Files
obikmer/DevDoc/architecture/numa_worker_pools/index.html
T

1593 lines
34 KiB
HTML
Raw Normal View History

2026-04-16 22:38:20 +02:00
<!doctype html>
<html lang="en" class="no-js">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width,initial-scale=1">
2026-08-15 20:56:29 +02:00
<link rel="prev" href="../siblings/">
2026-04-16 22:38:20 +02:00
2026-08-15 20:56:29 +02:00
<link rel="next" href="../numa_partition_runner/">
2026-04-16 22:38:20 +02:00
<link rel="icon" href="../../assets/images/favicon.png">
<meta name="generator" content="mkdocs-1.6.1, mkdocs-material-9.7.6">
2026-08-15 20:56:29 +02:00
<title>NUMA-aware worker pools - obikmer</title>
2026-04-16 22:38:20 +02:00
<link rel="stylesheet" href="../../assets/stylesheets/main.484c7ddc.min.css">
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
<link rel="stylesheet" href="https://fonts.googleapis.com/css?family=Roboto:300,300i,400,400i,700,700i%7CRoboto+Mono:400,400i,700,700i&display=fallback">
<style>:root{--md-text-font:"Roboto";--md-code-font:"Roboto Mono"}</style>
<script>__md_scope=new URL("../..",location),__md_hash=e=>[...e].reduce(((e,_)=>(e<<5)-e+_.charCodeAt(0)),0),__md_get=(e,_=localStorage,t=__md_scope)=>JSON.parse(_.getItem(t.pathname+"."+e)),__md_set=(e,_,t=localStorage,a=__md_scope)=>{try{t.setItem(a.pathname+"."+e,JSON.stringify(_))}catch(e){}}</script>
</head>
<body dir="ltr">
<input class="md-toggle" data-md-toggle="drawer" type="checkbox" id="__drawer" autocomplete="off">
<input class="md-toggle" data-md-toggle="search" type="checkbox" id="__search" autocomplete="off">
<label class="md-overlay" for="__drawer"></label>
<div data-md-component="skip">
2026-08-15 20:56:29 +02:00
<a href="#numa-aware-worker-pools-for-merge" class="md-skip">
2026-04-16 22:38:20 +02:00
Skip to content
</a>
</div>
<div data-md-component="announce">
</div>
<header class="md-header md-header--shadow" data-md-component="header">
<nav class="md-header__inner md-grid" aria-label="Header">
<a href="../.." title="obikmer" class="md-header__button md-logo" aria-label="obikmer" data-md-component="logo">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M12 8a3 3 0 0 0 3-3 3 3 0 0 0-3-3 3 3 0 0 0-3 3 3 3 0 0 0 3 3m0 3.54C9.64 9.35 6.5 8 3 8v11c3.5 0 6.64 1.35 9 3.54 2.36-2.19 5.5-3.54 9-3.54V8c-3.5 0-6.64 1.35-9 3.54"/></svg>
</a>
<label class="md-header__button md-icon" for="__drawer">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M3 6h18v2H3zm0 5h18v2H3zm0 5h18v2H3z"/></svg>
</label>
<div class="md-header__title" data-md-component="header-title">
<div class="md-header__ellipsis">
<div class="md-header__topic">
<span class="md-ellipsis">
obikmer
</span>
</div>
<div class="md-header__topic" data-md-component="header-topic">
<span class="md-ellipsis">
2026-08-15 20:56:29 +02:00
NUMA-aware worker pools
2026-04-16 22:38:20 +02:00
</span>
</div>
</div>
</div>
<script>var palette=__md_get("__palette");if(palette&&palette.color){if("(prefers-color-scheme)"===palette.color.media){var media=matchMedia("(prefers-color-scheme: light)"),input=document.querySelector(media.matches?"[data-md-color-media='(prefers-color-scheme: light)']":"[data-md-color-media='(prefers-color-scheme: dark)']");palette.color.media=input.getAttribute("data-md-color-media"),palette.color.scheme=input.getAttribute("data-md-color-scheme"),palette.color.primary=input.getAttribute("data-md-color-primary"),palette.color.accent=input.getAttribute("data-md-color-accent")}for(var[key,value]of Object.entries(palette.color))document.body.setAttribute("data-md-color-"+key,value)}</script>
</nav>
</header>
