Skip to content
#software testing Open access

GALAXY v0.7.0–v0.8.0: Barnes–Hut Self-Gravity, Parallel GPU Tree Construction, and Hardware Evidence Harness

Sep 2026 · Zenodo (CERN European Organization for Nuclear Research)

Abstract

# GALAXY v0.7.0–v0.8.0 GALAXY v0.7.0–v0.8.0 records two consecutive milestones in the development of GALAXY's resident Barnes–Hut self-gravity execution family. Version v0.7.0 established the correctness-first Barnes–Hut self-gravity and GPU tree-construction foundation. It introduced deterministic resident N-body self-gravity, independent direct-force verification, Morton-ordered flat-tree representations, GPU force traversal, persistent multi-step GPU evolution, and GPU-side Barnes–Hut tree construction. Version v0.8.0 advances that foundation to the BH #2D parallel GPU tree-construction phase and adds a fail-closed framework for future real-hardware scaling evidence. It also promotes the planar Barnes–Hut N-body simulator to the default GALAXY browser experience while preserving the previous prescribed-field rotation-law instrument separately. ## v0.7.0 — Barnes–Hut Self-Gravity and GPU Tree Construction GALAXY v0.7.0 introduced a separate resident self-gravity execution family in which all bodies participate in the mutually coupled gravitational system. The implementation includes: - deterministic planar Barnes–Hut self-gravity;- classical Barnes–Hut opening control;- Plummer-style gravitational softening;- kick–drift–kick leapfrog evolution;- deterministic disc and collision fixtures;- an independent O(N²) direct-force oracle;- bounded deterministic direct-force probes;- recursive CPU Barnes–Hut reference execution;- a Morton-ordered flat f64 CPU representation;- stable spatial ordering;- deterministic topology checksums;- explicit GPU transfer records;- Vulkan/WGSL Barnes–Hut force traversal;- persistent GPU-resident position, velocity, mass and acceleration state;- multi-step GPU self-gravity evolution;- bounded full-trajectory comparison against the f64 CPU reference;- GPU-side tree construction for BH #2C;- zero host particle readbacks during BH #2C evolution steps;- zero CPU tree rebuilds during BH #2C evolution steps;- deterministic repeated-tree checksum validation. The BH #2C builder in v0.7.0 is deliberately correctness-first. Its control-heavy tree-construction stages establish the device-ownership and scientific-validation boundary required before scalable parallel tree construction and hardware benchmarking. ## v0.8.0 — Parallel GPU Barnes–Hut Construction GALAXY v0.8.0 implements BH #2D, a separate parallel GPU tree-construction path. The BH #2D builder includes: - workgroup-parallel root-bounds reduction;- parallel Morton-code generation;- eight stable 4-bit LSD radix passes;- deterministic sparse level-order cells;- parallel range construction;- parallel topology construction;- reverse-depth parallel mass and centre-of-mass aggregation;- resident workloads up to 65,536 bodies;- a separate `galaxy-bh-gpu-tree-parallel` executable and verifier. The frozen tree representation is declared as: **Ordering** ```textstable-lsd-radix-4bit-morton-code; resident body index retained for equal Morton keys``` **Layout** ```textsparse-level-order-slot=depth*N+group_start``` These declarations are validated as part of the receipt contract. ## Executable Correctness Ladder BH #2D retains the earlier implementations as executable reference surfaces rather than replacing them. The validation ladder includes: - BH #2A flat f64 CPU Barnes–Hut reference;- BH #2B2 host-built evolving GPU path;- BH #2C serialized GPU tree builder;- bounded exact direct-force probes;- full trajectory comparisons where workload limits permit them. Large workloads explicitly record bounded-oracle skips rather than hiding expensive CPU work inside GPU performance measurements. ## Real-Hardware Evidence Harness Version v0.8.0 adds a fail-closed hardware-evidence harness: ```textscripts/bench-bh2d-hardware.py``` Real evidence capture is launched through dedicated clean-environment wrappers. **POSIX** ```bashsh scripts/bench-bh2d-hardware-launch.sh``` **Windows** ```textscripts\bench-bh2d-hardware-launch.cmd``` The default resident-body sweep covers: - 512- 1,024- 2,048- 4,096- 8,192- 16,384- 32,768- 65,536 The harness executes the BH #2D verifier with `--require-hardware` and binds accepted evidence to: - one Git source revision;- one GPU adapter identity;- enumerated adapter count;- GPU backend and device type;- workload parameters;- Cargo identity;- rustc identity;- compiler and linker identities;- executable hashes;- Cargo configuration;- dependency-source tree hashes;- Python executable identity;- receipt SHA-256 hashes;- raw run-log SHA-256 hashes;- benchmark samples;- correctness invariants;- deterministic tree invariants. The hardware scaling manifest schema is: ```textgalaxy.bh2d-hardware-scaling-manifest.v1``` ## Provenance and