Datasets:

myhapro
/

MuxaBench

id_aa stringlengths 15 30	title stringlengths 49 73	category stringclasses 5 values	prompt stringlengths 2.16k 3.66k	system_prompt stringclasses 1 value	rubric stringlengths 4.38k 7.28k	expected_deliverables stringclasses 1 value
render_texture_codec_001	Perceptual 8 bpp block codec + 1.3 GiB atlas residency	Rendering and content pipelines	Client review is Tuesday and the texture slice of the pipeline is still the placeholder we shipped in the prototype. Cut it properly this time. Three parts, pure Zig: (1) a fixed-rate 8 bpp codec over 4x4 RGBA8 blocks — mode set, endpoint quantisation and partition tables are yours, driven by a perceptual error metric...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	rate_budget_closes_at_128_bits: A fixed-rate 8 bpp codec over 4x4 RGBA8 blocks has exactly 128 bits per block to spend. Whether the submission's mode table actually closes that budget, and whether the decoder dispatches on the same fields the encoder writes. 0 — No per-mode bit accounting anywhere, or an accounting t...	source_bundle.md, verification.json, engineering_report.md
render_restir_di_002	ReSTIR direct illumination on a from-scratch BVH tracer	Rendering and content pipelines	The lighting client wants ReSTIR behaviour they can reason about before we commit to the DXR backend, so cut the resampling layer on a CPU reference path first. It needs to be in the shared tree by Friday. Scope, pure Zig: a BVH you build yourself (SAH split, stack-based traversal, watertight triangle intersection); a...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	reservoir_weight_algebra: ReSTIR is entirely its weight algebra. Whether the reservoir update, the contribution weight and the reuse path are the real estimator or a plausible-looking blend. 0 — Any of: reservoirs combined by averaging or summing radiance/colour; the contribution weight computed as w_sum / M with no ...	source_bundle.md, verification.json, engineering_report.md
render_vector_raster_003	Analytic-AA tiled vector rasteriser to replace NanoVG	Rendering and content pipelines	We are pulling NanoVG out of the client's HMI stack — their safety case cannot carry a C dependency. The fill and stroke path has to be proven on CPU before we cut the GPU backend, and the HMI team wants pixel evidence, not adjectives. Scope, pure Zig: adaptive flattening of cubic and quadratic Beziers to a device-spa...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	analytic_coverage_correctness: Analytic coverage is where a vector rasteriser is won or lost. Whether coverage is genuinely area-based and survives the cases the fixture set targets. 0 — Any of: coverage produced by 4x/16x supersampling or an MSAA-style sample mask while the report calls it analytic; the 0.35 px hair...	source_bundle.md, verification.json, engineering_report.md
render_city_navmesh_004	10 km² city + semantic multi-level navmesh for 10,000 agents	Rendering and content pipelines	The open-world client's streaming demo is in two weeks and their AI team still cannot path through our blockout — crowds walk through walls and nothing reaches a building interior. Cut the generator and the navmesh properly, pure Zig. Scope: a procedural city over 10 km^2 from seed 0xC17Y — road hierarchy (arterial / ...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	navmesh_construction_pipeline: Whether the navmesh is genuinely constructed from the generated geometry through a walkable-surface pipeline, or the source geometry relabelled. 0 — Any of: the navmesh is the ground/floor triangle soup re-emitted with no walkable filtering; a uniform grid of cells is called a navmesh w...	source_bundle.md, verification.json, engineering_report.md
render_occlusion_foveation_005	HZB occlusion + foveated culling scored against a visibility oracle	Rendering and content pipelines	The headset client's frame budget is gone: their draw list is about 4x what it should be and frustum culling is all they have. Build the visibility stage — engine-side, over our own scene data, no platform hooks — and score it against a brute-force oracle. I need the precision and recall numbers, not a demo capture. S...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	hzb_test_is_conservative: An occlusion culler is only useful if it can never remove a visible object. Whether the depth-pyramid construction and the occludee test are conservative in the right direction, case by case. 0 — Any of: the occludee is tested with its centre depth or centre pixel rather than its nearest-dep...	source_bundle.md, verification.json, engineering_report.md
render_graph_barriers_006	Render-graph GPU abstraction with independent hazard verification	Rendering and content pipelines	We have promised the client one abstraction over DX12 and Vulkan with no driver-specific code above it. Before either backend gets written, the graph layer and its hazard tracking have to be provably right — last engine we inherited shipped a missing barrier that only reproduced on one vendor. Scope, pure Zig: a rende...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	hazard_coverage: How much of the real synchronisation problem the inserter covers: dependency kinds, granularity, queues and aliasing. 0 — Barriers inserted for read-after-write only, resource state tracked as one value for the whole resource with no subresource ranges, and no aliasing barrier at all — so the mip-cha...	source_bundle.md, verification.json, engineering_report.md
