Stim

Stim is a tool for high performance simulation and analysis of quantum stabilizer circuits, especially those used in quantum error correction (QEC). It is typically used as a Python package, though command-line and C++ interfaces are also available.

Stim's key features include:

• Really fast simulation of large stabilizer circuits. It can analyze circuits with thousands of qubits and millions of operations quickly, producing objects that sample shots at kilohertz rates.
• Semi-automatic decoder configuration. It converts a noisy circuit into a detector error model, useful for configuring decoders. An option can suggest how hyper errors decompose into graphlike errors for matching-based decoders.
• Provides useful building blocks for working with stabilizers.

Stim prioritizes performance, aiming for absolute speed. Its design is bottom-up, providing low-level primitives for building higher-level abstractions. It guarantees backwards compatibility for its Python and command-line APIs within a major version.

Performance is achieved through techniques such as vectorized code using AVX instructions, Reference Frame Sampling for efficient bulk sampling, and an Inverted Stabilizer Tableau for faster measurements that commute with current stabilizers.

Note that Stim is limited to stabilizer operations (no non-Clifford gates) and has restricted noise/feedback support in its stim.Circuit class.