Repository navigation
[trainer throughput] Keep all experts on every GPU for Nemotron-3.5-Lightning 64K - #246
Merged
Micah Yong (micahtyong) merged 1 commit intoNov 17, 2026
Conversation
Micah Yong (micahtyong)
force-pushed
the
micah/nemotron-perf/trainer-3-expert-parallel-1
branch
from
November 17, 2026 02:22
eff6b35 to
91861f5
Compare
Base automatically changed from
micah/nemotron-perf/trainer-2-log-probs-chunk
to
main
November 17, 2026 02:33
Micah Yong (micahtyong)
force-pushed
the
micah/nemotron-perf/trainer-3-expert-parallel-1
branch
from
November 17, 2026 02:33
266c2f3 to
edc6695
Compare
Micah Yong (micahtyong)
marked this pull request as ready for review
November 17, 2026 02:42
Micah Yong (micahtyong)
requested a review
from Kailash Ranganathan (kailash109)
as a code owner
November 17, 2026 02:42
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Micah Yong (micahtyong)
force-pushed
the
micah/nemotron-perf/trainer-3-expert-parallel-1
branch
from
November 17, 2026 02:58
b4320f2 to
8a1d78a
Compare
Micah Yong (micahtyong)
deleted the
micah/nemotron-perf/trainer-3-expert-parallel-1
branch
November 17, 2026 03:31
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Nemotron-3.5-Lightning is a mixture-of-experts model with 128 experts per layer. The recipe split them across its two GPUs with
expert_model_parallel_size2, so each GPU held 64. In every layer each GPU then has to send tokens to the GPU that holds their expert and get the results back. A profile of one training step shows those exchanges, plus the gathers that go with them, take about 17% of GPU kernel time.The split is not needed for memory on H200. This PR sets
expert_model_parallel_sizeto 1 innemotron35_lightning_30b_a3b_lora_64k, so every GPU holds all 128 experts and no tokens move between GPUs. The 256k recipe sets its own value and is unchanged.Memory goes up. On the full stack, four clients each training a 64K-token row peak at 130 GB allocated on a 140 GB H200.
E2E results
Setup: 2x H200, one client, stock Tinker SDK, LoRA rank 32. Short rows are 128 rows of about 1,600 tokens per step, long rows are 8 rows of about 16,000 tokens, and 64K rows are 2 rows of 64K tokens. Per-token log probabilities match the previous PR to within run-to-run noise.