Create Profiler when caller injects a reporter, bypassing NCCL_CTRAN_TRANSPORT_PROFILER gate (#2005)#2005
Open
mlunar-meta wants to merge 2 commits intometa-pytorch:mainfrom
Open
Create Profiler when caller injects a reporter, bypassing NCCL_CTRAN_TRANSPORT_PROFILER gate (#2005)#2005mlunar-meta wants to merge 2 commits intometa-pytorch:mainfrom
mlunar-meta wants to merge 2 commits intometa-pytorch:mainfrom
Conversation
meta-cla
bot
added
the
CLA Signed
Contributor
|
@mlunar-meta has exported this pull request. If you are a Meta employee, you can view the originating Diff in D100086191. |
Summary: Add a new MCCL-level CVAR to control ctran transport profiling from MCCL. Defaults to true so MCCL enables profiling without requiring users to set the NCCL_CTRAN_TRANSPORT_PROFILER env var. Differential Revision: D100181757
…TRANSPORT_PROFILER gate (meta-pytorch#2005) Summary: MCCL injects its own McclAlgoProfilerReporter into ctran via ctranInit(), but the ctran::Profiler was only created when NCCL_CTRAN_TRANSPORT_PROFILER=true (default: false). This silently dropped the MCCL reporter, preventing all algo profiler data from reaching the mccl_operation_trace Scuba table. Fix: create the Profiler whenever a custom reporter is injected (reporter != nullptr), regardless of NCCL_CTRAN_TRANSPORT_PROFILER. The NCCLx path (no reporter, defaults to DefaultAlgoProfilerReporter) retains the existing CVAR gate. MCCL already has its own enable gate (MCCL_SCUBA_ENABLED) inside McclAlgoProfilerReporter, so requiring a second NCCL-prefixed CVAR is redundant. Differential Revision: D100086191
meta-codesync
bot
changed the title
mlunar-meta
force-pushed
the
export-D100086191
branch
from
489320a to
253bbdf
Compare
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
1 participant
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.
Summary:
MCCL injects its own McclAlgoProfilerReporter into ctran via ctranInit(),
but the ctran::Profiler was only created when NCCL_CTRAN_TRANSPORT_PROFILER=true
(default: false). This silently dropped the MCCL reporter, preventing all algo
profiler data from reaching the mccl_operation_trace Scuba table.
Fix: create the Profiler whenever a custom reporter is injected (reporter != nullptr),
regardless of NCCL_CTRAN_TRANSPORT_PROFILER. The NCCLx path (no reporter, defaults to
DefaultAlgoProfilerReporter) retains the existing CVAR gate.
MCCL already has its own enable gate (MCCL_SCUBA_ENABLED) inside
McclAlgoProfilerReporter, so requiring a second NCCL-prefixed CVAR is redundant.
Differential Revision: D100086191