qalgora-Q Docs Hub量子文档 ✦ Ask AI✦ 问问文档

Microsoft Azure Quantum

◐ Design-level API
This page documents qalgora-Q API design, architecture, or adaptation workflows. Code examples illustrate intended usage and are not guaranteed to run in the current reference implementation.

Azure Quantum brokers trapped-ion, superconducting, and neutral-atom hardware behind Azure identity and billing — part of the broader Microsoft Quantum program. qalgora-Q documents Azure provider targets; real submission requires the Azure Quantum SDK and an Azure subscription.

Providers

ProviderDevicesPlan
QuantinuumH2-1 / H2-2 (56q each), QCCD trapped ionStandard / Premium
IonQAria 1 (25q) / Forte 1 / Forte Enterprise 1 (36q)Pay-as-you-go
RigettiCepheus-1-108Q (108q) superconductingPay-as-you-go
PasqalFresnel (100q) neutral-atomPay-as-you-go
Quantum Circuits (QCI)Error-correcting superconducting — coming soon (private preview)

Treat this as indicative: the authoritative provider and device list is Azure's own target list. The configuration a device exposes on Azure — including qubit counts such as Quantinuum H2-1 / H2-2 — can differ from the vendor-direct / Nexus system. As of the documentation review date, public sources read this way; defer to the vendor's and Azure's official listings.

Microsoft's own hardware — Majorana

Microsoft's headline hardware is the topological qubit. Microsoft has announced Majorana 1 (Feb 2025), described as the first chip with a Topological Core (8 topological qubits) on an architecture Microsoft says is meant to scale to a million, and Majorana 2 (Microsoft Build, June 2026), which Microsoft reports gives ~1,000× better qubit reliability (parity lifetimes >20 s) by switching the material from aluminium to lead. Microsoft states a roadmap target of a scalable, fault-tolerant machine by 2029 — a roadmap goal, not a delivered capability. As of the documentation review date, these are vendor claims; specifics defer to Microsoft's official announcements.

Majorana evidence remains debated
The underlying Majorana evidence is still contested in the physics community — new critiques in Nature (reported June 2026) again questioned it. Present these as Microsoft-reported roadmap milestones, not established hardware capability. The Majorana / topological route remains at an early validation-and-debate stage and should not be equated with the cloud QPUs already open for access.

Workspace & credentials

Azure Quantum jobs run inside an Azure Quantum workspace (a resource in your subscription). Authenticate with the Azure CLI (az login); billing flows through the provider's plan (pay-as-you-go, or Quantinuum Standard/Premium).

Submission is a planned adapter workflow
Submitting to Azure Quantum is a planned adapter workflow, not a native qalgora-Q capability. The open reference build runs locally; the remote-submission lines below are commented to document the intended Azure flow, not runnable code. (bell is a predefined kernel from earlier in these docs.)
import qalgora
# Planned Azure Quantum adapter — NOT in the open reference build.
# Real submission needs the Azure Quantum SDK + an Azure subscription.
# az login first, then point at your workspace:
# qalgora.set_target("azure", machine="quantinuum.qpu.h2-1",
#                    workspace="my-quantum-workspace", location="eastus")
# job = qalgora.sample_async(bell, shots_count=2000)   # bell = predefined kernel
# print(job.status())
# print(job.get())
Reference build runs locally — not real hardware
This example targets a remote cloud/QPU, but the open reference build does not submit to a real machine — it warns and falls back to the local CPU statevector simulator, returning simulated results. Submitting to an actual device (e.g. 天衍/TianYan, IBM, IonQ, Origin) requires the vendor’s own SDK and credentials.

Resource estimation

Azure's Resource Estimator projects the physical qubits and runtime a fault-tolerant algorithm would need. As of 2026 it ships as the Python qdk.qre module (the old VS Code estimator is deprecated). qalgora-Q's estimate_resources gives the logical-circuit cost that feeds such an estimate.

res = qalgora.estimate_resources(shor_kernel, N)
print(res.gate_count, res.depth, res.num_qubits)
Specification API — not in the open reference build yet
This example shows a qalgora-Q specification API (or a third-party library) that the open reference build does not bundle today. It documents the intended interface; to run code now, use the reference build’s supported core API.
QDK, Q# & Copilot
Microsoft's native language is Q#, but the QDK now spans Q#, OpenQASM, Qiskit, Cirq, QIR, and Python, with a Copilot that writes and explains circuits. qalgora-Q runs alongside it — submitting kernels to IonQ / Quantinuum / Rigetti / Pasqal is a planned adapter workflow, or exchange via OpenQASM with existing Q# / QDK projects in the same workspace.
Logical qubits
Through qubit-virtualization, Microsoft + Atom Computing entangled 24 logical qubits (2024, a record at the time), and the partner Quantinuum Helios processor reports up to 48 logical qubits — but that figure is a benchmark/code-specific demonstration (see Quantinuum), not 48 general-purpose, long-lived, fault-tolerant application qubits. Treat both as vendor-reported milestones.

