc.sqfs |
103 |
real 0m57.380s 0m44.697s |
104 |
user 1m45.136s 1m20.377s |
105 |
sys 0m9.059s 0m6.116s |
106 |
---> ~1/4 faster |
107 |
|
108 |
|
109 |
|
110 |
Some memory shuffling: |
111 |
$ dd if=/dev/urandom of=random bs=1M count=500 |
112 |
real 2m0.306s 1m49.348s |
113 |
user 0m0.003s 0m0.000s |
114 |
sys 2m0.292s 1m49.315s |
115 |
---> 1/12 faster |
116 |
|
117 |
$ cp random r2 |
118 |
real 0m1.069s 0m0.917s |
119 |
user 0m0.000s 0m0.004s |
120 |
sys 0m1.067s 0m0.908s |
121 |
---> 1/10 faster |
122 |
|
123 |
|
124 |
|
125 |
Compile Bach's Christmas Oratorio, Tenor part, 16 pages A5: |
126 |
$ lilypond wo.ly |
127 |
real 0m31.430s 0m23.737s |
128 |
user 0m30.711s 0m23.129s |
129 |
sys 0m0.717s 0m0.592s |
130 |
---> 1/3 faster |
131 |
|
132 |
Compile Oratorio de Noël by Saint-Saëns, 4 voices, 16 pages A4: |
133 |
$ lilypond noel.lyk |
134 |
real 0m41.575s 0m26.494s |
135 |
user 0m41.177s 0m25.870s |
136 |
sys 0m0.390s 0m0.604s |
137 |
---> >1/3 faster |
138 |
|
139 |
|
140 |
|
141 |
Optimising a PNG (photo of Orion nebula, 1400x1050 pixel): |
142 |
$ optipng -o9 Orion.png |
143 |
real 0m23.491s 0m21.337s |
144 |
user 0m23.465s 0m21.281s |
145 |
sys 0m0.027s 0m0.008s |
146 |
---> 1/10 faster |
147 |
|
148 |
|
149 |
Encoding a video file to x264 (1280x960, 1600 frames, no sound): |
150 |
First pass: |
151 |
$ mencoder bike.flv -ovc x264 -x264encopts bitrate=2000:pass=1 -nosound -o /dev/null |
152 |
real 1m57.379s 1m44.500s |
153 |
user 3m48.048s 3m19.728s |
154 |
sys 0m0.837s 0m0.796s |
155 |
---> 1/8 faster |
156 |
|
157 |
Second pass: |
158 |
$ mencoder bike.flv -ovc x264 -x264encopts bitrate=2000:pass=2 -nosound -o bike.avi |
159 |
real 4m53.233s 4m15.533s |
160 |
user 9m41.169s 8m26.660s |
161 |
sys 0m0.990s 0m1.144s |
162 |
---> 1/8 faster |
163 |
|
164 |
|
165 |
Rendering a small test scene in Blender 2.63 (http://www.eofw.org/bench/test.blend): |
166 |
(file's scene defaults: 800x600, one thread) |
167 |
win32 2:59 |
168 |
Gentoo32 2:57 |
169 |
Gentoo64 2:09 |
170 |
---> > 1/3 faster |
171 |
-- |
172 |
Gruß | Greetings | Qapla' |
173 |
Please do not share anything from, with or about me with any Facebook service. |
174 |
|
175 |
“Microsoft isn't evil, they just make really crappy operating systems.” |
176 |
– Linus Torvalds |