physics_gjk_epa_ccd_001	Deterministic GJK/EPA + CCD rigid-body core, 1,000-frame run	Physics and motion simulation	The sandbox client's stacks jitter apart and their 320 m/s test round goes straight through the plate. We are replacing the collision and solver core wholesale, pure Zig, and this is the slice: get it deterministic and get it stable. Scope: GJK for convex overlap and closest distance; EPA for penetration depth and con...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	gjk_implementation: Whether GJK is the real algorithm — support mapping, simplex subdistance, a principled termination test — or an overlap test wearing its name. 0 — Any of: overlap decided by separating-axis tests over face normals with no simplex or support function anywhere; distance between hulls approximated as...	source_bundle.md, verification.json, engineering_report.md
physics_xpbd_cloth_rope_002	XPBD rope and cloth with tearing, wind and buoyancy	Physics and motion simulation	Soft bodies are the next slice of the sandbox. The client wants rope and cloth in the same deterministic loop as the rigid core, with tearing that looks like tearing, and they want it to survive the wind tunnel instead of exploding on frame 30. Scope, pure Zig: an XPBD solver with compliance-based constraints — distan...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	xpbd_compliance_formulation: XPBD's whole point is that stiffness stops depending on iteration and substep count, via the compliance term and the per-constraint Lagrange multiplier. Whether the submission implements that or plain PBD relabelled. 0 — Constraints projected with a stiffness factor in [0,1] applied per i...	source_bundle.md, verification.json, engineering_report.md
physics_active_ragdoll_003	Active ragdoll: Hill-type muscles with a balance and reflex controller	Physics and motion simulation	The client's stunt system is canned animation and it shows the moment anything unexpected hits a character. Cut the active-ragdoll layer: procedural, online, no clips and no motion matching. Scope, pure Zig, three layers with the seams explicit. Physical: a 16-body skeleton with 15 joints (ball and hinge, with limits ...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	muscle_model_is_hill_type: Whether the muscular layer is a real Hill-type actuator or a joint PD controller renamed. 0 — Joint torques computed directly as a PD term on joint angle error (or a target-pose spring) with muscles appearing only as names, field comments or an unused struct; or a single linear spring per j...	source_bundle.md, verification.json, engineering_report.md
physics_fullbody_ik_004	Full-body IK for 128 characters with limits and pole vectors	Physics and motion simulation	Crowd characters in the client's build snap their elbows backwards when a handhold moves and their feet slide straight through the heightfield. We own the IK layer now — cut it so it holds 128 characters at 120 Hz without looking wrong. Scope, pure Zig: a full-body solver over a 56-bone skeleton — state per chain whet...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	joint_limits_via_swing_twist: The backwards elbow is the failure the client is complaining about. Whether limits are enforced in a rotation-representation that cannot produce it. 0 — Limits clamped as per-axis Euler angles (so the clamp depends on decomposition order and permits hyperextension), or limits applied onl...	source_bundle.md, verification.json, engineering_report.md
physics_eulerian_fluid_005	128³ Eulerian smoke solver with a multigrid Poisson projection	Physics and motion simulation	The VFX client wants smoke they can art-direct at 128^3 and a conservation report their TD will actually read. Our current solver is a Jacobi loop that mushes out every plume by frame 200. Re-cut it, pure Zig, CPU, with the Poisson solve done properly. Scope: a MAC-grid solver on 128^3 with density, velocity and tempe...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	multigrid_is_real: Whether the pressure solve is a genuine multigrid cycle with the operators named, or an iterative smoother wearing the label. 0 — A fixed number of Jacobi or Gauss-Seidel sweeps on the finest grid with no coarse levels, while the report calls it multigrid; or a residual is never computed so converg...	source_bundle.md, verification.json, engineering_report.md
physics_drone_control_006	Quadcopter flight controller and simulator over a 12-waypoint mission	Physics and motion simulation	The survey-drone client loses the line every time it gusts above about 8 m/s, and their autopilot vendor will not show us the loop. We are re-cutting the controller and the simulator that proves it, pure Zig, so the flight-test crew has something to argue with before the next campaign. Scope: quadcopter dynamics in X ...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	plant_model_fidelity: Whether the simulated vehicle is a real quadcopter model or a kinematic stand-in that makes control trivial. 0 — Any of: thrust and torques applied directly from controller output with no motor dynamics; attitude integrated from commanded rates rather than from torques and inertia; a point-mass ...	source_bundle.md, verification.json, engineering_report.md
ml_training_step_001	Fused transformer training step under a 64 MiB activation budget	Numerical and ML kernels	The client's research team wants a training loop they can audit line by line — nothing from the Python stack inside their trust boundary. This slice is one small transformer, CPU, end to end, and it has to fit their memory envelope. Model spec, fixed — do not change it: 4 layers, d_model 256, 8 heads, head_dim 32, Swi...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	spec_fidelity_and_param_count: The spec is fixed, so the parameter count is a single number a correct implementation must reach. With tied embeddings and no biases it is 4,262,144: embedding 4,096 x 256 = 1,048,576, plus 4 layers x 803,328 (attention 4 x 256 x 256 = 262,144; SwiGLU 3 x 256 x 704 = 540,672; two RMSNorm ...	source_bundle.md, verification.json, engineering_report.md
