1 |
On Fri, 2003-11-28 at 19:43, Steve Arnold wrote: |
2 |
> I think you've hit the nail on the head here, as the kernel seems to |
3 |
> be the one thing in Gentoo I haven't been happy with lately; I've |
4 |
> tried many, including making that Planet CCRMA sources ebuild in an |
5 |
> attempt to get something like the performance I get with rh9 and the |
6 |
> binary CCRMA kernel. I still haven't done it yet... |
7 |
|
8 |
Intriguing. Has it occured to you that maybe the reason you won't get |
9 |
equivalent performance is because the Redhat distro is suppoting Ulrich |
10 |
Drepper's very fast, scaleable POSIX threads implementation (NPTL) in |
11 |
the glibc shipped by Redhat (as well as the kernel of course)? I |
12 |
revamped by desktop system to support NPTL recently and I can tell you |
13 |
... it's fast. The thing is, I don't believe any Gentoo user has yet |
14 |
attempted to bootstrap a glibc against 2.4 sources (probably because the |
15 |
redhat/fedora kernel is the only available 2.4 sources that support it). |
16 |
Also, building sys-libs/glibc with USE="nptl" expects 2.6 sources in |
17 |
place. That is something I want to try soon; I think I mentioned this in |
18 |
a previous post. Basically, this benefits anything that links to |
19 |
pthread.so (many things). To give you but one example, I've heard of |
20 |
300% performance improvements in MySQL. |
21 |
|
22 |
Interestingly, Redhat have dropped the RMAP VM from the latest set of |
23 |
sources (certainly the Fedora 1 sources). This is somewhat |
24 |
dissapointing. I believe that Andrea's stock VM may appear faster under |
25 |
certain useage patterns, but Rik's RMAP will hold up better under |
26 |
serious stressloads (i.e. servers). That is why I hope I am capable of |
27 |
patching it in, as Marc did with the WOLK kernel. |
28 |
|
29 |
> My one recommendation now would be to drop the lm_sensor patches (or |
30 |
> don't include them) but make sure i2c is current. Maybe you could |
31 |
> even do a local use flag for stable/dev branches (I don't know what |
32 |
> the i2c state is right now, so maybe that's a moot point). |
33 |
> |
34 |
> I'm not sure about 2.4.22, but the earlier kernel tree included a |
35 |
> rather stale set of i2c modules (and no lm_sensors), whereas wolk |
36 |
> includes updated versions of both. The latter approach seems handy |
37 |
> from one perspective, however, it somewhat collides with the |
38 |
> lm_sensors package (since both the i2c and lm_sensors ebuilds |
39 |
> over-write any existing kernel modules). |
40 |
> |
41 |
> I'm not sure if that's very clear, but the lm_sensor ebuild includes |
42 |
|
43 |
Yes, I follow - the lm_sensors ebuild wants to build modules against the |
44 |
current kernel sources/headers. |
45 |
|
46 |
> both the user-space tools and modules; since a set of kernel |
47 |
> patches only includes the sensor modules, it's only partially |
48 |
> usable. |
49 |
|
50 |
> |
51 |
> Depending on the age/version of the kernel i2c code, I think |
52 |
> adding a set of i2c patches (maybe one for each branch, since both |
53 |
> branches are in gentoo already) is the best way to go. Most of the |
54 |
> kernel sources I've tried in the last few months update at least |
55 |
> i2c, if not both update i2c/add lm_sensors. |
56 |
|
57 |
Interesting. I might hit you up for some advice on how to actually use |
58 |
the things! I'm not very experienced with lm_sensors. But yes, I'll keep |
59 |
an eye open for any patches in accordance with your wishes. |
60 |
|
61 |
> |
62 |
> Well, that's my $.02 (other than offering encouragement). Of |
63 |
> course, I can't offer much in the way of real help until my |
64 |
> crap-load of work gets at least a little smaller. |
65 |
|
66 |
Being an experienced programmer you could perhaps help me out and show |
67 |
me how to use CVS _properly_ ;) |
68 |
|
69 |
--Kerin Francis Millar |