<div class="md-container" data-md-component="container">
<main class="md-main" data-md-component="main">
<div class="md-main__inner md-grid">
<div class="md-sidebar md-sidebar--primary" data-md-component="sidebar" data-md-type="navigation" >
<div class="md-sidebar__scrollwrap">
<div class="md-sidebar__inner">
<nav class="md-nav md-nav--primary" aria-label="Navigation" data-md-level="0">
<label class="md-nav__title" for="__drawer">
<a href="../.." title="obikmer" class="md-nav__button md-logo" aria-label="obikmer" data-md-component="logo">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M12 8a3 3 0 0 0 3-3 3 3 0 0 0-3-3 3 3 0 0 0-3 3 3 3 0 0 0 3 3m0 3.54C9.64 9.35 6.5 8 3 8v11c3.5 0 6.64 1.35 9 3.54 2.36-2.19 5.5-3.54 9-3.54V8c-3.5 0-6.64 1.35-9 3.54"/></svg>
</a>
obikmer
</label>
<ul class="md-nav__list" data-md-scrollfix>
<li class="md-nav__item">
<a href="../.." class="md-nav__link">
<span class="md-ellipsis">
Home
</span>
</a>
</li>
2026-08-15 20:56:29 +02:00
<li class="md-nav__item">
<a href="../../installation/" class="md-nav__link">
<span class="md-ellipsis">
2026-04-16 22:38:20 +02:00
2026-08-15 20:56:29 +02:00
Installation
</span>
</a>
</li>
2026-04-16 22:38:20 +02:00
2026-08-15 20:56:29 +02:00
<li class="md-nav__item md-nav__item--nested">
2026-04-16 22:38:20 +02:00
2026-08-15 20:56:29 +02:00
<input class="md-nav__toggle md-toggle " type="checkbox" id="__nav_3" >
2026-04-16 22:38:20 +02:00
2026-08-15 20:56:29 +02:00
<label class="md-nav__link" for="__nav_3" id="__nav_3_label" tabindex="0">
2026-04-16 22:38:20 +02:00
<span class="md-ellipsis">
Theory
</span>
<span class="md-nav__icon md-icon"></span>
</label>
2026-08-15 20:56:29 +02:00
<nav class="md-nav" data-md-level="1" aria-labelledby="__nav_3_label" aria-expanded="false">
<label class="md-nav__title" for="__nav_3">
2026-04-16 22:38:20 +02:00
<span class="md-nav__icon md-icon"></span>
Theory
</label>
<ul class="md-nav__list" data-md-scrollfix>
<li class="md-nav__item">
<a href="../../kmers/" class="md-nav__link">
2026-04-16 22:38:20 +02:00
<span class="md-ellipsis">
Kmers and super-kmers
</span>
</a>
</li>
<li class="md-nav__item">
2026-08-15 20:56:29 +02:00
<a href="../../theory/encoding/" class="md-nav__link">
2026-04-16 22:38:20 +02:00
<span class="md-ellipsis">
DNA encoding
</span>
</a>
</li>
2026-08-15 20:56:29 +02:00
<li class="md-nav__item">
<a href="../../theory/entropy/" class="md-nav__link">
2026-04-16 22:38:20 +02:00
<span class="md-ellipsis">
Entropy filter
</span>
</a>
</li>
<li class="md-nav__item">
2026-08-15 20:56:29 +02:00
<a href="../../theory/minimizer/" class="md-nav__link">
<span class="md-ellipsis">
Minimizer selection
</span>
</a>
</li>
2026-04-16 22:38:20 +02:00
<li class="md-nav__item">
2026-08-15 20:56:29 +02:00
<a href="../../theory/indexing/" class="md-nav__link">
2026-04-16 22:38:20 +02:00
<span class="md-ellipsis">
Partitioning architecture
</span>
</a>
</li>
2026-08-15 20:56:29 +02:00
<li class="md-nav__item">
<a href="../../theory/evolutionary_distances/" class="md-nav__link">
<span class="md-ellipsis">
Central-position SNP distance (discussion)
</span>
</a>
</li>
2026-04-16 22:38:20 +02:00
</ul>
</nav>
</li>
<li class="md-nav__item md-nav__item--nested">
2026-08-15 20:56:29 +02:00
<input class="md-nav__toggle md-toggle " type="checkbox" id="__nav_4" >
2026-04-16 22:38:20 +02:00
2026-08-15 20:56:29 +02:00
<label class="md-nav__link" for="__nav_4" id="__nav_4_label" tabindex="0">
2026-04-16 22:38:20 +02:00
<span class="md-ellipsis">
Implementation
</span>
<span class="md-nav__icon md-icon"></span>
</label>
2026-08-15 20:56:29 +02:00
