1 |
Hey list, |
2 |
|
3 |
I'm observing on this olde Centrino laptop that emerges take much, much longer |
4 |
for certain packages than they did in the past. There was a bigger update on |
5 |
25. Jan (the first since 20. Oct) and another one on 22. Feb. Examples |
6 |
can be found at the bottom, they show increases from 15 to more than 100 |
7 |
percent. It's not with every package and there's always some fluctuation |
8 |
of course as packet versions get bigger. But I really noticed it when I |
9 |
was rebuilding gcc. Usually on bigger updates, I let it sit in the |
10 |
corner for the day, not using it, since it's a puny single-core. |
11 |
|
12 |
I'm using the same CXXFLAGS on my netbook, where gcc also took a step up |
13 |
from 4.6 to 4.7 (increas by 200%), so it's inherent in the version |
14 |
upgrade. OTOH, the first gcc-4.7 installation in the Centrino took only |
15 |
half as long as any of the following. |
16 |
(On a sidenote, a 64 bit crossdev gcc on the netbook -- same version and |
17 |
same useflags as world's -- took 51 minutes, whereas the system compiler |
18 |
needed 2h 40. What gives?) |
19 |
|
20 |
Can you give me an insight based on the little knowledge I can provide |
21 |
about the environment? My usual CXXFLAGS are "-O2 -march=native |
22 |
-fomit-frame-pointer -pipe". At some point in the past I added |
23 |
-fno-unwind-tables -fno-asynchronous-unwind-tables. My browser history |
24 |
tells me that I was researching that on 25. October 2013. And recently I |
25 |
experimented with -mpfmath=sse, but when gcc wouldn't build with that on |
26 |
the Netbook, I removed it again yesterday. |
27 |
|
28 |
I removed the extra CXXFLAGS yesterday and rebuilt gcc twice over last |
29 |
night, but the results aren't as hoped. |
30 |
|
31 |
|
32 |
Example compile times (output of qlop -g, with the installed versions |
33 |
added by me) |
34 |
|
35 |
gcc: Sat Apr 27 20:06:24 2013: 5338u seconds # 4.6.3 |
36 |
gcc: Sun Oct 20 21:05:53 2013: 7878u seconds # 4.7.3-r1 |
37 |
gcc: Sun Jan 26 06:18:34 2014: 7274u seconds # 4.7.3-r1 with USE=fortran until here (profile default), and probably introducing fno-...-tables |
38 |
gcc: Sat Feb 22 22:58:16 2014: 15814u seconds # 4.7.3-r1 did some stuff while compiling -> not pure compile time. But from here USE=-fortran |
39 |
gcc: Sun Feb 23 05:05:02 2014: 12952u seconds # 4.7.3-r1 removed -fno-unwind-tables -fno-asynchronous-unwind-tables -mpfmath=sse |
40 |
gcc: Sun Feb 23 09:50:35 2014: 13118u seconds # 4.7.3-r1 |
41 |
|
42 |
mc: Sun Apr 28 05:58:04 2013: 218u seconds # 4.8.7 |
43 |
mc: Sun Jan 26 08:37:50 2014: 238u seconds # 4.8.9 |
44 |
mc: Sun Feb 23 03:41:36 2014: 586u seconds # 4.8.11 |
45 |
|
46 |
php: Sun Jan 26 08:42:19 2014: 1994u seconds # 5.5.7 |
47 |