ml_quantized_inference_002	Int8 inference engine with GQA and a paged KV cache	Numerical and ML kernels	The edge client ships a sealed appliance: no Python, no cuBLAS, no vendor runtime. We need the quantised inference path proven on CPU first, with the numerics written down, before anyone talks about a GPU backend. Model spec, fixed: 12 layers, d_model 768, 12 query heads and 4 KV heads (grouped-query), head_dim 64, Sw...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	quantisation_scheme_is_specified_and_correct: Whether int8 quantisation is defined precisely enough to be reproducible, and implemented consistently between the quantiser and the kernels. 0 — Any of: weights stored as int8 with a single global scale for a whole tensor while the report claims per-channel; scales compu...	source_bundle.md, verification.json, engineering_report.md
ml_spirv_codegen_003	Runtime SPIR-V compute-shader emitter with its own validator	Numerical and ML kernels	The Vulkan backend needs its compute shaders generated at runtime from our kernel specs — we cannot ship glslang or any offline toolchain into the client's build. Cut the emitter, and cut something that can tell us when the emitter is wrong, because no driver is going to give us a useful error message. Scope, pure Zig...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	module_binary_structure: A SPIR-V module has a fixed five-word header — magic 0x07230203, version, generator magic, id bound, schema 0 — and a mandated logical layout order. Whether the emitter produces a structurally valid module. 0 — Any of: magic word absent or wrong (not 0x07230203); the header not five words; th...	source_bundle.md, verification.json, engineering_report.md
ml_flash_attention_004	Tiled causal attention, forward and backward, at 8k context	Numerical and ML kernels	Prefill at 8k context is eating our entire latency budget and the current kernel materialises the score matrix, which at this shape is 256 MiB per head. Cut the tiled kernel — forward and backward, because the client trains as well as serves. Shape, fixed: batch 1, 8 heads, sequence 8,192, head_dim 128, causal mask, b...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	online_softmax_recurrence: The streaming softmax is the kernel's core: as each key block moves the running max, the accumulated output and the running denominator must both be rescaled. 0 — Any of: the score matrix materialised per row-block and softmaxed in two passes (which is the thing being replaced); the running...	source_bundle.md, verification.json, engineering_report.md
ml_spatial_audio_005	512-voice physically modelled spatial audio, one hour without an underrun	Numerical and ML kernels	The simulator client wants physically modelled audio — not sample playback — with sources placed in 3D, and their certification people will not accept an audio thread that can underrun. 512 voices, and it has to hold for an hour. Scope, pure Zig: Karplus-Strong string voices with fractional-delay tuning and a document...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	fractional_delay_tuning: At 48 kHz, A4 = 440 Hz needs a delay line of 109.09 samples. An integer delay of 109 gives 440.367 Hz (+1.44 cents) and 110 gives 436.364 Hz (-14.4 cents), so the 1 cent target is only reachable with a real fractional-delay element. 0 — The delay line is integer-length only (so the reported f...	source_bundle.md, verification.json, engineering_report.md
sys_kv_wal_recovery_001	Crash-safe embedded KV engine: WAL recovery under torn writes	Systems infrastructure	The client's field devices lose power mid-write and their current store comes back with a corrupt index roughly one time in fifty. We are replacing the storage core, and the deliverable that matters is the recovery evidence. Scope, pure Zig: a single-file page store over a block-device abstraction you write — 4 KiB ch...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	durability_argument: Whether the ordering and detection story is airtight, or the usual hand-wave that a checksum will probably catch it. 0 — Any of: data pages written before their WAL record is durable; fsync called only at the end of the ingest or not at all; a page checksum covering less than the whole page; or t...	source_bundle.md, verification.json, engineering_report.md
sys_hft_matching_002	Deterministic matching engine at 10 M orders/s with FIFO fairness	Systems infrastructure	The venue client's risk desk will not sign off without deterministic replay, and their throughput requirement is 10 million orders per second on one core. Cut the matching engine, pure Zig. Scope: a limit order book with strict price-time priority — a fixed tick ladder for the near band with an overflow structure for ...	You are a staff systems engineer at Northgate Runtime, a five-person shop that builds from-scratch, zero-dependency subsystems in Zig for clients who cannot take third-party code into their tree. House rules, on every brief: - Target Zig 0.15.1. Zig standard library only: no external packages, no build.zig.zon depende...	order_type_semantics: The brief states the exchange semantics explicitly; each is a place implementations quietly diverge. 0 — Any of: FOK allowed to fill partially; post-only silently repriced or accepted as a crossing order instead of rejected; replace always keeping or always losing time priority regardless of the...	source_bundle.md, verification.json, engineering_report.md