<nav class="md-nav" data-md-level="1" aria-labelledby="__nav_4_label" aria-expanded="false">
<label class="md-nav__title" for="__nav_4">
2026-04-16 22:38:20 +02:00
<span class="md-nav__icon md-icon"></span>
Implementation
</label>
<ul class="md-nav__list" data-md-scrollfix>
<li class="md-nav__item">
<a href="../../implementation/superkmer/" class="md-nav__link">
<span class="md-ellipsis">
SuperKmer
</span>
</a>
</li>
<li class="md-nav__item">
<a href="../../implementation/kmer/" class="md-nav__link">
<span class="md-ellipsis">
Kmer
</span>
</a>
</li>
<li class="md-nav__item">
<a href="../../implementation/chunkreader/" class="md-nav__link">
<span class="md-ellipsis">
Chunk reader
</span>
</a>
</li>
<li class="md-nav__item">
<a href="../../implementation/pipeline/" class="md-nav__link">
<span class="md-ellipsis">
Construction pipeline
</span>
</a>
</li>
<li class="md-nav__item">
<a href="../../implementation/obipipeline/" class="md-nav__link">
<span class="md-ellipsis">
obipipeline library
</span>
</a>
</li>
2026-04-16 22:38:20 +02:00
<li class="md-nav__item">
<a href="../../implementation/storage/" class="md-nav__link">
<span class="md-ellipsis">
On-disk storage
</span>
</a>
</li>
<li class="md-nav__item">
<a href="../../implementation/mphf/" class="md-nav__link">
<span class="md-ellipsis">
MPHF selection
</span>
</a>
</li>
<li class="md-nav__item">
<a href="../../implementation/unitig_evidence/" class="md-nav__link">
<span class="md-ellipsis">
Unitig evidence encoding
</span>
</a>
</li>
<li class="md-nav__item">
<a href="../../implementation/evidence_elimination/" class="md-nav__link">
<span class="md-ellipsis">
Evidence elimination (discussion)
</span>
</a>
</li>
<li class="md-nav__item">
<a href="../../implementation/obilayeredmap/" class="md-nav__link">
<span class="md-ellipsis">
obilayeredmap crate
</span>
</a>
</li>
<li class="md-nav__item">
<a href="../../implementation/persistent_compact_int_vec/" class="md-nav__link">
<span class="md-ellipsis">
PersistentCompactIntVec
</span>
</a>
</li>
<li class="md-nav__item">
<a href="../../implementation/persistent_bit_vec/" class="md-nav__link">
<span class="md-ellipsis">
PersistentBitVec
</span>
</a>
</li>
<li class="md-nav__item">
<a href="../../implementation/merge/" class="md-nav__link">
<span class="md-ellipsis">
Merge command
</span>
</a>
</li>
<li class="md-nav__item">
2026-08-15 20:56:29 +02:00
<a href="../../implementation/merge_parallelism/" class="md-nav__link">
<span class="md-ellipsis">
2026-08-15 20:56:29 +02:00
Merge parallelism & memory
</span>
</a>
</li>
<li class="md-nav__item">
<a href="../../implementation/filtering/" class="md-nav__link">
<span class="md-ellipsis">
Kmer filtering
</span>
</a>
</li>
<li class="md-nav__item">
<a href="../../implementation/select/" class="md-nav__link">
<span class="md-ellipsis">
Select command
</span>
</a>
</li>
<li class="md-nav__item">
<a href="../../implementation/obitaxonomy/" class="md-nav__link">
<span class="md-ellipsis">
obitaxonomy crate
</span>
</a>
</li>
2026-04-16 22:38:20 +02:00
</ul>
</nav>
</li>
2026-08-15 20:56:29 +02:00
2026-04-16 22:38:20 +02:00
2026-08-15 20:56:29 +02:00
<li class="md-nav__item md-nav__item--active md-nav__item--nested">
2026-04-16 22:38:20 +02:00
2026-08-15 20:56:29 +02:00
<input class="md-nav__toggle md-toggle " type="checkbox" id="__nav_5" checked>
2026-04-16 22:38:20 +02:00
2026-08-15 20:56:29 +02:00
<label class="md-nav__link" for="__nav_5" id="__nav_5_label" tabindex="0">
2026-04-16 22:38:20 +02:00
<span class="md-ellipsis">
Architecture
</span>
<span class="md-nav__icon md-icon"></span>
</label>
