---
type: "post"
id: "harmonic-adds-multiclock-latch-and-clock-gating-verification-85dc21"
slug: "harmonic-adds-multiclock-latch-and-clock-gating-verification-85dc21"
title: "Harmonic adds multiclock, latch and clock-gating verification to Aristotle"
description: "Harmonic says Aristotle now verifies SystemVerilog designs with multiple clock domains, latches, and clock gating, removing a prior single-clock restriction."
retrieval_nugget: "Harmonic says Aristotle now verifies SystemVerilog designs with multiple clock domains, latches, and clock gating, removing a prior single-clock restriction. Its asynchronous FIFO examples show both proofs that hold across independent clock schedules and counterexamples for metastability and a missed wakeup. For hardware teams, this makes formal checks applicable to more realistic control and synchronization logic."
published_at: "2026-11-16"
updated_at: "2026-11-16"
record_date: "2026-11-16"
date_kind: "published_at"
topics: ["verification"]
entities: ["Harmonic","Aristotle","SystemVerilog"]
related_patterns: []
novelty_label: "new_fact"
what_changed: "Aristotle now supports multiclock designs, latches and clock gating, and can verify asynchronous FIFOs with their clocking logic modeled."
prior_refs: ["https://newruntime.com/posts/harmonic-extends-aristotle-to-formal-verification-of-c-db045f/"]
source_url: "https://www.harmonic.fun/news/multiclock-verification"
source_urls: ["https://www.harmonic.fun/news/multiclock-verification"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:harmonic-adds-multiclock-latch-and-clock-gating-verification-85dc21"
status: "published"
visuals: []
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/harmonic-adds-multiclock-latch-and-clock-gating-verification-85dc21/","markdown":"https://newruntime.com/posts/harmonic-adds-multiclock-latch-and-clock-gating-verification-85dc21.md","json":"https://newruntime.com/posts/harmonic-adds-multiclock-latch-and-clock-gating-verification-85dc21.json"}
markdown_url: "https://newruntime.com/posts/harmonic-adds-multiclock-latch-and-clock-gating-verification-85dc21.md"
---

# Harmonic adds multiclock, latch and clock-gating verification to Aristotle

## Retrieval answer

Harmonic says Aristotle now verifies SystemVerilog designs with multiple clock domains, latches, and clock gating, removing a prior single-clock restriction. Its asynchronous FIFO examples show both proofs that hold across independent clock schedules and counterexamples for metastability and a missed wakeup. For hardware teams, this makes formal checks applicable to more realistic control and synchronization logic.

## Summary

Harmonic says Aristotle now verifies SystemVerilog designs with multiple clock domains, latches, and clock gating, removing a prior single-clock restriction. Its asynchronous FIFO examples show both proofs that hold across independent clock schedules and counterexamples for metastability and a missed wakeup. For hardware teams, this makes formal checks applicable to more realistic control and synchronization logic.

## What happened

In an engineering announcement dated November 12, 2026, Harmonic says Aristotle’s SystemVerilog support has moved beyond designs with a single ungated clock. The system now supports multiple clock domains, latches, and clock gating, with clocking logic modeled in the verification. Harmonic demonstrates the change using an asynchronous FIFO, where a writer and reader operate on separate clocks and exchange data through synchronized pointers. The example uses Gray-coded pointers and two-stage synchronizers.

For its baseline FIFO, the company describes a design holding four 8-bit words. Its properties check that the FIFO does not overflow or underflow, returns the correct data, and eventually makes queued data visible to the reader. Harmonic reports that all properties were proved in about 20 minutes. The checker counts accepted writes and reads, tracks an arbitrary word, and evaluates relevant conditions on the appropriate clock: capacity on the write clock and data correctness on the read clock. The example’s assumptions include coordinated initial reset, no transfers during reset, no later resets, clocks that continue running, and a producer that keeps its request and data stable while stalled.

Harmonic then scales the example to 4096 words of 32 bits, retaining the same properties. It assigns fixed clock periods in a 2000:3 write-to-read ratio while leaving their phases independent. The announcement says Aristotle proved all seven properties, including six SystemVerilog assertions and a non-vacuity check, in about 10 minutes. These are results for this example and its stated assumptions, not a general runtime guarantee.

The announcement also gives two failure demonstrations. In one, the FIFO’s Gray-coded pointers are replaced with one-hot pointers. Harmonic models digital metastability by making each changing bit in the first synchronizer stage nondeterministically take its old or new value. Under that model, Aristotle finds a counterexample to a read-count assertion: the receiver can capture two set pointer bits during a transition and infer more queued data than exists. The company contrasts this with Gray code, where only one bit changes per increment, so the modeled receiver sees the old or new pointer.

