The original value of 256 was under the assumption that you're a batch buffer which is likely going to have a large number of relocations. However, pipeline objects on Gen7 will have at most 6 relocations (one per shader stage and one for the workaround BO) so this is a lot of per-pipeline wasted space. Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>tags/19.3-branchpoint
| @@ -118,7 +118,7 @@ anv_reloc_list_grow(struct anv_reloc_list *list, | |||
| if (list->num_relocs + num_additional_relocs <= list->array_length) | |||
| return VK_SUCCESS; | |||
| size_t new_length = MAX2(256, list->array_length * 2); | |||
| size_t new_length = MAX2(16, list->array_length * 2); | |||
| while (new_length < list->num_relocs + num_additional_relocs) | |||
| new_length *= 2; | |||