2026-08-15 20:56:29 +02:00
<nav class="md-nav" data-md-level="1" aria-labelledby="__nav_5_label" aria-expanded="true">
<label class="md-nav__title" for="__nav_5">
2026-04-16 22:38:20 +02:00
<span class="md-nav__icon md-icon"></span>
Architecture
</label>
<ul class="md-nav__list" data-md-scrollfix>
<li class="md-nav__item">
2026-08-15 20:56:29 +02:00
<a href="../sequences/invariant/" class="md-nav__link">
2026-04-16 22:38:20 +02:00
<span class="md-ellipsis">
Sequences
</span>
</a>
</li>
<li class="md-nav__item">
2026-08-15 20:56:29 +02:00
<a href="../index_architecture/" class="md-nav__link">
<span class="md-ellipsis">
Kmer index
</span>
</a>
</li>
2026-08-15 20:56:29 +02:00
<li class="md-nav__item">
<a href="../siblings/" class="md-nav__link">
<span class="md-ellipsis">
Sibling annex (discussion)
</span>
</a>
</li>
<li class="md-nav__item md-nav__item--active">
<input class="md-nav__toggle md-toggle" type="checkbox" id="__toc">
<label class="md-nav__link md-nav__link--active" for="__toc">
<span class="md-ellipsis">
NUMA-aware worker pools
</span>
<span class="md-nav__icon md-icon"></span>
</label>
<a href="./" class="md-nav__link md-nav__link--active">
<span class="md-ellipsis">
NUMA-aware worker pools
</span>
</a>
<nav class="md-nav md-nav--secondary" aria-label="Table of contents">
<label class="md-nav__title" for="__toc">
<span class="md-nav__icon md-icon"></span>
Table of contents
</label>
<ul class="md-nav__list" data-md-component="toc" data-md-scrollfix>
<li class="md-nav__item">
<a href="#problem" class="md-nav__link">
<span class="md-ellipsis">
Problem
</span>
</a>
</li>
<li class="md-nav__item">
<a href="#strategy" class="md-nav__link">
<span class="md-ellipsis">
Strategy
</span>
</a>
</li>
<li class="md-nav__item">
<a href="#rayon-thread-pool-isolation" class="md-nav__link">
<span class="md-ellipsis">
Rayon thread pool isolation
</span>
</a>
</li>
<li class="md-nav__item">
<a href="#thread-pinning" class="md-nav__link">
<span class="md-ellipsis">
Thread pinning
</span>
</a>
</li>
<li class="md-nav__item">
<a href="#memory-locality" class="md-nav__link">
<span class="md-ellipsis">
Memory locality
</span>
</a>
</li>
<li class="md-nav__item">
<a href="#adaptive-spawn-criterion" class="md-nav__link">
<span class="md-ellipsis">
Adaptive spawn criterion
</span>
</a>
</li>
<li class="md-nav__item">
<a href="#required-changes" class="md-nav__link">
<span class="md-ellipsis">
Required changes
</span>
</a>
</li>
<li class="md-nav__item">
<a href="#platform-guard" class="md-nav__link">
<span class="md-ellipsis">
Platform guard
</span>
</a>
</li>
<li class="md-nav__item">
<a href="#open-questions" class="md-nav__link">
<span class="md-ellipsis">
Open questions
</span>
</a>
</li>
</ul>
</nav>
</li>
<li class="md-nav__item">
<a href="../numa_partition_runner/" class="md-nav__link">
<span class="md-ellipsis">
NUMA-aware partition runner
</span>
</a>
</li>
2026-04-16 22:38:20 +02:00
</ul>
</nav>
</li>
</ul>
</nav>
</div>
</div>
</div>
<div class="md-sidebar md-sidebar--secondary" data-md-component="sidebar" data-md-type="toc" >
<div class="md-sidebar__scrollwrap">
<div class="md-sidebar__inner">
<nav class="md-nav md-nav--secondary" aria-label="Table of contents">
<label class="md-nav__title" for="__toc">
<span class="md-nav__icon md-icon"></span>
Table of contents
</label>
<ul class="md-nav__list" data-md-component="toc" data-md-scrollfix>
<li class="md-nav__item">
2026-08-15 20:56:29 +02:00
<a href="#problem" class="md-nav__link">
2026-04-16 22:38:20 +02:00
<span class="md-ellipsis">
2026-08-15 20:56:29 +02:00
Problem
2026-04-16 22:38:20 +02:00