References

  • Microsoft Azure Quantum, provider global availability. learn.microsoft.com
  • Microsoft, "Majorana 1" topological-qubit announcement (Feb 2025). The underlying Majorana evidence remains contested in the physics community; renewed Nature critiques were reported in June 2026. reuters.com

Microsoft Azure Quantum

◐ 设计接口
本页描述的是 qalgora-Q 的接口设计、架构设计或适配工作流。相关代码用于说明预期用法,当前参考实现不保证可以直接运行。

Azure Quantum 把离子阱、超导和中性原子等多种硬件统一接入,共用一套 Azure 身份认证与计费体系,是更大的 Microsoft Quantum 计划的一部分。qalgora-Q 记录 Azure 服务商目标;真正提交需 Azure Quantum SDK 与 Azure 订阅。

服务商

服务商设备套餐
QuantinuumH2-1 / H2-2(各 56 比特),QCCD 离子阱标准版 / 高级版
IonQAria 1(25 比特)/ Forte 1 / Forte Enterprise 1(36 比特)按量付费
RigettiCepheus-1-108Q(108 比特)超导按量付费
PasqalFresnel(100 比特)中性原子按量付费
Quantum Circuits (QCI)纠错超导——即将上线(私有预览)

以上仅供参考:权威的服务商与设备清单应以 Azure 官方的 target list 为准。设备在 Azure 上暴露 的配置(包括 Quantinuum H2-1 / H2-2 等的比特数)与厂商直连 / Nexus 系统能力可能不同。截至文档核对日期,公开资料如此;具体以厂商与 Azure 官方公告为准。

微软自研硬件——Majorana

微软如今的旗舰硬件是拓扑量子比特。据微软宣布,Majorana 1(2025 年 2 月)是首款带拓扑核心(8 个拓扑比特)的芯片,微软称其架构面向百万比特扩展;微软报告,Majorana 2(2026 年 6 月 Microsoft Build)通过将材料由铝改为铅,宣称比特可靠性提升约 1,000 倍(宇称寿命 >20 秒)。微软将可扩展容错机的路线图目标定在 2029 年——这是路线图目标,而非已交付的能力。截至文档核对日期,以上均为厂商宣称,具体以微软官方公告为准。

马约拉纳证据仍有争议
底层马约拉纳证据在物理学界仍有争议;《Nature》上的新批评(2026 年 6 月经报道)再次提出质疑。应将其视为 Microsoft 宣称的路线图进展,而非已确证的硬件能力。马约拉纳 / 拓扑路线仍处于早期验证和争议阶段,不应与已开放访问的云端 QPU 等同。

工作区与凭证

Azure Quantum 任务在 Azure Quantum 工作区(订阅中的一个资源)内运行。使用 Azure CLI(az login)进行认证;计费通过服务商套餐结算(按量付费,或 Quantinuum 标准版/高级版)。

import qalgora
# 规划中的 Azure Quantum 适配器流程 —— 开放参考实现中不包含。
# 真正提交需 Azure Quantum SDK 与 Azure 订阅。
# 先执行 az login,再指向你的工作区:
# qalgora.set_target("azure", machine="quantinuum.qpu.h2-1",
#                    workspace="my-quantum-workspace", location="eastus")
# job = qalgora.sample_async(bell, shots_count=2000)   # bell 为预定义内核
# print(job.status())
# print(job.get())
参考实现仅本地运行 · 非真机
此示例虽指向远程云端/QPU,但开放参考实现不会真正提交到真机 —— 它会给出告警并回退到本地 CPU 态矢量模拟器,返回模拟结果,不会真正提交到天衍等真机。真正提交真机需使用对应厂商的 SDK 与凭证。

资源估算

Azure 资源估算器可预测容错算法所需的物理量子比特数与运行时。2026 年起它以 Python qdk.qre 模块形式提供(旧的 VS Code 估算器已弃用)。qalgora-Q 的 estimate_resources 给出逻辑电路开销,作为此类估算的输入。

res = qalgora.estimate_resources(shor_kernel, N)
print(res.gate_count, res.depth, res.num_qubits)
规范接口 · 参考实现暂未包含
此示例展示的是 qalgora-Q 规范中的接口(或第三方库),开放参考实现目前尚未内置,仅用于说明预期用法;如需立即运行,请使用参考实现已支持的核心 API。
QDK、Q# 与 Copilot
微软的原生语言是 Q#,但 QDK 现已覆盖 Q#、OpenQASM、Qiskit、Cirq、QIR 和 Python,并配有可编写、解释电路的 Copilot。qalgora-Q 可与其并行运行——向 IonQ / Quantinuum / Rigetti / Pasqal 提交 qalgora 内核属规划中的适配工作流,或通过 OpenQASM 与同一工作区内的现有 Q# / QDK 项目互通。
逻辑量子比特
借助量子比特虚拟化,微软与 Atom Computing 纠缠了 24 个逻辑比特(2024 年,当时的纪录);合作伙伴 Quantinuum 的 Helios 处理器报告可达 48 个逻辑比特——但该数字是特定编码与基准任务下的演示结果(参见 Quantinuum),并不等同于 48 个可任意编程、长寿命、通用容错的应用逻辑比特。两者均应视为厂商宣称的进展。

参考文献

  • Microsoft Azure Quantum, provider global availability. learn.microsoft.com
  • Microsoft, "Majorana 1" topological-qubit announcement (Feb 2025). The underlying Majorana evidence remains contested in the physics community; renewed Nature critiques were reported in June 2026. reuters.com