| 1 |
On Wed, Nov 23, 2011 at 11:55 AM, Stefan G. Weichinger <lists@×××××.at> wrote: |
| 2 |
> |
| 3 |
> Ad CFLAGS for i7-2600: |
| 4 |
> |
| 5 |
> Is that too much ricer-style? --> |
| 6 |
> |
| 7 |
> ### gcc -march=native -E -v - </dev/null 2>&1 | sed -n 's/.* -v - //p' |
| 8 |
> |
| 9 |
> CFLAGS="-O2 -D_FORTIFY_SOURCE=2 -march=core2 -mcx16 -msahf -maes |
| 10 |
> -mpclmul -mpopcnt -mavx --param l1-cache-size=32 --param |
| 11 |
> l1-cache-line-size=64 --param l2-cache-size=8192 -mtune=generic" |
| 12 |
> |
| 13 |
> gcc-4.5.3-r1 ... |
| 14 |
|
| 15 |
That's equivalent to, what, "-O2 -D_FORTIFY_SOURCE=2 -march=native" ? |
| 16 |
I don't see anything to complain about. |
| 17 |
|
| 18 |
Actually, it's pretty interesting seeing what that processor comes |
| 19 |
down to for -march=native. I wish there were a database of processors |
| 20 |
and their decomposed compiler tuning flags for comparison. That would |
| 21 |
be *very* interesting, from the standpoint of proc shopping and |
| 22 |
looking at the evolution of CPUs. |
| 23 |
|
| 24 |
-- |
| 25 |
:wq |
Archives