Create a path toward sharing memory from a numpy buffer #1061
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
hmaarrfk
force-pushed
the
from_buffer
branch
3 times, most recently
from
1c7415a to
615bfa5
Compare
|
@jlaine I didn't ask but are you interested in something like this? I've been using it with quite a bit of success more recently, I can add checks for continuous buffers too if you want. |
hmaarrfk
force-pushed
the
from_buffer
branch
2 times, most recently
from
6c6d398 to
3fbb392
Compare
|
anycomments on the desire to even support this kind of feature? |
hmaarrfk
force-pushed
the
from_buffer
branch
2 times, most recently
from
9b176c7 to
a78a054
Compare
hmaarrfk
force-pushed
the
from_buffer
branch
2 times, most recently
from
e4da4df to
172d8a8
Compare
This avoids the use of hstack, which inevitiably copies the data to a new memory locaiton. The speed up is small, but measurable. I can go from 185 fps decoding of a ~3000 x 2000 video to 200 fps decoding.
hmaarrfk
force-pushed
the
from_buffer
branch
from
172d8a8 to
635a6c5
Compare
|
I've rebased this. this is ready for review. |
|
For the future, I'm okay with having numpy as a compile time dep if that meant PyAV was faster or smaller. |
Maybe we can discuss this in an issue |
WyattBlue
added a commit
that referenced
this pull request
Feb 25, 2024
The YUV fastpath did not preserve the underlying byte-order, which causes issues for applications that rely on it.
WyattBlue
added a commit
that referenced
this pull request
Feb 25, 2024
The YUV fastpath did not preserve the underlying byte-order, which causes issues for applications that rely on it.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I know there is some concern about having the memory aligned.
I didn't know if it needed to be aligned to 16 bits, or 16 bytes.
If you know let me know which it is, I can add a check for it. If you don't want to support this, then that is fine too.
Closes: #1060