Skip to content

docs: add exact dedup tuning params to v26.04 docs and release notes - #1699

Merged
lbliii merged 3 commits into
NVIDIA-NeMo:26.04-stagingfrom
lbliii:lbliii/davis-v2-docs
Apr 3, 2026
Merged

docs: add exact dedup tuning params to v26.04 docs and release notes#1699
lbliii merged 3 commits into
NVIDIA-NeMo:26.04-stagingfrom
lbliii:lbliii/davis-v2-docs

Conversation

@lbliii

@lbliii lbliii commented Mar 31, 2026

Copy link
Copy Markdown
Contributor

Description

Documents three new ExactDeduplicationWorkflow parameters (total_nparts, rmm_pool_size, spill_memory_limit) exposed in PR #1561 for fine-grained control over large-scale cluster runs. Updates the configuration table and performance best practices in exact.mdx, and adds an Enhancements entry to the 26.04 release notes.

Usage

workflow = ExactDeduplicationWorkflow(
    input_path="input_data/",
    output_path="./results",
    total_nparts=512,
    rmm_pool_size=72 * 1024 * 1024 * 1024,  # 72 GiB
    spill_memory_limit="auto",
)

Checklist

  • I am familiar with the Contributing Guide.
  • New or Existing tests cover these changes.
  • The documentation is up to date with these changes.

@lbliii
lbliii requested a review from a team as a code owner March 31, 2026 19:28
@lbliii
lbliii requested review from ayushdg and removed request for a team March 31, 2026 19:28
Comment thread fern/versions/v26.04/pages/curate-text/process-data/deduplication/exact.mdx Outdated
@@ -163,6 +166,7 @@ The workflow produces these output files:
- Use `input_blocksize="2GiB"` for optimal performance

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

more for the #1369 docs but we should change the guidance here

lbliii and others added 3 commits April 3, 2026 10:48
Documents three new ExactDeduplicationWorkflow parameters exposed in
PR NVIDIA-NeMo#1561: total_nparts, rmm_pool_size, and spill_memory_limit.
Updates the configuration table and performance best practices in
exact.mdx, and adds an Enhancements entry to the 26.04 release notes.

Signed-off-by: Lawrence Lane <llane@nvidia.com>
…ion/exact.mdx

Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com>
Signed-off-by: Lawrence Lane <llane@nvidia.com>
Update total_nparts guidance per ayushdg's feedback: explicitly
recommend smaller values (256/512) for better shuffle performance
on large runs.

Signed-off-by: Logan Lane <llane@nvidia.com>
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Signed-off-by: Lawrence Lane <llane@nvidia.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants