|
|
|
|
|
|
by masklinn
5812 days ago
|
|
|
Maybe you should provide both graphs on a full scale (from 0 to 4.8) to show just how little, in your benchmark, -fomit-frame-pointer brings to the table in 32b (under half a percent using your mean cycles count) versus how much is lost due to it in 64b (nearly +30% cycles) |
|
|