</span>
</a>
</li>
<li class="md-nav__item">
2026-08-15 20:56:29 +02:00
<a href="#strategy" class="md-nav__link">
2026-04-16 22:38:20 +02:00
<span class="md-ellipsis">
2026-08-15 20:56:29 +02:00
Strategy
2026-04-16 22:38:20 +02:00
</span>
</a>
</li>
<li class="md-nav__item">
2026-08-15 20:56:29 +02:00
<a href="#rayon-thread-pool-isolation" class="md-nav__link">
2026-04-16 22:38:20 +02:00
<span class="md-ellipsis">
2026-08-15 20:56:29 +02:00
Rayon thread pool isolation
2026-04-16 22:38:20 +02:00
</span>
</a>
</li>
<li class="md-nav__item">
2026-08-15 20:56:29 +02:00
<a href="#thread-pinning" class="md-nav__link">
2026-04-16 22:38:20 +02:00
<span class="md-ellipsis">
2026-08-15 20:56:29 +02:00
Thread pinning
2026-04-16 22:38:20 +02:00
</span>
</a>
</li>
<li class="md-nav__item">
2026-08-15 20:56:29 +02:00
<a href="#memory-locality" class="md-nav__link">
2026-04-16 22:38:20 +02:00
<span class="md-ellipsis">
2026-08-15 20:56:29 +02:00
Memory locality
2026-04-16 22:38:20 +02:00
</span>
</a>
</li>
<li class="md-nav__item">
2026-08-15 20:56:29 +02:00
<a href="#adaptive-spawn-criterion" class="md-nav__link">
2026-04-16 22:38:20 +02:00
<span class="md-ellipsis">
2026-08-15 20:56:29 +02:00
Adaptive spawn criterion
2026-04-16 22:38:20 +02:00
</span>
</a>
</li>
<li class="md-nav__item">
2026-08-15 20:56:29 +02:00
<a href="#required-changes" class="md-nav__link">
2026-04-16 22:38:20 +02:00
<span class="md-ellipsis">
2026-08-15 20:56:29 +02:00
Required changes
</span>
</a>
</li>
<li class="md-nav__item">
<a href="#platform-guard" class="md-nav__link">
<span class="md-ellipsis">
Platform guard
</span>
</a>
</li>
<li class="md-nav__item">
<a href="#open-questions" class="md-nav__link">
<span class="md-ellipsis">
Open questions
2026-04-16 22:38:20 +02:00
</span>
</a>
</li>
</ul>
</nav>
</div>
</div>
</div>
<div class="md-content" data-md-component="content">
<article class="md-content__inner md-typeset">
2026-08-15 20:56:29 +02:00
<h1 id="numa-aware-worker-pools-for-merge">NUMA-aware worker pools for merge</h1>
<h2 id="problem">Problem</h2>
<p>The merge command's bottleneck is <code>compute_degrees</code> in <code>obidebruinj</code>: a random pointer-chase over 2070 M node hash maps that saturates DRAM bandwidth. When multiple partition workers run concurrently, they contend for the shared memory bus, causing super-linear slowdown (measured: 0.016 µs/node solo → 0.95 µs/node with 45 concurrent workers, ×60 degradation).</p>
<p>Modern HPC nodes are multi-socket NUMA machines (observed: 2 sockets × 4 NUMA nodes × 24 cores = 192 cores). Cross-NUMA memory traffic compounds the contention:</p>
2026-04-16 22:38:20 +02:00
<ul>
2026-08-15 20:56:29 +02:00
<li>Full 192-core run: ~15 min/partition (×10 worse than M3 Mac)</li>
<li><code>taskset</code> restricted to 4 NUMA nodes (96 cores): ~90 s/partition</li>
<li>OAR job on 1 NUMA node (24 cores): ~80 s/partition, same throughput as 96 cores</li>
</ul>
<p><strong>Conclusion</strong>: the bottleneck is memory bandwidth per NUMA node, not core count. 24 cores on one NUMA node achieve the same throughput as 96 cores across four.</p>
<h2 id="strategy">Strategy</h2>
<p>Run N worker groups in parallel, one per NUMA node, each with its own Rayon thread pool whose threads are pinned to the NUMA node's CPUs. Linux's first-touch policy then places graph allocations on local DRAM automatically — no explicit NUMA allocator needed.</p>
<p>Expected throughput: N × single-NUMA throughput. On the 8-NUMA-node HPC: 8 × ~80 s = 910 min total instead of &gt;60 min with the current single-pool approach.</p>
