Runnable, SVA-clean SystemVerilog verification IP and simulator-labelled preference data for hardware-verification code models. Every package ships a testplan, assertions, a self-checking testbench and a regression certificate. Pay in USDC on Base — no account, no KYC.
Runnable UART verification package: testplan + SVA + self-checking testbench + regression certificate. Catches the RX last-wins overrun corner. MIT DUT.
axis_fifo verification package: ordering, tlast, full/backpressure, random stimulus. PASS + SVA-clean.
Arbiter verification package: grant invariants, round-robin fairness, priority stability, reset. PASS + SVA-clean.
Data integrity + ordering + tlast under backpressure, no loss or duplication. PASS + SVA-clean.
Toggle-based pulse CDC: two-clock self-checking scoreboard (no loss/duplication/stretch), one-cycle width + reset SVA, testplan + certificate. CDC is a premium verification corner. MIT DUT (ON1).
Simulator-verified Verilog preference pairs (iverilog PASS/FAIL labels) + reproducible generation scripts + HF dataset card. Uplift proof: QLoRA DPO on this data lifts Qwen2.5-Coder-7B testbench pass@1 from 0.0% to 17.5% (+17.5 pts, 40 simulator-verified generations); uplift.json + LoRA adapter included.
Ready-to-run formal sign-off: 4 SVA properties + environment assumptions + SymbiYosys/Jasper/VC Formal run scripts + formal top. Verilator-clean. For formal contractors who hate setup. MIT.
Ready-to-run formal sign-off for the AXI-Stream UART: SVA (AXI stability, txd idle, reset) + assumptions + SymbiYosys/Jasper/VC Formal scripts. Verilator-clean. RX-path quoted separately. MIT.
Ready-to-run formal sign-off for the AXI-Stream FIFO: SVA (AXI stability, depth bound, reset) + universal-property assumptions + SymbiYosys/Jasper/VC Formal scripts. Verilator-clean. MIT.
Ready-to-run formal sign-off for the AXI-Stream register slice / skid buffer: SVA (AXI stability, reset) + universal-property assumptions + SymbiYosys/Jasper/VC Formal scripts. Verilator-clean. MIT.
Predicts a SystemVerilog testbench's outcome (pass/fail/compile_error) BEFORE simulation to skip wasted sims. 87.5% held-out accuracy, trained on simulator-verified data. CPU-only CLI + model. MIT.
Accelerates RTL verification: parallel sim scheduling across cores + content-hash result cache (repeat runs free) + optional ML pre-filter. Wraps Icarus. 3.4x speedup, instant cached re-runs. MIT.
Instant rule-based triage of failing Icarus/Verilator logs: compile errors (file:line), assertion failures, VERDICT, failure type + likely-cause hint. No LLM/GPU. Human + JSON output. MIT.
Closes the coverage loop: runs Verilator coverage, finds uncovered points, generates targeted stimulus hints (hold branches, round-robin, reset, rare conditions) to close coverage faster. CLI + JSON. MIT.
9 simulator-verified RL tasks (write/fix a self-checking testbench across 5 DUTs). Icarus Verilog is the free objective verifier; reward gates on passing reference RTL and scales by detecting compilable mutant RTLs (anti-gaming). Includes RTL, 18 mutants, verify_tb.py + reward spec. 9 tasks at the ~$200/task market anchor for simulator-verified RL tasks. MIT.
5 contamination-controlled simulator-verified eval tasks across 5 DUTs + verifier + 18 mutants + eval harness + base leaderboard. Reward gates on reference-RTL PASS AND mutant detection (null / always-PASS / always-FAIL = 0), addressing the shared-state reward-hacking flaws audited in SWE-bench / Terminal-Bench. MIT.
LoRA r16 DPO adapter trained on the 1130 verified pairs; measured pass@1 uplift 0.0% → 17.5% (+17.5 pts). Includes uplift.json + SERVE.md (peft / vLLM / merge). Drop-in verification-testbench writer. MIT.
Safety-critical I&C example: 2-out-of-3 reactor trip voter RTL + self-checking TB + 4 detectable mutants + RLVR task + simulator verifier. For nuclear FPGA/PLD qualification (NQA-1 / IEEE 7-4.3.2) and safety-logic agent evals. MIT, ON1-authored.
Country-year nuclear panel: generation TWh, share of electricity, consumption, per-capita, YoY change. 10,192 records, 247 countries/regions, 1965–2025. OWID-derived (CC-BY, attribution), ON1 packaging. For SMR market sizing, fleet/retirement analysis, regulatory-AI grounding. jsonl + csv + stats.
Closed-loop RTL verification acceleration: coverage-director + testbench-prefilter (trained model) + sim-orchestrator + sim-log-summarizer. Find holes → target → triage → parallel run → debug. Save $97.
Everything above in one purchase: UART, FIFO, arbiter, register-slice and CDC verification IP plus the 1130-pair verified DPO dataset. Save 205 USDC vs buying separately.
0x4d3166D744Ecfab85f804bc1ed52D52618f8af18
Live once the seller endpoint is deployed and listed on agentic.market.
Works with only static hosting — no server or facilitator key required.
Contact: o@Origo.ON1 · Prices are proposals set by the owner. Receive wallet is public, receive-only.