Nvimgcodec + tensor ipc #38
Merged
Conversation
This file contains hidden or 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
Make tensor IPC datapath optional/config-based Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Missed as part of rebase. This suggestion makes sense Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> Signed-off-by: Brandon Pelfrey <brandonpelfrey@gmail.com> Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Author
|
|
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
…atype, cleanup on scheduler finished_req_ids Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
Signed-off-by: Brandon Pelfrey <bpelfrey@nvidia.com>
soodoshll
force-pushed
the
nvimgcodec-ipc
branch
from
a4046db to
517e7e4
Compare
wangshangsam
approved these changes
Feb 9, 2026
wangshangsam
added a commit
that referenced
this pull request
Feb 10, 2026
This reverts commit bf71fe5.
wangshangsam
added a commit
that referenced
this pull request
Feb 10, 2026
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
New Requirements:
nvidia-nvjpegnvidia-nvimgcodec-cu13==0.7.0.11nvimgcodec is enabled by default. Note that this change does not affect the image loader in
dynamo.vllm, which has its own image processs logic. Will come up with a PR to dynamo later.Enable tensor ipc with
--multimodal-tensor-ipc=torch_shm**Note: ** images sent to the api server will be decoded even before being added to the waiting queue, so please throttle the max concurrency at the client side (now I use 384 * num_backend), which potentially requires modifying
task.py.cc @wangshangsam