<h2 id="rayon-thread-pool-isolation">Rayon thread pool isolation</h2>
<p>Rayon provides <code>ThreadPool::install(|| { ... })</code>: any Rayon call (<code>par_iter</code>, <code>current_num_threads</code>, etc.) inside the closure uses <em>that</em> pool exclusively. Wrapping <code>merge_partition</code> in <code>pool.install()</code> redirects all downstream Rayon calls — including those in <code>debruijn.rs</code> and <code>partition.rs</code> — without touching those crates.</p>
<div class="highlight"><pre><span></span><code><span class="c1">// worker thread, assigned to NUMA pool `pool`</span>
<span class="n">pool</span><span class="p">.</span><span class="n">install</span><span class="p">(</span><span class="o">||</span><span class="w"> </span><span class="p">{</span>
<span class="w"> </span><span class="n">dst_partition</span><span class="p">.</span><span class="n">merge_partition</span><span class="p">(</span><span class="n">i</span><span class="p">,</span><span class="w"> </span><span class="n">srcs</span><span class="p">,</span><span class="w"> </span><span class="n">mode</span><span class="p">,</span><span class="w"> </span><span class="n">n_dst_genomes</span><span class="p">,</span><span class="w"> </span><span class="n">block_bits</span><span class="p">,</span><span class="w"> </span><span class="n">evidence</span><span class="p">)</span>
<span class="p">})</span>
</code></pre></div>
<p><code>rayon::current_num_threads()</code> inside <code>merge_partition</code> will return the pool size (e.g. 24), not the global thread count — which is the right value for buffer sizing.</p>
<h2 id="thread-pinning">Thread pinning</h2>
<p><code>ThreadPoolBuilder::spawn_handler</code> provides a hook executed for each thread at creation. Inside, <code>libc::sched_setaffinity</code> pins the thread to a CPU set:</p>
<div class="highlight"><pre><span></span><code><span class="kd">let</span><span class="w"> </span><span class="n">cpus</span><span class="p">:</span><span class="w"> </span><span class="nb">Vec</span><span class="o">&lt;</span><span class="kt">usize</span><span class="o">&gt;</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">numa_node_cpus</span><span class="p">(</span><span class="n">node</span><span class="p">);</span><span class="w"> </span><span class="c1">// from /sys/devices/system/node/nodeN/cpulist</span>
<span class="n">rayon</span><span class="p">::</span><span class="n">ThreadPoolBuilder</span><span class="p">::</span><span class="n">new</span><span class="p">()</span>
<span class="w"> </span><span class="p">.</span><span class="n">num_threads</span><span class="p">(</span><span class="n">cpus</span><span class="p">.</span><span class="n">len</span><span class="p">())</span>
<span class="w"> </span><span class="p">.</span><span class="n">spawn_handler</span><span class="p">(</span><span class="k">move</span><span class="w"> </span><span class="o">|</span><span class="n">thread</span><span class="o">|</span><span class="w"> </span><span class="p">{</span>
<span class="w"> </span><span class="kd">let</span><span class="w"> </span><span class="k">mut</span><span class="w"> </span><span class="n">b</span><span class="w"> </span><span class="o">=</span><span class="w"> </span><span class="n">std</span><span class="p">::</span><span class="n">thread</span><span class="p">::</span><span class="n">Builder</span><span class="p">::</span><span class="n">new</span><span class="p">();</span>
