1 |
Simon Stelling wrote: |
2 |
> Luca Barbato wrote: |
3 |
>> Alternatives: |
4 |
>> |
5 |
>> - as PPC we provide a default cflags & use tuned per certain cpu |
6 |
>> families using profiles, amd64 could provide a nocona profile that bans |
7 |
>> 3dnow* useflags. |
8 |
> |
9 |
> Not really. There are athlon64s and opterons with and without sse3 support. The |
10 |
> users who have an sse3-enabled CPU mostly stick -msse3 into their CFLAGS, as |
11 |
> they expect the flag to do what it says. The problem is even worse for x86: |
12 |
> You'd have to provide own profiles for: |
13 |
> |
14 |
> * i386 |
15 |
|
16 |
there is already isn't it? |
17 |
|
18 |
> * pentium-mmx,k6 |
19 |
|
20 |
people know what they are |
21 |
|
22 |
> * athlon xp,pentium3 |
23 |
|
24 |
ditto |
25 |
|
26 |
> * pentium4,athlon64/opteron (starting from 'Paris' core),celeron (starting from |
27 |
> 'willamette' core) |
28 |
|
29 |
those are relatively new |
30 |
|
31 |
> * celeron D/M, core solo/duo,core 2 duo, core 2 extreme,athlon64 (starting from |
32 |
> venice stepping E3 or san diego stepping E4), athlon64 X2, athlon64 FX (starting |
33 |
> from san diego stepping E4), sempron (starting from palermo stepping E3), |
34 |
|
35 |
Don't complain with me about marketing |
36 |
|
37 |
> pentium4 (starting from 'prescott') |
38 |
|
39 |
and so on... |
40 |
|
41 |
> etc... |
42 |
> |
43 |
> and now you want to let your pentium4(paris)-server, which is running 24/7, |
44 |
> compile a binpkg for your pentium4(prescott), using SSE3 |
45 |
> |
46 |
> have fun 8-) |
47 |
|
48 |
I take the base profile and turn sse3 useflag on, tune the cflags |
49 |
properly and then I issue the emerge -B foo as usual |
50 |
|
51 |
specific profiles are good just for disoriented users that need |
52 |
something working quickly, people knowing their system and what they |
53 |
want would be much happy with the useflags. |
54 |
|
55 |
Having better descriptions for flags could be of help too. |
56 |
|
57 |
Keep in mind that discover what your cpu or another cpu could take as |
58 |
cflags or simdflags requires the same (look at the arch faq about it, |
59 |
parse the cpuinfo, google a bit) |
60 |
|
61 |
> I'm not sure whether I understand this correctly. If we use flameeyes logic |
62 |
> anyway, why keeping the use flags? |
63 |
> |
64 |
|
65 |
Because you may not want all the flags the gcc guess would set for a |
66 |
reason or another (and there are many, including having gcc-X.ugly doing |
67 |
the worst with -msimd_du_jour but having some nicely tuned routines for |
68 |
that simd triggered by --enable-simd_du_jour) |
69 |
|
70 |
lu |
71 |
|
72 |
-- |
73 |
|
74 |
Luca Barbato |
75 |
|
76 |
Gentoo/linux Gentoo/PPC |
77 |
http://dev.gentoo.org/~lu_zero |
78 |
|
79 |
-- |
80 |
gentoo-dev@g.o mailing list |