https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68303

--- Comment #6 from Jonathan Wakely <redi at gcc dot gnu.org> ---
But it doesn't make a huge difference and iteration is still much faster for <=
17 elements:

| GCC 9 |       | GCC 10 |      | 10 wrapped |
0=0.696880      0=0.700247      0=0.110166
1=0.705601      1=0.676875      1=0.177209
2=1.006746      2=0.728425      2=0.228726
3=0.920503      3=0.778136      3=0.379848
4=1.044559      4=0.830860      4=0.378648
5=1.205183      5=0.882847      5=0.330908
6=0.950419      6=0.930731      6=0.363111
7=1.004634      7=1.001264      7=0.359892
8=1.070675      8=1.030736      8=0.368559
9=1.117331      9=1.100524      9=0.408855
10=1.192385     10=1.138487     10=0.424364
11=1.409491     11=1.181265     11=0.481035
12=0.958946     12=1.231175     12=0.517279
13=0.987015     13=1.328534     13=0.579356
14=1.024184     14=0.970562     14=0.629750
15=1.049774     15=0.991556     15=0.703401
16=1.082365     16=1.017649     16=0.754023
17=1.098181     17=1.037775     17=0.878584
18=1.127022     18=1.060267     18=1.137607
19=1.233192     19=1.085448     19=1.193701
20=1.215447     20=1.104254     20=1.001146
21=1.257106     21=1.128671     21=1.169435
22=1.278363     22=1.149634     22=1.185851
23=1.320331     23=1.171409     23=1.206449
24=0.988053     24=1.194045     24=1.229551
25=1.002762     25=1.218135     25=1.253025
26=1.006531     26=1.239955     26=1.276393
27=1.013021     27=1.263612     27=1.296041
28=1.027757     28=1.286184     28=1.360461
29=1.035280     29=1.308196     29=1.355412
30=1.055340     30=0.996437     30=1.040220

Reply via email to