Fail-Closed Validation The v0.8.0 evidence path was subjected to extensive adversarial review. The resulting contract includes: - clean-environment process launch;- isolated Python execution;- Linux-, macOS- and Windows-aware tool provenance;- recorded and hashed Cargo, rustc, compiler and linker binaries;- rejection of loader-injection environment variables;- rejection of Vulkan ICD, layer and driver overrides;- rejection of Cargo/Rust compiler and wrapper overrides;- source-revision binding;- explicit Git checkout binding;- disabled Git replacement objects;- raw tracked-byte validation against Git index blobs;- executable-mode validation;- Cargo configuration validation;- dependency-source tree hashing;- dependency file-type and permission-mode hashing;- registry/Git dependency hashing regardless of cache location;- strict UTF-8 JSON parsing;- rejection of duplicate JSON object keys;- rejection of non-standard NaN and infinity constants;- integer-type enforcement for execution counters;- hardware/software adapter classification checks;- adapter index/count consistency;- Rust-compatible numeric-versus-text adapter selection semantics;- allowed GPU backend validation;- frozen tree representation declarations;- topology fan-out checks;- bucket occupancy checks;- per-depth quadtree occupancy checks;- shared root-to-leaf path-capacity checks;- force-error consistency checks;- trajectory state-error consistency checks;- benchmark sample and median recomputation;- failure-manifest retention. Failed or unusable verifier runs retain cryptographically bound diagnostics. The manifest records the raw run-log path and SHA-256, and records the receipt path and SHA-256 whenever a receipt artifact exists. ## Browser N-Body Simulator Version v0.8.0 also makes the planar resident Barnes–Hut N-body simulator the default browser entrypoint. The default scene contains 768 mutually interacting bodies, with selectable resident populations from 128 to 2,048. The browser interface provides: - binary encounter preset;- rotating-disc preset;- cold-collapse preset;- pause and resume;- deterministic single-step;- seed control;- orbit and zoom controls;- Barnes–Hut tree overlay;- direct-force auditing;- glow sprites;- short trajectory-history trails;- reduced-motion handling;- hidden-tab suspension;- fixed-rate physics independent of monitor refresh rate. Physics uses a fixed 60 Hz schedule. Overload drops excess catch-up work rather than enlarging the physical timestep. The previous prescribed-field rotation-law application is preserved at: ```textrotation-lab.html``` The detailed Barnes–Hut laboratory remains available at: ```textbarnes-hut.html``` ## Validation The v0.7.0 and v0.8.0 implementation sequence was validated through repository CI and deterministic reference checks including: - Rust Barnes–Hut unit and integration tests;- CPU direct-force comparison;- CPU flat-tree verification;- GPU transfer ABI validation;- Vulkan/WGSL Barnes–Hut execution;- CUDA source/layout parity checks;- persistent-state GPU evolution;- GPU-side tree construction;- BH #2D parallel-tree contract tests;- source and dependency provenance tests;- clean-launch validation;- native Windows launcher failure propagation;- browser N-body integration tests;- refresh-rate invariance;- pause and single-step behavior;- visibility handling;- reduced-motion handling;- native CPU runtime validation;- retro CPU portability;- CPU memory-wall validation;- host-auto execution validation. Software Vulkan is used in CI as correctness and integration evidence only. ## Scientific and Performance Claim Boundary This joint record does not claim measured real-GPU performance or production-quality BH #2D scaling. In particular, it does not claim: - measured hardware GPU speedup;- production promotion of BH #2D;- multi-GPU Barnes–Hut partitioning;- distributed deterministic tree construction;- logical-u64 mutually interacting self-gravity;- hydrodynamics;- gas evolution;- star formation;- a complete evolving baryonic/dark-matter density model. A completed real-hardware BH #2D scaling manifest remains a separate evidence milestone. Future hardware-performance claims are intended to apply only to the source revision, host, toolchain, dependency tree, GPU adapter, workload and benchmark samples recorded in that manifest. ## Version Relationship This Zenodo record jointly documents the consecutive GALAXY v0.7.0 and v0.8.0 development milestones. **v0.7.0** establishes the GPU Barnes–Hut correctness and device-tree-construction baseline. **v0.8.0** extends that baseline with parallel GPU tree construction, a hardened hardware-evidence capture contract, and the default browser N-body simulator. ## Current Software Releases **GALAXY v0.7.0** https://github.com/QSOLKCB/GALAXY/releases/tag/v0.7.0 **GALAXY v0.8.0** https://github.com/QSOLKCB/GALAXY/releases/tag/v0.8.0 ## Source Repository https://github.com/QSOLKCB/GALAXY

