1 |
· Florian Philipp <f.philipp@××××××.de>: |
2 |
|
3 |
> You see, they are not compatible and even if some code works I wouldn't bet |
4 |
> multimedia apps will perform well. |
5 |
> |
6 |
> With -mtune the instruction set stays the same. It is just "rearranged". |
7 |
|
8 |
Hm. Allright. When using just -mtune (ie. without -march), the |
9 |
docs at http://gcc.gnu.org/onlinedocs/gcc-4.1.2/gcc/i386-and-x86_002d64-Options.html |
10 |
say: |
11 |
|
12 |
| While picking a specific cpu-type will schedule things appropriately |
13 |
| for that particular chip, the compiler will not generate any code that |
14 |
| does not run on the i386 without the -march=cpu-type option being used. |
15 |
|
16 |
If -mtune=athlon-xp is used, code is generated which may make |
17 |
use of 3dNOW!. 3dNOW! is, of course, not to be found on 386 :) |
18 |
If the instruction set stays the same, code generated with |
19 |
-mtune=athlon-xp would not be executable on 386 machines, if |
20 |
I understand you correctly. |
21 |
|
22 |
Hm. With -mtune, the set of available instructions (ie. |
23 |
stuff like 3dNOW!, I suppose?) is NOT changed from the default |
24 |
of i386, is it? Or what does "Tune to cpu-type everything applicable |
25 |
about the generated code, except for the ABI and the set of available |
26 |
instructions." mean - especially note the "except for [...] the set of |
27 |
available instructions" part. |
28 |
|
29 |
So with "-mtune=pentium-m -march=athlon-xp" I'm making the compiler |
30 |
generate code which is "ordered" the way it's best for pentium-m |
31 |
machines while allowing it to use athlon-xp instruction set? Is |
32 |
that what I'm doing? |
33 |
|
34 |
If so, then it seems you're right - code will run, but maybe not |
35 |
so well. |
36 |
|
37 |
Is that understanding correct? If so, then I really should think |
38 |
twice about using "-mtune=pentium-m -march=athlon-xp", shouldn't |
39 |
I? |
40 |
|
41 |
Curious, |
42 |
|
43 |
Alexander Skwar |
44 |
-- |
45 |
No matter how many resources you have, it is never enough. |
46 |
-- Murphy's Computer Laws n°1 |
47 |
|
48 |
|
49 |
-- |
50 |
gentoo-user@g.o mailing list |