php: Sat Feb 1 16:37:25 2014: 2321u seconds # 5.5.7 |
48 |
php: Sun Feb 23 08:42:49 2014: 4037u seconds # 5.5.9 |
49 |
|
50 |
qtcore: Tue Jul 16 23:28:28 2013: 1424u seconds # 4.8.4-r5 |
51 |
qtcore: Mon Oct 21 00:23:09 2013: 1518u seconds # 4.8.5 |
52 |
qtcore: Sat Jan 25 23:34:12 2014: 2097u seconds # 4.8.5-r1 |
53 |
|
54 |
openssl: Sun Apr 28 00:54:07 2013: 378u seconds # 1.0.1c |
55 |
openssl: Mon Oct 21 13:21:48 2013: 404u seconds # 1.0.1e-r1 |
56 |
openssl: Sat Jan 25 21:41:01 2014: 546u seconds # 1.0.1f |
57 |
|
58 |
xorg-server: Sun Apr 28 05:37:58 2013: 430u seconds # 1.13.4 |
59 |
xorg-server: Mon Oct 21 15:24:02 2013: 501u seconds # 1.14.3 |
60 |
xorg-server: Sun Jan 26 16:48:22 2014: 544u seconds # 1.14.3-r2 |
61 |
|
62 |
gegl: Wed Jul 17 19:59:39 2013: 204u seconds # 0.1.6 |
63 |
gegl: Mon Oct 21 16:43:45 2013: 239u seconds # 0.2.0-r2 |
64 |
gegl: Sat Feb 1 17:31:07 2014: 273u seconds # 0.2.0-r2 |
65 |
|
66 |
git: Sun Apr 28 06:37:01 2013: 228u seconds # 1.8.1.5 |
67 |
git: Mon Oct 21 16:47:44 2013: 226u seconds # 1.8.1.5-r1 |
68 |
git: Sun Jan 26 00:36:17 2014: 314u seconds # 1.8.3.2-r1 |
69 |
|
70 |
kopete: Fri Sep 13 07:16:13 2013: 1495u seconds # 4.11.1 |
71 |
kopete: Mon Oct 21 23:29:32 2013: 1531u seconds # 4.11.2 |
72 |
kopete: Sat Feb 1 05:36:00 2014: 2025u seconds # 4.11.2-r1 |
73 |
|
74 |
kmail: Fri Sep 13 09:15:12 2013: 2015u seconds # 4.11.1 |
75 |
kmail: Tue Oct 22 02:20:00 2013: 2059u seconds # 4.11.2 |
76 |
kmail: Sat Feb 1 07:46:49 2014: 2489u seconds # 4.11.2-r1 |
77 |
|
78 |
portage: Wed Jul 17 01:26:02 2013: 55u seconds # 2.1.12 |
79 |
portage: Thu Sep 12 18:08:27 2013: 72u seconds # 2.2.1 |
80 |
portage: Sat Jan 25 19:12:30 2014: 126u seconds # 2.2.7 |
81 |
|
82 |
curl: Thu Sep 12 21:55:12 2013: 201u seconds # 7.31.0 |
83 |
curl: Sat Jan 25 22:26:28 2014: 259u seconds # 7.34.0-r1 |
84 |
curl: Sat Feb 22 20:17:06 2014: 608u seconds # 7.35.0 # Possible did some other stuff at the same time. |
85 |
|
86 |
qtsql: Mon Oct 21 05:12:27 2013: 124u seconds # 4.8.5 |
87 |
qtsql: Sun Jan 26 00:54:54 2014: 522u seconds # 4.8.5 |
88 |
|
89 |
qtdbus: Mon Oct 21 04:44:45 2013: 194u seconds # 4.8.5 |
90 |
qtdbus: Sun Jan 26 01:03:36 2014: 223u seconds # 4.8.5 |
91 |
|
92 |
lcms: Sat Jan 25 21:50:07 2014: 85u seconds # 2.5 |
93 |
lcms: Sat Feb 1 11:55:35 2014: 81u seconds # 2.5 |
94 |
lcms: Sat Feb 22 21:15:26 2014: 196u seconds # 2.5-r1 |
95 |
|
96 |
json-c: Sat Sep 8 21:39:56 2012: 30u seconds # 0.9 |
97 |
json-c: Sat Jan 25 21:12:03 2014: 58u seconds # 0.11 |
98 |
json-c: Sat Feb 22 21:13:33 2014: 113u seconds # 0.11 |
99 |
|
100 |
-- |
101 |
Gruß | Greetings | Qapla’ |
102 |
Please do not share anything from, with or about me with any Facebook service. |
103 |
|
104 |
I am not half as intelligent as you are! |