View source

Similar papers

#computer vision Review Sep 2017

Agile Software Development Methods: Review and Analysis

This publication proposes a definition and a classification of agile software development approaches and analyses ten software development methods that can be characterized as being "agile" against the defined criterion.

P. Abrahamsson, O. Salo, Jussi Ronkainen et al. · 727 citations · ⚡54
#computer vision Jun 2008

The impact of agile practices on communication in software development

The study shows that agile practices improve both informal and formal communication, but indicates that, in larger development situations involving multiple external stakeholders, a mismatch of adequate communication mechanisms can sometimes even hinder the communication.

M. Pikkarainen, Jukka Haikara, O. Salo et al. · 401 citations · ⚡48
#machine learning Review Open access Oct 2014

Software development in startup companies: A systematic mapping study

The results indicate that software engineering work practices are chosen opportunistically, adapted and configured to provide value under the constrains imposed by the startup context.

Nicolò Paternoster, Carmine Giardino, M. Unterkalmsteiner et al. · 394 citations · ⚡54
#computer vision Review Mar 2008

Agile methods in European embedded software development organisations: a survey on the actual use and usefulness of Extreme Programming and Scrum

The results show that the embedded industry has been able to apply agile methods in its development processes and that the appreciation of the agile methods and their individual practices appears to increase once adopted and applied in practice.

O. Salo, P. Abrahamsson · 238 citations · ⚡9
#computer vision Open access Jul 2017

What happens when software developers are (un)happy

Consequences of happiness and unhappiness that are beneficial and detrimental for developers' mental well-being, the software development process, and the produced artifacts are found.

D. Graziotin, Fabian Fagerholm, Xiaofeng Wang et al. · 236 citations · ⚡13
#computer vision Open access Oct 2004

Mobile-D: an agile approach for mobile application development

The Mobile-D approach is briefly outlined here and the experiences gained from four case studies are discussed, which helped develop an agile development approach for mobile application development.

P. Abrahamsson, Antti Hanhineva, H. Hulkko et al. · 225 citations · ⚡18

Related blog posts

MIT News · Artificial Intelligence Oct 2, 2026

Documenting the tech worker movement

Writing as a participant and researcher, PhD student JS Tan SM ’22 has co-authored a new book about the rise of tech worker protests and the employer backlash that followed.

GPT-Lab Sep 23, 2026

Requirements Don’t Live in Isolation: What We’re Exploring with Req-Space

Requirements in large systems rarely exist in isolation. Their meaning depends on the wider project context - other requirements, policies, decisions, tests, and implementation details. That becomes especially important when AI is used for review, because spotting a possible conflict or gap is only the beginning. ReqSpace explores how AI, visualisation, and connected project context can help reviewers understand those findings, trace the relationships behind them, and focus on the questions that…

GPT-Lab Sep 17, 2026

Beyond Prompt Engineering: The Role of Tacit Knowledge in Software Engineering

AI is making software generation faster, but speed does not remove the need for expertise. As more work is delegated to AI, tacit knowledge may become one of the most important human advantages in software engineering. The post Beyond Prompt Engineering: The Role of Tacit Knowledge in Software Engineering appeared first on GPT-Lab.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.