IB CI fixes: correct CQ-full warning bound, drain in-flight posts on stopProxy, expand multi-QP perf tests#808
Merged
Merged
Conversation
Contributor
Author
|
/azp run |
|
Azure Pipelines successfully started running 5 pipeline(s). |
chhwang
force-pushed
the
chhwang/ib-ci-fix
branch
from
June 23, 2026 17:02
2d85947 to
4c624ce
Compare
chhwang
changed the title
DefaultMaxCqPollNum to 256 & minor IB CI fixes
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Corrects the send-CQ overflow warning, closes a proxy-shutdown leak, and reworks the multi-QP perf tests. No production API change.
Changes
IbQp(src/core/ib.cc,src/core/include/ib.hpp): track the send CQ size explicitly via newmaxSendCqSize_and use it for the "CQ almost full" warning instead ofsendCq_->cqe. Fix warning message spacing.ProxyService::stopProxy(src/core/port_channel.cc): after draining pendingTriggerSyncflushes, flush remaining in-flight signaled posts per connection (5 s timeout) so shutdown leaves no unflushed sends. Reword the drain comment.test/mp_unit/port_channel_tests.cu): rewritekernelMultiQpBandwidthto one thread per QP with parallel waits; split the combined{1,2,4}loops into individually namedPERF_TESTs (Single/Two/Three/Multi QP); tunenItersand extend the size sweeps; report aggregate GB/s in the flush-stress test.