In the clock-gating example, a latch controls a gate intended to stop the read-side clock when the FIFO appears empty. The mistake is gating the incoming write-pointer synchronizer. A wake detector does not prevent a missed wakeup, and Aristotle refutes the property that queued data eventually becomes visible. Harmonic says moving that synchronizer to the ungated root read clock repairs this example: the pointer copy can update while the read side sleeps, allowing the empty flag to reopen the gate. The company invites users to submit SystemVerilog and assertions or describe properties in English, with clock and reset assumptions stated.

## What the archive adds

New Runtime’s earlier coverage, [Harmonic’s extension of Aristotle to C and Verilog verification](https://newruntime.com/posts/harmonic-extends-aristotle-to-formal-verification-of-c-db045f/), described the move from mathematical proofs into software and hardware verification. This is a new capability announcement on that known story: Aristotle’s hardware scope now includes multiclock designs, latches, and clock gating, with concrete asynchronous FIFO checks and failure examples.

## Trend signal

This does not yet fit a pattern New Runtime tracks. Harmonic’s examples show formal verification applied to more complex hardware timing behavior, but they do not establish a broader shift in agent systems, agent runtimes, or AI software delivery. The result is best read as a product capability and a specific verification demonstration rather than evidence for a general trend.

## Why now

Harmonic frames the change as removing a restriction that existed when it launched SystemVerilog support: the compiler then handled a single ungated clock. The company’s examples explain why that boundary matters for large designs, which may use multiple clocks and asynchronous data exchange, and for power-saving clock gates often built with latches. The announcement does not give a separate market or release schedule rationale.

## Connections to other events

New Runtime’s earlier coverage of Aristotle’s C and Verilog work provides the closest connection: both announcements concern formal verification, but this one focuses on clock-domain assumptions and control behavior rather than the initial expansion into software and hardware. The connection is practical as well as chronological. Teams evaluating the earlier scope can now consider whether their designs depend on independent clocks, synchronizers, or gated clocks—cases the new announcement explicitly demonstrates.

## Why it matters

Multiclock hardware creates verification conditions that a single-clock abstraction can miss. A design may preserve its ordinary data-path behavior while failing during a pointer transition, or it may stop making progress after a clock gate prevents a synchronizer from updating. Checking safety properties such as “no overflow” is not enough if the system can leave queued data invisible indefinitely; the examples pair data and capacity checks with a liveness property for eventual visibility.

For hardware teams, the useful change is not simply that Aristotle accepts more syntax. Harmonic demonstrates checks against independent clock phase and period assumptions, an explicit digital metastability model, and a latch-based gate. Those details make the examples relevant to CDC and low-power design review. The results remain demonstrations, however: they do not establish that Aristotle will prove every production design quickly, or that a proof is sound under assumptions that omit real reset, clock, or environmental behavior.

## Limits and uncertainty

Harmonic reports proof outcomes and runtimes for its FIFO examples, but does not provide independent replication, a comparison against another verification tool, or results across a broader set of production designs. The clock and reset assumptions materially bound what the proofs establish: the example assumes coordinated initial reset, no later reset, and continuously running root clocks. Its metastability treatment is an explicit digital nondeterminism model, not a measurement of analog behavior. The announcement also does not quantify usability, compute requirements, or how proof time changes with other designs and property sets.

## Skill to build now

For a SystemVerilog asynchronous FIFO, write down the environment contract before asking a formal tool to prove it: which clocks continue running, how reset is coordinated, whether traffic can occur during reset, and what the producer must hold while stalled. Then state separate properties for capacity, underflow, data ordering, and eventual visibility, sampling each on the clock that owns the behavior. This mirrors Harmonic’s example and makes it easier to distinguish a real design failure from a mismatch between the design and its assumptions.

## Reader takeaway

Aristotle’s new multiclock support is demonstrated on more than a happy-path proof: it exposes how pointer encoding and clock-gating choices can violate safety or progress properties. Hardware teams should focus on the assumptions behind each result—especially clock behavior, reset, metastability modeling, and eventual visibility—before treating a successful proof as a guarantee about the physical system.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Verification](https://newruntime.com/topics/verification.md): Explore the verification topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Sources](https://newruntime.com/sources/): See the original sources this archive cites.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)
