dssim 1.2.0

Discrete-event simulation framework

DSSim — Discrete-Event Simulation Framework

DSSim is a Python framework for discrete-event simulation (DES). It advances time by jumping from one scheduled event to the next, making it well-suited for modelling systems where behavior is driven by events: hardware protocols, queuing systems, resource contention, multi-process coordination, and more.

Full documentation: majvan.github.io/DSSim

Who It Is For

DSSim targets a wide audience — not just software engineers:

• Hardware engineers — describe bus protocols, pipeline stages, and signal flows using component endpoints wired together, close to how hardware is assembled.
• Scientists and researchers — model experiments, queuing phenomena, or multi-agent systems without writing a scheduler from scratch.
• Software engineers — build process-oriented models using Python generators and coroutines with familiar async idioms.

Key Features

• Two layer profiles — LiteLayer2 for maximum throughput; PubSubLayer2 for routing-rich pub/sub with delivery tiers, condition filtering, and composable filter circuits.
• Any Python object as an event — no base class, no wrapping required.
• Timeout on every blocking call — wait, get, put, sleep all accept timeout; None is the universal "nothing happened" sentinel.
• Condition filtering — wake a process only when an event matches a predicate, a filter, a circuit of filters (AND/OR/reset), or a future's completion. No spurious wakeups in PubSubLayer2.
• Publisher-subscriber with tiers — PRE (observe), CONSUME (first-accept wins), POST_HIT, POST_MISS.
• Built-in components — DSQueue, DSResource, DSPriorityResource, DSContainer, DSState, DSTimer, and hardware models (UART, etc.).
• Probes and statistics — attach QueueStatsProbe / ResourceStatsProbe without modifying component code; custom probe types supported.
• SimPy / salabim / asyncio shims — migrate existing models without rewriting; migrated code and native DSSim code share the same event loop.
• No external dependencies — stdlib only at the core.

Quick Start

pip install dssim

from dssim import DSSimulation, LiteLayer2

sim = DSSimulation(layer2=LiteLayer2)

async def producer():
    for i in range(3):
        print(f"t={sim.time}  produce {i}")
        await sim.wait(2)

async def consumer():
    for _ in range(3):
        event = await sim.wait(3)
        print(f"t={sim.time}  tick")

sim.schedule(0, producer())
sim.schedule(0, consumer())
sim.run(until=10)

For local development from source, use editable mode:

pip install -e .

More examples in examples/lite/ and examples/pubsub/.

Choosing a Profile

Situation	Profile
Small or exploratory project	PubSubLayer2 (default)
Debugging, probes, condition filtering needed	PubSubLayer2
Maximum throughput is the primary goal	LiteLayer2
Unsure	Start with LiteLayer2; switch when you hit the first blocker

sim = DSSimulation()                   # PubSubLayer2 — default
sim = DSSimulation(layer2=LiteLayer2)  # LiteLayer2

Switching is a one-line change at construction time.

Documentation

Full documentation: majvan.github.io/DSSim

• Getting Started
• User Guide
• Feature Comparison with SimPy and salabim

License

Apache 2.0