Back to browse
GitHub Repository

High-performance, differentiable quantum state-vector & tensor network simulator in 100% pure JAX (no classical framework overhead). Accelerated on NVIDIA GPUs and Google Cloud TPU v6e-64/v5e VM clusters up to 37 qubits! Supported by Google's TPU Research Cloud (TRC) program.

110 starsJupyter Notebook

TPU-accelerated quantum circuit simulation in Jax

by ashitesh_12·Jul 18, 2026·2 points·0 comments

AI Analysis

●●SolidWizardryNiche Gem

Scales to 37 qubits on TPUs using pure JAX, bypassing heavy frameworks like Cirq.

Strengths
  • Eliminates framework overhead by compiling directly to monolithic XLA kernels.
  • Implements reverse-mode auto-differentiation for efficient variational algorithm training.
  • Demonstrates impressive multi-device sharding across 64-chip TPU meshes.
Weaknesses
  • Highly specialized utility restricted to researchers with access to TPU pods.
  • Lacks the extensive gate library and ecosystem tooling of established frameworks.
Category
Target Audience

Quantum computing researchers and high-performance computing engineers

Similar To

Cirq · Qiskit · PennyLane

Similar Projects

Developer Tools●●●Banger

Interactive fluid simulation in Jax using Brinkman penalization

Draw obstacles with your mouse and get gradient-based inverse design without adjoint solvers.

WizardrySolve My ProblemNiche Gem
arriemeijer
202mo ago