1 |
On 9/14/06, Duncan <1i5t5.duncan@×××.net> wrote: |
2 |
> "Mark Knecht" <markknecht@×××××.com> posted |
3 |
> 5bdc1c8b0609140715t2c135c9q50bf7ff8cbf1d82e@××××××××××.com, excerpted |
4 |
> below, on Thu, 14 Sep 2006 07:15:42 -0700: |
5 |
> |
6 |
> > I'm just curious whether anyone besides me is noticing their machine |
7 |
> > feeling somewhat sluggish since doing the gcc-4.1 upgrade? Mine seems ot |
8 |
> > be using a lot of memory. Alt-tabbing between windows seems slow. |
9 |
> > Ethernet traffic in my browser is causing pretty noticeable |
10 |
> > interruptions in things like MythTV. |
11 |
> |
12 |
> > The machine is still quite usable, but it doesn't feel as snappy as it |
13 |
> > did last week. |
14 |
> > |
15 |
> > I made no changes in /etc/make.conf for the upgrade. Everything is |
16 |
> > pretty basic as far as I can tell: |
17 |
> > |
18 |
> > CFLAGS="-march=k8 -O2 -pipe" |
19 |
> |
20 |
> > CXXFLAGS="${CFLAGS}" |
21 |
> |
22 |
> I've noticed rather the opposite, here. gcc-4.1.1 compiled binaries are |
23 |
> /dramatically/ faster and more efficient than 3.x. However, I'm using a |
24 |
> rather more elaborate CFLAGS/CXXFLAGS, and it's my conviction that gcc-4.1 |
25 |
> does better at optimizing exactly the way you've told it to. That is, if |
26 |
> you've given it inefficient optimizations, I'm convinced it makes a bad |
27 |
> thing worse, while if you've chosen your optimizations well, it makes a |
28 |
> good thing dramatically better. |
29 |
> |
30 |
> Here's my CFLAGS/CXXFLAGS: |
31 |
> |
32 |
> CFLAGS="-march=k8 -Os -pipe -frename-registers -fweb -freorder-blocks |
33 |
> -freorder-blocks-and-partition -combine -funit-at-a-time -ftree-pre |
34 |
> -fgcse-sm -fgcse-las -fgcse-after-reload -fmerge-all-constants" |
35 |
> |
36 |
> CXXFLAGS="-march=k8 -Os -pipe -frename-registers -fweb -freorder-blocks |
37 |
> -funit-at-a-time -ftree-pre -fgcse-sm -fgcse-las -fgcse-after-reload |
38 |
> -fmerge-all-constants" |
39 |
> |
40 |
<SNIP> |
41 |
> |
42 |
> As I said, with the above, there's a /dramatic/ improvement in |
43 |
> performance between gcc-3.x and gcc-4.1.x. |
44 |
> |
45 |
> -- |
46 |
> Duncan - List replies preferred. No HTML msgs. |
47 |
|
48 |
Hi Duncan, |
49 |
As always, very deep thanks for the answer. Very informative and interesting. |
50 |
|
51 |
Now, you are very adept at this. You're explanations make sense to |
52 |
the level I've considered them. (Not very far right now...) Main |
53 |
questions: |
54 |
|
55 |
1) What can be done to test this out at my end without making a 2-day |
56 |
commitment to rebuild the complete machine. Is it possibly to rebuild |
57 |
only portions of the machine using a different set of flags or is it a |
58 |
system wide commitment requiring that I rebuild 575 packages as I did |
59 |
last weekend? |
60 |
|
61 |
2) What about building the kernel? How do the standard |
62 |
|
63 |
make && make modules_install |
64 |
|
65 |
command make any use of the flags in /etc/make.conf? |
66 |
|
67 |
This machine is a fairly standard desktop running Xorg-7, Gnome and |
68 |
just a few apps most of the time. However I am an audio oriented |
69 |
person so my kernel is rt-sources from the proaudio overlay. (Ingo |
70 |
Molnar's patches to the kernel.org kernels and not a Gentoo kernel.) I |
71 |
need to ensure that the audio stuff (Jack, Ardour, Aqualung, 1394 hard |
72 |
drives) continue to work well. |
73 |
|
74 |
Your ideas are most welcome. |
75 |
|
76 |
Thanks, |
77 |
Mark |
78 |
-- |
79 |
gentoo-amd64@g.o mailing list |