| Time |
Nick |
Message |
| 03:30 |
|
VanessaE joined #minetest-dev |
| 04:28 |
|
celeron55 joined #minetest-dev |
| 05:19 |
|
SpeedProg joined #minetest-dev |
| 05:31 |
|
Calinou joined #minetest-dev |
| 10:09 |
|
EdB joined #minetest-dev |
| 15:08 |
|
PilzAdam joined #minetest-dev |
| 15:28 |
|
hmmmm joined #minetest-dev |
| 16:09 |
|
rubenwardy joined #minetest-dev |
| 16:11 |
|
rubenwardy1 joined #minetest-dev |
| 16:12 |
|
PilzAdam joined #minetest-dev |
| 16:43 |
|
rubenwardy joined #minetest-dev |
| 16:53 |
|
VanessaE joined #minetest-dev |
| 16:59 |
|
rubenwardy joined #minetest-dev |
| 18:20 |
|
pfaff joined #minetest-dev |
| 18:26 |
|
rubenwardy1 joined #minetest-dev |
| 18:52 |
hmmmm |
so i'm designing the new perlin functions, and i have some questions |
| 18:52 |
hmmmm |
they require intermediary buffers |
| 18:53 |
hmmmm |
would you like it for the caller to provide it, or should i have the perlin noise function allocate it via alloca() |
| 18:54 |
hmmmm |
since all the hardcore computation happens in the gradient function, using alloca shouldn't affect things much from a performance standpoint since SFO will still be usable in gradient |
| 18:54 |
hmmmm |
however, when requesting maps that are too large, it'll cause a stack overflow |
| 18:55 |
hmmmm |
having the caller provide the intermediate buffer is pretty ugly though |
| 18:55 |
hmmmm |
for the typical use case (16x16 areas) it wouldn't overflow at all |
| 18:55 |
hmmmm |
16x16x16 as well |
| 18:57 |
|
rubenwardy joined #minetest-dev |
| 19:11 |
|
rubenwardy left #minetest-dev |
| 21:13 |
hmmmm |
:/ |
| 21:14 |
hmmmm |
celeron55 |
