1 |
On Mon, Apr 28, 2014 at 5:46 PM, Andrew Savchenko <bircoph@×××××.com> wrote: |
2 |
> Hello, |
3 |
> |
4 |
> On Sun, 27 Apr 2014 07:23:11 -0400 Rich Freeman wrote: |
5 |
>> And yet, in the same paragraph you mention -O3, which is |
6 |
>> tantamount to just setting a flag and walking away. That turns |
7 |
>> on 14 things you probably don't really need. |
8 |
> |
9 |
> Why 14 things? ... |
10 |
> |
11 |
> From my experience only three of them are harmful: |
12 |
... |
13 |
> All other -O3 option have either no effect or measurable |
14 |
> performance enhancements in the range of several percent. |
15 |
|
16 |
You missed my point. I think running batch optimizations like -O2/3 |
17 |
only makes sense. The argument was that -flto doesn't always help, |
18 |
and thus shouldn't always be used. My point was that convenience |
19 |
options like -O2/3 were used because while the options don't always |
20 |
help, they usually do, and nobody wants to bother with micromanaging |
21 |
them. |
22 |
|
23 |
Personally I use -O2 or -Os with a few additional options that are |
24 |
less space-expensive than full -O3, on the premise that cache and |
25 |
memory conservation probably buys you more than avoiding some jumps. |
26 |
But, short of profiling every package any selection is going to be a |
27 |
suboptimal choice based on averages. |
28 |
|
29 |
I wasn't trying to say that there was something wrong with -O3/2/etc. |
30 |
|
31 |
Rich |