Measure the CPU/GPU crossover on each machine

未关闭
  • 哪里不对？
#20 0 条评论 0 个 reaction 已指派 0 人 在 GitHub 查看

还没有人认领这个 Issue。

评估

难度
4/5
预计耗时
3-5 天
新手友好度
48/100
Issue 类型
功能
描述清晰度
基本清楚
活跃度
活跃
技术栈
rust
领域
backend, performance

调研方向

Start with the shortcut: logic in src/lib.rs and trace how backend="auto" chooses CPU or GPU. Read the environment variable and setter paths first; the issue suggests adding a threshold setting before calibration. Done means the threshold can be set by environment variable or setter, and CPU is chosen on ties; calibration is a possible follow-up.

由索引模型根据 Issue 内容生成。

描述

enhancement
What you want to do

backend="auto" switches to the GPU at 14 qubits (and 20 for circuits that measure part-way), values
measured on one Ryzen 5 5500 + RX 6600 (src/lib.rs, marked shortcut:). On other machines the
best crossover may be a few qubits away, so "auto" picks the slower backend in that range.

Implementation notes

Start with the environment variable and setter, which are the cheapest. Calibration can follow.
Keep the CPU on a tie, for its 64-bit precision.

What it could look like
qsengine.calibrate()                    # times both backends, saves the crossover
qsengine.set_gpu_threshold(16)          # or set it by hand
# or the environment variable QSENGINE_GPU_MIN_QUBITS=16
Contributing
  • I would like to open a pull request for this (see CONTRIBUTING.md).
主要语言
Rust
星标
1
派生
0
平均合并
12 分钟
30 天内合并 PR
4

环境准备

  • 没有 Dockerfile 或 Docker Compose 文件
  • 没有 Pull Request 模板
  • 阅读贡献指南

从这里开始

  1. 先读完整个 Issue，再读项目的贡献指南。
  2. 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
  3. Fork 仓库，在一个分支上完成修改。
  4. 提交 Pull Request，并在描述里引用这个 Issue 编号。

kaiwave/qsengine 的其他 Issue

查看 kaiwave/qsengine 的全部 Issue

相似的 Issue

更多 Rust Issue

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。