---
id: "20260820-0221-do-microcosmoss-scaling-benchmark"
title: "Do Microcosmos's scaling benchmark (linear to ~500k particles on one L40S) and its four experimental results hold up against the primary paper's own figures?"
type: "capture"
status: "promoted"
origin: "batch"
promoted_to: ["30-notes/claim-microcosmos-chemotaxis-is-fitness-rewarded-not-emergent.md (new claim: chemotaxis is fitness-rewarded via an explicit energy term, not spontaneously emergent)","30-notes/claim-microcosmos-four-experiments-handdesigned-to-emergent.md (updated: Correction history block narrowing point 3's 'emergent chemotaxis' framing, plus wikilinks to new entity pages)","30-notes/claim-microcosmos-linear-scaling-avoids-quadratic-gpu-populations.md (updated: audit_status corroboration appending this session's Figure 5 visual-read detail)","30-notes/claim-microcosmos-elastic-filaments-differentiable-gpu-fluid.md (updated: wikilinks to new entity-bert-wang-chak-chan and entity-lattice-boltzmann-method pages)","40-entities/entity-edward-purcell.md (new hub)","40-entities/entity-jean-baptiste-mouret.md (new hub)","40-entities/entity-jeff-clune.md (new hub)","40-entities/entity-kenneth-o-stanley.md (new hub)","40-entities/entity-mark-tensen.md (new hub)","40-entities/entity-bert-wang-chak-chan.md (new hub)","40-entities/entity-lattice-boltzmann-method.md (new hub)","50-questions/question-verify-microcosmos-benchmarks-and-experiments.md (ruled: status -> answered, answered_log added)"]
not_promoted: ["Scaling benchmark claim (linear to ~500k particles, L40S) — collision with claim-microcosmos-linear-scaling-avoids-quadratic-gpu-populations.md, already promoted 2026-07-09 and cross-model audited. This capture's Figure 5 visual-read numbers (measured/O(N) ~1.4-1.6ms/step, O(N²) ~4ms/step at 500k) were appended to that note's audit_status as corroboration, not written as a new note — same claim, richer read, no new atomic content.","Experiment 1 (hand-designed locomotion, Purcell scallop theorem) — collision with claim-microcosmos-four-experiments-handdesigned-to-emergent.md point 1, already confirmed by the 2026-07-09 audit. No new note; Purcell wikilinked to his new entity page instead.","Experiment 2 (1000-node filament, Adam lr 0.04/300 steps, MNIST folding) — collision with the same note's point 2, already confirmed. This capture's extra hyperparameters (250-step rollout, 128x128 target resolution) are minor supporting detail, not a distinct atomic claim; left in the capture rather than promoted.","Figure 4 MAP-Elites archive-coverage curve (399/400 cells by generation 300) — a direct plot-read with no accompanying data table or text quote; the capture itself flags this as a future dedicated-note candidate, not a claim ready for promotion now. The headline number was folded into claim-microcosmos-four-experiments-handdesigned-to-emergent.md's point 4 as supporting detail rather than promoted standalone.","Tom Mohr / Particle Life (2023) entity candidate — single passing mention as the O(n^2) comparison baseline, thin and unconfirmed beyond this capture's own assertion of authorship. Per the entity promotion test's 'when unsure, don't promote,' left as a plain-text mention rather than a hub.","ALIFE (conference) entity candidate — this is a publication venue, not an actor in the arguments (it doesn't fund, build, decide, or block anything the claims rest on). Under the entity spec's stricter org bar, does not qualify; left as a mention.","Particle Life's own O(n^2) scaling and the Microcosmos GitHub code release — both explicitly flagged by the capture as unpursued further leads, not claims; left in the capture for a future session."]
writer_model: "claude-sonnet-5"
date_created: "2026-08-20T00:00:00.000Z"
provenance: "Batch research run, 2026-08-20, closing [[question-verify-microcosmos-benchmarks-and-experiments]]"
derived_from: []
tags: ["artificial-life","gpu-simulation","computational-scaling","benchmark-verification","differentiable-physics","neuroevolution","quality-diversity","map-elites","primary-source-verification","microcosmos"]
source_url: "https://arxiv.org/html/2607.02954v1"
source_sha: "ff318efa170e596c4b8c92f5f601fffb5efd31f81c8ca61c9ce7221ef4abd01b"
source_title: "Microcosmos: Reimagining Artificial Life for the GPU Era"
source_author: "Mark Tensen, Ciaran Regan, Bert Wang-Chak Chan, Mizuki Oka, Kenneth O. Stanley, Grisha Szep"
source_date: "2026-07-03T00:00:00.000Z"
source_venue: "arXiv:2607.02954 [cs.NE]; accepted at ALIFE 2026"
source_tier: 1
source_quote: "We validate the engine through four experiments... Linear scaling with particle count confirms the engine supports large-scale simulation."
seek_code_commit: "17d9798"
---


Short answer, on direct re-read of the primary paper: **yes.** Every specific
figure carried by the two existing seedling notes —
[[claim-microcosmos-linear-scaling-avoids-quadratic-gpu-populations]] and
[[claim-microcosmos-four-experiments-handdesigned-to-emergent]] — is present,
verbatim-groundable, and consistent with the paper's own text and plots. One
nuance surfaces on close reading of Experiment 3 (below) that the prior
paraphrase-only capture could not have caught.

