1 |
Simon Strandman posted <4405E164.90501@×××××.com>, excerpted below, on |
2 |
Wed, 01 Mar 2006 19:01:08 +0100: |
3 |
|
4 |
Robbert Young wrote... |
5 |
>> USE="x86 3dnow X acpi alsa apache2 audiofile avi bitmap-fonts bzip2 |
6 |
>> crypt cups dga eds emboss encode exif expat fam fbcon foomaticdb |
7 |
>> fortran gd gdbm gif glut gpm gstreamer gtk gtk2 imagemagick imlib ipv6 |
8 |
>> java jikes jpeg junit lcms libg++ libwww mad mailwrapper mhash mikmod |
9 |
>> mmx mng mozilla mp3 mpeg mysql ncurses nptl ogg oggvorbis opengl pam |
10 |
>> pcre pdflib png python quicktime readline samba sdl slang spell sse |
11 |
>> ssl tcpd tetex tiff truetype truetype-fonts type1-fonts udev |
12 |
>> userlocales vorbis xml xml2 xmms xv zlib userland_GNU kernel_linux |
13 |
>> elibc_glibc" |
14 |
>> |
15 |
> 1. Check you RAM with memtest86+. |
16 |
|
17 |
Excellent advice, but it won't catch all errors, particularly bus errors |
18 |
that happen only under stress. I know, as I have just such a problem |
19 |
here. After a BIOS came out that had memory speed tweaking, I reduced the |
20 |
speed from the rated 200 (ddr to 400=pc3200) to 183 (ddr to 366, pc3000) |
21 |
and it's now rock-stable, as Linux /should/ be. |
22 |
|
23 |
Of course, an insufficient power supply is the next candidate in the |
24 |
hardware failure realm. |
25 |
|
26 |
> 2. Drop your CFLAGS to "-march=opteron -O2 -pipe". Extreme CFLAGS wont |
27 |
> improve your performance anyway and I'm a bit suprised you're using such |
28 |
> CFLAGS on a server. |
29 |
|
30 |
Well, it /is/ glibc we are talking, and if you check the glibc ebuild, it |
31 |
strips virtually all flags /other/ than the above "-march=opteron -O2 |
32 |
-pipe". Thus, in this particular case, unless the ebuild itself has been |
33 |
hacked to kill that strip, I don't believe this is the problem. |
34 |
|
35 |
On most ebuilds it could be. You are absolutely right on that, as most |
36 |
don't stripflags to such a small subset. Also see below. |
37 |
|
38 |
> 3. You aren't supposed to use a x86 profile on a x86_64-pc-linux-gnu |
39 |
> system. Change to an amd64 profile instead. |
40 |
|
41 |
If the issue isn't purely hardware, I'd lay money on this. glibc in |
42 |
particular is not something one wants to be compiling toward the wrong |
43 |
arch. It could very easily be a section of 32-bit hand-tuned assembly, |
44 |
that the compiler is trying to unite with a 64-bit overall glibc. That |
45 |
*WILL* produce problems. |
46 |
|
47 |
In addition, flags like 3dnow and mmx (just tried to write xmms =8^) are |
48 |
masked on amd64, because they tend to invoke 32-bit only assembly code. |
49 |
amd64 has those by default, so no need for those USE flags. As I |
50 |
mentioned, however, it's not likely that such is the problem with glibc, |
51 |
given its stripflags. That's anyway on amd64. It's possible they still |
52 |
get thru on the bastardized amd64 system but x86 profile we see here. |
53 |
|
54 |
-- |
55 |
Duncan - List replies preferred. No HTML msgs. |
56 |
"Every nonfree program has a lord, a master -- |
57 |
and if you use the program, he is your master." Richard Stallman in |
58 |
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html |
59 |
|
60 |
|
61 |
-- |
62 |
gentoo-amd64@g.o mailing list |