1 |
On Fri, Aug 17, 2012 at 2:53 PM, Florian Philipp <lists@×××××××××××.net> wrote: |
2 |
> Am 17.08.2012 10:58, schrieb Jorge Almeida: |
3 |
>> |
4 |
>> 1) Is this strategy right? If so, any other flags to add? (or any |
5 |
>> flags to remove from the list?) |
6 |
>> |
7 |
>> 2) The --param flags are the ones of the computer that will do the |
8 |
>> compiling. I'm guessing the produced binaries are compatible with cpu |
9 |
>> with different --param flags. Is this right? |
10 |
>> |
11 |
>> TIA |
12 |
>> |
13 |
>> Jorge Almeida |
14 |
>> |
15 |
> |
16 |
> 1) Yes. But as you can see, -march=prescott is basically a subset of |
17 |
> atom. In fact, before there was a -march=atom option, prescott was the |
18 |
> best flag for atoms. I think you can avoid some hassle by simply |
19 |
> enabling "-march=prescott --param l1-cache-line-size=64 --param |
20 |
> l2-cache-size=512". |
21 |
> |
22 |
> 2) Yes, the param flags do not affect compatibility. Using the lower |
23 |
> value will probably be better but this is just an educated guess. |
24 |
> |
25 |
|
26 |
What about: |
27 |
|
28 |
CFLAGS="-O2 -pipe -fomit-frame-pointer -march=prescott -mtune=atom --param |
29 |
l1-cache-size=16 --param l1-cache-line-size=64 --param l2-cache-size=1024" |
30 |
|
31 |
If prescott were exactly a subset of atom, this would yield the best of both |
32 |
worlds. Can it still be safe? |
33 |
|
34 |
I read in http://forums.gentoo.org/viewtopic-t-895104.html : |
35 |
atom |
36 |
Intel Atom CPU with 64-bit extensions, MMX, SSE, SSE2, SSE3 and SSSE3 |
37 |
instruction set support. |
38 |
|
39 |
Does this mean that these flags are pulled by -mtune=atom, or do we need to |
40 |
ask for them explicitly? The WiKi shows how to find which flags are pulled by |
41 |
-march=native, but not the other cases. |
42 |
|
43 |
Thanks |
44 |
|
45 |
Jorge Almeida |