<span class="w"> </span><span class="n">std</span><span class="p">::</span><span class="n">thread</span><span class="p">::</span><span class="n">Builder</span><span class="p">::</span><span class="n">new</span><span class="p">().</span><span class="n">spawn</span><span class="p">(</span><span class="k">move</span><span class="w"> </span><span class="o">||</span><span class="w"> </span><span class="p">{</span>
<span class="w"> </span><span class="n">pin_to_cpus</span><span class="p">(</span><span class="o">&amp;</span><span class="n">cpus</span><span class="p">);</span><span class="w"> </span><span class="c1">// sched_setaffinity via libc</span>
<span class="w"> </span><span class="n">thread</span><span class="p">.</span><span class="n">run</span><span class="p">()</span>
<span class="w"> </span><span class="p">})</span><span class="o">?</span><span class="p">;</span>
<span class="w"> </span><span class="nb">Ok</span><span class="p">(())</span>
<span class="w"> </span><span class="p">})</span>
<span class="w"> </span><span class="p">.</span><span class="n">build</span><span class="p">()</span><span class="o">?</span>
</code></pre></div>
<p>NUMA topology is read from <code>/sys/devices/system/node/node*/cpulist</code> — no <code>libnuma</code> dependency required. If the <code>numa</code> crate is linked, <code>numa_available()</code> / <code>numa_run_on_node()</code> are an alternative.</p>
<h2 id="memory-locality">Memory locality</h2>
<p>Linux allocates pages on the NUMA node of the thread that first writes them (first-touch policy). Once Rayon threads are pinned to node N, all graph data built by those threads lands on node N's DRAM. No changes to the allocator, no explicit <code>numa_alloc_onnode</code> calls.</p>
<h2 id="adaptive-spawn-criterion">Adaptive spawn criterion</h2>
<p>The current criterion uses <code>std::thread::available_parallelism()</code> (returns total cores = 192) and <code>max_workers = n_cores / 2</code>. With NUMA pools:</p>
<ul>
<li><code>n_cores</code> per pool = cores per NUMA node (e.g. 24)</li>
<li><code>max_workers</code> per pool = pool size / 2 (e.g. 12)</li>
<li>CPU efficiency is measured per pool, not globally</li>
</ul>
<p>Each NUMA group runs its own independent adaptive pool. Workers are distributed across NUMA groups round-robin or by workload (partition assignment can be pre-split by NUMA group index).</p>
<h2 id="required-changes">Required changes</h2>
<table>
<thead>
<tr>
<th>File</th>
<th>Change</th>
</tr>
</thead>
<tbody>
<tr>
<td><code>obikindex/src/merge.rs</code></td>
<td>Detect NUMA topology; build N <code>ThreadPool</code>s with pinned threads; assign each pre-spawned worker to a pool; wrap <code>merge_partition</code> in <code>pool.install()</code></td>
</tr>
<tr>
<td><code>obikindex/src/merge.rs</code></td>
<td>Replace <code>available_parallelism()</code> with per-NUMA core count for spawn criterion</td>
</tr>
<tr>
<td><code>obikpartitionner/src/merge_layer.rs</code></td>
<td>No change — <code>merge_partition</code> already works inside any Rayon context</td>
</tr>
<tr>
<td><code>obidebruinj/src/debruijn.rs</code></td>
<td>No change — <code>par_iter</code> and <code>current_num_threads</code> are pool-context-aware</td>
</tr>
<tr>
<td><code>obikpartitionner/src/partition.rs</code></td>
<td>No change — same reason</td>
</tr>
</tbody>
</table>
<h2 id="platform-guard">Platform guard</h2>
<p>NUMA pinning is Linux-only. The fallback is the current single global pool:</p>
<div class="highlight"><pre><span></span><code><span class="cp">#[cfg(target_os = </span><span class="s">&quot;linux&quot;</span><span class="cp">)]</span>