End of preview. Expand in Data Studio

MuxaBench: Systems Programming Benchmark

MuxaBench is an evaluation benchmark designed to assess LLMs and coding agents on 26 staff/principal-level systems programming tasks in pure Zig 0.15.1.

Benchmark Overview

The benchmark contains 26 tasks divided into 5 core engineering domains:

Rendering and content pipelines (6 tasks): ReSTIR DI on BVH, perceptual 8 bpp texture codec, analytic-AA vector rasteriser, 10 km² city + navmesh, HZB occlusion culling.
Physics and motion simulation (6 tasks): GJK/EPA CCD rigid body, XPBD rope/cloth, active ragdoll, full-body IK, 128³ Eulerian multigrid smoke solver, quadcopter simulator.
Numerical and ML kernels (5 tasks): Fused transformer step, Int8 GQA engine with paged KV-cache, runtime SPIR-V shader generator, FlashAttention causal attention, spatial audio synthesizer.
Systems infrastructure (5 tasks): Crash-safe WAL KV engine, deterministic matching engine (10M ops/s), x86_64 microkernel slice (4-level VMM/page allocator), Raft under 30% packet loss, JIT compiler with generational GC.
Security and cryptography (4 tasks): Constant-time ML-KEM-768, ML-DSA-65 signatures, R1CS SHA-256 prover, static PE heuristic analyzer.

Evaluation Methodology (Score Against Rubric)

Each task in MuxaBench is scored against a multi-criteria rubric with discrete scoring levels (0 to 2 or 3) across ~6 explicit criteria per task. Models are evaluated on:

Strict correctness and mathematical formulation
Enforcing zero dynamic allocation in hot loops after initialization
Deterministic hashing across runs and threads
metric_provenance: verifying that reported metrics trace directly to executable test harness code

Downloads last month: -

Number of rows:

Total file size:

277 kB