Repository navigation
Conversation
|
This pull request has merge conflicts that must be resolved before it can be |
8ff3441 to
3326236
Compare
|
/ci run |
|
❌ This PR is 2 commits behind upstream |
|
/ci run --allow-stale |
|
✅ Triggered Buildkite CI #94658 for commit
|
janeyx99
left a comment
There was a problem hiding this comment.
Seeing if I can update the PR with comments
|
Thanks for running this @njhill, but the PR that this was pointing to just got merged so I'm going to update it the the vllm fork instead of my own. Can you rerun it again after that, or add the ready tag and I can rerun it? |
Signed-off-by: Chris Leonard <chleonar@redhat.com>
Signed-off-by: Chris Leonard <chleonar@redhat.com>
Signed-off-by: Chris Leonard <chleonar@redhat.com>
3326236 to
2a29cc5
Compare
|
✅ @cleonard530, CI is now available for this PR.
|
|
/ci run |
|
✅ Triggered Buildkite CI #94660 for commit |
Signed-off-by: Shengqi Chen <harry-chen@outlook.com>
|
/ci run |
|
✅ Triggered Buildkite CI #94667 for commit |
Harry-Chen
left a comment
There was a problem hiding this comment.
LGTM. Sorry for the CI breakage in the previous PR.
|
FYI, @Harry-Chen, @ywang96, @janeyx99. I did find another bug. If |
|
Hi @cleonard530 @Harry-Chen can we land this PR again after testing with torch 2.14? The CI in this PR ran before torch 2.14 upgrade, so it merged and clashed with torch 2.14 upgrade on main (landed like 12 hours ago) and caused all image build on CI to fail. More details here https://vllm-dev.slack.com/archives/C07R5PAL2L9/p1792014117298589. Sorry about this |
…ject#61491) Signed-off-by: Chris Leonard <chleonar@redhat.com> Signed-off-by: Shengqi Chen <harry-chen@outlook.com> Co-authored-by: Shengqi Chen <harry-chen@outlook.com> Signed-off-by: Andreas Karatzas <AndreasKaratzas@users.noreply.github.com>
…ject#61491) Signed-off-by: Chris Leonard <chleonar@redhat.com> Signed-off-by: Shengqi Chen <harry-chen@outlook.com> Co-authored-by: Shengqi Chen <harry-chen@outlook.com>
Overview
Update DeepGEMM to the top of this PR vllm-project/DeepGEMM#35.
The current DeepGEMM pin computes balanced SM locality domains during deep_gemm initialization. On a 60-SM MIG device, that reaches an existing balancing error during vLLM import, even when locality features are unused (see #59482 (comment)). This PR updates both the CMake dependency and standalone installer to cleonard530/DeepGEMM@6af2d3a, which restores lazy C++ locality initialization using stable PyTorch APIs.
The legacy balancing issue remains and can still surface when balanced locality domains are requested.cc @Harry-Chen @janeyx99 @khluu
Pull Request Checklist
I used vLLM's
/pr-checklistskill. (Mandatory for agents, optional for humans).AI assistance was used during the creation of this PR.
Design Fit: Minimizes impact on core components, reuses existing functionality, and justifies added complexity.
Testing and Validation: Validates the change and ensures any added tests are meaningful and reliable, with CI coverage or documented CI resource constraints and validation performed outside CI.
Code Quality and Style: Keeps code and comments clear and concise, and updates relevant documentation and examples.
Pull Request Contents: Includes a brief summary and relevant links, supports claims with evidence, explains root causes and implementation trade-offs, and follows the contributing guide.