<span class="k">fn</span><span class="w"> </span><span class="nf">build_numa_pools</span><span class="p">()</span><span class="w"> </span><span class="p">-&gt;</span><span class="w"> </span><span class="nb">Option</span><span class="o">&lt;</span><span class="nb">Vec</span><span class="o">&lt;</span><span class="n">rayon</span><span class="p">::</span><span class="n">ThreadPool</span><span class="o">&gt;&gt;</span><span class="w"> </span><span class="p">{</span><span class="w"> </span><span class="o">..</span><span class="p">.</span><span class="w"> </span><span class="p">}</span>
<span class="cp">#[cfg(not(target_os = </span><span class="s">&quot;linux&quot;</span><span class="cp">))]</span>
<span class="k">fn</span><span class="w"> </span><span class="nf">build_numa_pools</span><span class="p">()</span><span class="w"> </span><span class="p">-&gt;</span><span class="w"> </span><span class="nb">Option</span><span class="o">&lt;</span><span class="nb">Vec</span><span class="o">&lt;</span><span class="n">rayon</span><span class="p">::</span><span class="n">ThreadPool</span><span class="o">&gt;&gt;</span><span class="w"> </span><span class="p">{</span><span class="w"> </span><span class="nb">None</span><span class="w"> </span><span class="p">}</span>
</code></pre></div>
<p>When <code>build_numa_pools()</code> returns <code>None</code> (macOS, UMA, or single-socket), <code>merge.rs</code> uses the existing code path unchanged.</p>
<h2 id="open-questions">Open questions</h2>
<ul>
<li><strong>Partition assignment</strong>: split partitions by NUMA group up-front (static) or use a shared queue with per-group workers stealing from a common pool? Static split is simpler; stealing is better for load balance when partitions vary widely in size.</li>
<li><strong>Intra-NUMA adaptive criterion</strong>: with 24 cores and ~35 effective workers per NUMA node, the current marginal-gain criterion needs re-tuning or can be left as-is with per-pool <code>n_cores = 24</code>.</li>
<li><strong>I/O</strong>: partition data (unitig files) is on a shared filesystem. With 8 concurrent NUMA groups, I/O concurrency increases 8× — need to verify the filesystem (Lustre or local SSD) can absorb it without becoming the new bottleneck.</li>
2026-04-16 22:38:20 +02:00
</ul>
</article>
</div>
<script>var target=document.getElementById(location.hash.slice(1));target&&target.name&&(target.checked=target.name.startsWith("__tabbed_"))</script>
</div>
</main>
<footer class="md-footer">
<div class="md-footer-meta md-typeset">
<div class="md-footer-meta__inner md-grid">
<div class="md-copyright">
Made with
<a href="https://squidfunk.github.io/mkdocs-material/" target="_blank" rel="noopener">
Material for MkDocs
</a>
</div>
</div>
</div>
</footer>
</div>
<div class="md-dialog" data-md-component="dialog">
<div class="md-dialog__inner md-typeset"></div>
</div>
<script id="__config" type="application/json">{"annotate": null, "base": "../..", "features": [], "search": "../../assets/javascripts/workers/search.2c215733.min.js", "tags": null, "translations": {"clipboard.copied": "Copied to clipboard", "clipboard.copy": "Copy to clipboard", "search.result.more.one": "1 more on this page", "search.result.more.other": "# more on this page", "search.result.none": "No matching documents", "search.result.one": "1 matching document", "search.result.other": "# matching documents", "search.result.placeholder": "Type to start searching", "search.result.term.missing": "Missing", "select.version": "Select version"}, "version": null}</script>
<script src="../../assets/javascripts/bundle.79ae519e.min.js"></script>
<script src="https://unpkg.com/mathjax@3/es5/tex-mml-chtml.js"></script>
</body>
</html>