This capture rests on two independent fetches of the same primary document,
both TLS-verified: the arXiv HTML rendering (`archive_page`, sha256
`ff318efa170e596c4b8c92f5f601fffb5efd31f81c8ca61c9ce7221ef4abd01b`, used for
all quotes below, matching the existing notes' `source_url` for citation
continuity) and the arXiv PDF (`extract_pdf`, sha256
`619559f78659974148e2d7d9e0c1588e598d2ac05f2ce845213367088e2e846f`), read as
rendered pages to inspect Figure 5 itself, since the HTML-to-text conversion
carries only figure captions, not plot geometry. The arXiv abstract page
(`archive_page`, sha256
`806c826d3f0a35fa363e3f2ca9aa17959a259a5456deaaa2ae3d54c03461c062`) confirms
publication metadata: submitted 3 Jul 2026, "Accepted at ALIFE 2026." All
quotes below were checked verbatim (normalized) against the cited text file
via `quote_check` before being recorded.

## Claim: The scaling benchmark holds up — linear wall-clock scaling to 500k particles on one L40S GPU, confirmed in the paper's own Methods text and Figure 5 plot

The paper states its own methodology plainly: "To demonstrate scalability, we
measure wall clock time as we increase the number of particles at a fixed
grid resolution," "running 1000 steps with up to 500k particles" (grid fixed
at 256×256), with a footnote reading "Ran with 1 NVIDIA L40S GPU, JAX version
0.8.1." The result: "As shown in fig.5, Microcosmos scales linearly, unlike
simulators that rely on pairwise interactions such as Particle Life," which
the paper says scale as O(n²).

Figure 5 was visually inspected directly (PDF page 8 of the extracted
document). It plots wall time per step (ms) against particle count (0 to
500k) with three series: a "Measured" line (blue, circular markers), an
"O(N)" dashed reference (orange), and an "O(N²)" dashed reference (green).
The measured line tracks closely with the O(N) reference across the full
range, both reaching roughly 1.4–1.6 ms/step at 500k particles, while the
O(N²) reference diverges sharply, reaching roughly 4 ms/step at 500k. This is
a direct visual read of the primary source's own figure, not an independent
statistical fit — no raw data table accompanies the plot, so a precise
linear-regression check (e.g. R²) is not possible from what the paper
publishes. Within that limit, the plot corroborates the linear-scaling text
claim; it does not show a super-linear tail breaking away from the O(N) line
within the tested range.

This is the paper's own reported benchmark, run once, on the authors'
hardware, with no data released for independent recomputation beyond the
plot itself. No independent (different-author, different-venue)
reproduction of this specific figure was found in this session — the claim
sits in the same "self-reported, unreplicated benchmark" pattern as
[[claim-no-independent-benchmark-of-encharge-en100-efficiency-figures]] and
[[claim-rapid-rapidx-speedups-fall-short-of-crisp-press-figure]] elsewhere in
the vault, though nothing here contradicts the figure the way those cases'
comparisons did.

## Claim: Hand-designed locomotion (Experiment 1) holds up — five geometries, Reynolds ~10³ to 1, matching Purcell's scallop theorem

The paper: "We demonstrate five hand-designed geometries and locomotion
strategies that validate the physics engine's fluid coupling." Viscosity is
swept across a range such that "this spans approximately Reynolds number
Re ∼ 10³ to" "1 depending on the swimmer, covering the transition from"
inertia-dominated to viscosity-dominated swimming. The result: "Consistent
with Purcell's scallop theorem, time-reversible strategies (e.g. ray) produce
zero net displacement at high viscosity, while non-reciprocal strategies
(e.g. cilia) maintain viable locomotion." The five named geometries (worm,
tadpole, ray, cilia, jellyfish) and the displacement-vs-viscosity plot
(Figure 2) match the quoted text exactly — this is a physics sanity check
the paper's own hand-picked parameters are built to pass, not a discovered
result, and it reads that way in the text.

## Claim: Filament folding (Experiment 2) holds up — a 1000-node filament optimized by Adam (lr 0.04, 300 steps) reconstructs all ten MNIST digit classes

The paper: "The angles and lengths of a 1000-node linear filament are
optimized directly using Adam" with "learning rate 0.04, 300 parameter
update" steps, "targeting MNIST digit shapes." Result: "Gradient descent
reliably recovers parameters that fold the filament into recognizable digit
shapes across all ten classes." This is the paper's differentiability
demonstration — gradients flow through the PBD solver and LBM fluid step via
JAX autodiff, per the Methods section — and the specific hyperparameters
(1000 nodes, Adam, lr 0.04, 300 steps, 250-step rollout, 128×128 target
resolution) all check out against the primary text.

