mirror of
https://codeberg.org/dnkl/foot.git
synced 2026-02-04 04:06:06 -05:00
doc: foot.ini: document tweak.min-stride-alignment
This commit is contained in:
parent
bd994eda1c
commit
e43ea3676f
1 changed files with 26 additions and 0 deletions
|
|
@ -2031,6 +2031,32 @@ any of these options.
|
|||
|
||||
Default: _512_. Maximum allowed: _2048_ (2GB).
|
||||
|
||||
*min-stride-alignment*
|
||||
This option controls the minimum stride alignment, in bytes, when
|
||||
allocating SHM buffers.
|
||||
|
||||
In some circumstances, a compositor can import foot's SHM buffers
|
||||
directly to the GPU, without copying the buffer to GPU memory
|
||||
(typically on integrated graphics). Different drivers have
|
||||
different requirements for this, and one of those requirements is
|
||||
typically the stride alignment. At the time of writing, AMD GPUs
|
||||
require 256-byte alignment.
|
||||
|
||||
Note that doing a direct import typically disables immediate
|
||||
buffer release (if the compositor supports that), which means foot
|
||||
has to double buffer. This adds a performance penalty in foot, but
|
||||
the overall system performance should still be better.
|
||||
|
||||
If you are not using integrated graphics, or if the compositor
|
||||
does not support GPU direct imports, this option has close to zero
|
||||
impact. You can save a small amount of memory by setting this to
|
||||
0.
|
||||
|
||||
Ultimately, it is up to the compositor to decide whether to do
|
||||
immediate buffer releases, or try to optimize GPU imports.
|
||||
|
||||
Default: _256_
|
||||
|
||||
*sixel*
|
||||
Boolean. When enabled, foot will process sixel images. Default:
|
||||
_yes_
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue