Meet or beat the C# v2-scale simulation baseline #51

New issue

Closed

opened 2026-08-06 03:56:48 +00:00 by coilyco-ops · 1 comment

coilyco-ops commented

2026-08-06 03:56:48 +00:00

Member

Problem

At the current 100x100 implementation point, the Rust simulation is reported to run slower than the historical C# implementation. Longer radar, deployment, and sustained-power proofs would amplify that regression and make their validation loop impractical.

This issue is the performance gate for the v2-scale expansion tracked by #49.

Scope

Recover and run an apples-to-apples C# reference at the matching v2 development point.
Add a reproducible Rust benchmark for the same seed, population, tick horizon, build mode, and output work.
Separate simulation cost from compilation, viewer rendering, and snapshot serialization.
Measure the major tick phases, including dispatch, pathfinding and movement, power balancing, production, mutation, and snapshot projection.
Optimize the measured bottlenecks without weakening deterministic ordering or gameplay behavior.
Add a stable regression budget derived from the measured C# baseline.

Acceptance

A Ward command runs the paired benchmark and records enough configuration to reproduce it.
The before and after report includes raw timings, ticks per second, scenario seed, population, tick count, build profile, and host context.
Rust median tick throughput meets or exceeds the C# reference under the same measured workload.
The deterministic v2-world state and existing gameplay assertions remain unchanged.
The regression check avoids flaky single-sample wall-clock assertions.
The full Ward gate and release Wasm build pass.

Blocks

#52 radar-equivalent targeting.
#53 remote coal-plant deployment.
#54 deployed-generator and power-network integration.
#55 sustained post-battery operation.

## Problem At the current 100x100 implementation point, the Rust simulation is reported to run slower than the historical C# implementation. Longer radar, deployment, and sustained-power proofs would amplify that regression and make their validation loop impractical. This issue is the performance gate for the v2-scale expansion tracked by #49. ## Scope * Recover and run an apples-to-apples C# reference at the matching v2 development point. * Add a reproducible Rust benchmark for the same seed, population, tick horizon, build mode, and output work. * Separate simulation cost from compilation, viewer rendering, and snapshot serialization. * Measure the major tick phases, including dispatch, pathfinding and movement, power balancing, production, mutation, and snapshot projection. * Optimize the measured bottlenecks without weakening deterministic ordering or gameplay behavior. * Add a stable regression budget derived from the measured C# baseline. ## Acceptance * A Ward command runs the paired benchmark and records enough configuration to reproduce it. * The before and after report includes raw timings, ticks per second, scenario seed, population, tick count, build profile, and host context. * Rust median tick throughput meets or exceeds the C# reference under the same measured workload. * The deterministic v2-world state and existing gameplay assertions remain unchanged. * The regression check avoids flaky single-sample wall-clock assertions. * The full Ward gate and release Wasm build pass. ## Blocks * #52 radar-equivalent targeting. * #53 remote coal-plant deployment. * #54 deployed-generator and power-network integration. * #55 sustained post-battery operation.

coilyco-ops referenced this issue

2026-08-06 03:56:58 +00:00

Model radar-equivalent deployment targeting at v2 scale #52

coilyco-ops added the

autonomy

headless

priority

labels

2026-08-06 03:56:58 +00:00

coilyco-ops referenced this issue

2026-08-06 03:57:11 +00:00

Produce, haul, and deploy remote coal plants #53