## Claim: Neuroevolution and MAP-Elites (Experiments 3–4) hold up, with one correction — the "chemotaxis" result is fitness-*rewarded*, not spontaneously emergent

Two related searches, both confirmed:

- **NEAT/CPPN neuroevolution.** "The fitness function is defined as the
  total displacement traveled. We found that the highest fitness swimmers
  predominantly perform sinusoidal locomotion." A second run adds a food-like
  energy field, and here the paper is explicit about the mechanism: "fitness
  becomes the total displacement traveled plus the total energy collected,
  rewarding both movement and chemotaxis." This is a **direct reward-shaping
  mechanism**, not unprompted or spontaneous behavior — the evolutionary
  search is explicitly selecting for energy-seeking because the fitness
  function pays for it. The predecessor capture's paraphrase ("emergent
  chemotaxis") is defensible against the paper's own abstract language
  ("produce a wide range of swimming and chemotaxis behaviors
  automatically") but understates that the specific behavior was directly
  selected-for by an explicit reward term, not discovered incidentally under
  a displacement-only objective. This resolves item 2 of
  [[question-verify-microcosmos-benchmarks-and-experiments]]: the mechanism
  is rewarded chemotaxis via an added fitness term, described by the authors
  as happening "automatically" in the sense of evolutionary search
  discovering the *strategy* rather than a human hand-designing it — not in
  the sense of the objective failing to reward it.
- **MAP-Elites quality-diversity search.** "we use Multi-dimensional Archive
  of Phenotypic Elites (MAP-Elites)" — citation [38] in the HTML rendering,
  identified from the paper's own reference list as Mouret, J.-B. and Clune,
  J. (2015), "Illuminating search spaces by mapping elites" — "to search for
  diverse locomotion strategies," using bending effort and stretching effort
  as behavioral descriptors. Figure 4 reports archive coverage reaching 399
  of 400 cells by generation 300 (read directly off the plot, not quoted
  from text — a further-lead figure, not verified against a data table).

## Further leads

- Particle Life (Mohr, 2023) — cited as the O(n²) baseline Microcosmos claims to beat; not independently checked in this session whether Particle Life's own scaling has been separately benchmarked.
- Code release at https://github.com/alife-institute/microcosmos — not fetched this session; would let a future capture check whether the 500k-particle benchmark is independently reproducible rather than resting solely on the paper's own plot.
- Figure 4's MAP-Elites archive-coverage curve (399/400 cells by generation 300) — read off the plot directly, not text-quoted; a clean target for a dedicated QD-search note.
- Kenneth Stanley's broader open-endedness/POET lineage, and NEAT (Stanley & Miikkulainen, 2002) / CPPN (Stanley, 2007) as the specific prior methods the neuroevolution experiment builds on — flagged in the predecessor capture, still unpursued.
- No independent (different-author) benchmark or replication of Figure 5 was found in this session; if one surfaces later it would directly corroborate or contradict a currently single-source figure.

> [!note] Seek's commentary:
> The interesting finding here isn't that the benchmark checks out — self-reported figures from a paper's own plot usually do check out against that same plot, that's a low bar. The more useful catch is the chemotaxis nuance: "automatically" (abstract) and "emergent" (the prior capture's paraphrase) are doing real work to make a fitness-shaped, directly-rewarded behavior sound more like spontaneous discovery than the Methods section actually supports. Neither word is false, exactly — the *strategy* for collecting energy was found by search, not hand-coded — but a reader skimming only the abstract would come away thinking the creatures started foraging without being paid to. Worth remembering next time a paper's own abstract is the only source read.
> — Seek

## Entity candidates

- Edward Purcell — person — "Life at Low Reynolds Number" (1977) supplies the scallop theorem that Experiment 1 is explicitly validated against; the correctness claim for the hand-designed locomotion experiment rests on matching his 1977 prediction, not on anything original to Microcosmos. The foundational figure this capture's core claim measures itself against — flag first, per the vault's known blind spot.
- Jean-Baptiste Mouret — person — co-originator (with Jeff Clune) of MAP-Elites (2015), the QD-search algorithm Experiment 4 runs unmodified; the method Microcosmos's quality-diversity result is built on and measured against.
- Jeff Clune — person — co-originator of MAP-Elites (2015); see above.
- Kenneth O. Stanley — person — Microcosmos co-author, but also the originator of NEAT (2002) and CPPN (2007), the specific prior methods Experiment 3 (neuroevolution) directly reuses — his own earlier work is the substrate his co-authored paper measures against.
- Mark Tensen — person — first/corresponding author of Microcosmos (Artificial Life Institute, Japan).
- Tom Mohr — person — creator of "Particle Life" (2023), cited as the O(n²)-scaling comparison baseline the linear-scaling claim is measured against.
- Bert Wang-Chak Chan — person — Microcosmos co-author; also author of Lenia (2019), the abstract cellular-automaton substrate the paper's introduction argues against.
- ALIFE (conference) — concept — Microcosmos was "Accepted at ALIFE 2026" per the arXiv listing; the paper's peer-review venue.
- Lattice Boltzmann Method — concept — the D2Q9 fluid solver underlying the scaling claim's O(n) local-update mechanism; candidate for its own definitional note.