惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

Last Week in AI
Last Week in AI
D
DataBreaches.Net
腾讯CDC
Recent Announcements
Recent Announcements
有赞技术团队
有赞技术团队
A
About on SuperTechFans
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC
Google DeepMind News
Google DeepMind News
Microsoft Security Blog
Microsoft Security Blog
云风的 BLOG
云风的 BLOG
罗磊的独立博客
月光博客
月光博客
MyScale Blog
MyScale Blog
U
Unit 42
Martin Fowler
Martin Fowler
Stack Overflow Blog
Stack Overflow Blog
T
Tailwind CSS Blog
Engineering at Meta
Engineering at Meta
N
Netflix TechBlog - Medium
G
Google Developers Blog
博客园 - 【当耐特】
D
Docker
I
InfoQ
雷峰网
雷峰网

C++ Team Blog

Bringing Correctly Rounded Math to Production with LLVM-libc - C++ Team Blog C++ in the Age of AI: Visual Studio at CppCon 2026 - C++ Team Blog What’s New for C++ Developers in Visual Studio 2026 (18.7 – 18.10) - C++ Team Blog Microsoft at CppCon 2026 - C++ Team Blog MSVC C++23: constexpr cmath with LLVM Libc - C++ Team Blog MSVC Build Tools Preview updates – September 2026 - C++ Team Blog Clearer Build Optimization Results with GitHub Copilot - C++ Team Blog MSVC Build Tools Preview updates - August 2026 - C++ Team Blog What's New in vcpkg (Jul 2026) - C++ Team Blog Pure Virtual C++ 2026 Is a Wrap - C++ Team Blog Pure Virtual C++ 2026 Is Now Live! - C++ Team Blog C++ Dependencies Without the Headache: vcpkg + Copilot CLI - C++ Team Blog Pure Virtual C++ 2026 Is Tomorrow and On-Demand Sessions Are Now Available - C++ Team Blog Faster C++ iterative builds with GitHub Copilot - C++ Team Blog Pure Virtual C++ 2026 [Meet the Speakers, Part 3]: Modernizing C++ - C++ Team Blog MSVC Build Tools Preview updates - July 2026 - C++ Team Blog Rethinking C++ Performance: Faster Code Navigation and GitHub Copilot Tools with Whole Codebase Indexing - C++ Team Blog Pure Virtual C++ 2026 [Meet the Speakers, Part 2]: The AI-Native C++ Developer Workflow - C++ Team Blog Pure Virtual C++ 2026 [Meet the Speakers, Part 1]: Build Faster, Run Faster - C++ Team Blog What's New in vcpkg (June 2026) - C++ Team Blog Pure Virtual C++ 2026 Talks Announced - C++ Team Blog Save the Date: Pure Virtual C++ 2026 - C++ Team Blog Streamline C++ Code Intelligence Setup in Copilot CLI - C++ Team Blog Boosting Adobe Photoshop’s Performance with MSVC and SPGO - C++ Team Blog GitHub Copilot modernization for C++ is out of preview MSVC Build Tools Preview updates – June 2026 What’s New in vcpkg (May 2026) What’s New for C++ Developers in Visual Studio 2026 (18.1 – 18.6) Introducing Sample Profile Guided Optimization in MSVC NuGet PackageReference for C++ Projects in Visual Studio
Project-Specific Build Optimizations with GitHub Copilot
David Li · 2026-05-08 · via C++ Team Blog

We are excited to announce that GitHub Copilot build performance for Windows now supports project-specific builds! Available in the latest Visual Studio Insiders, you can target a single MSBuild project or CMake target instead of analyzing your entire solution. For game developers and teams working with large codebases, this eliminates the need to wait for a full-solution build when you only want to optimize one project.

Since launching GitHub Copilot build performance for Windows in Public Preview, we have received a lot of feedback from game studios working on large engines and enterprise monorepos. Rebuilding an entire solution to analyze one project is too slow and too disruptive. With project-specific builds, you can now point the agent at the project you care about and get your results faster.

A screenshot of the C++ Build Performance agent landing page in GitHub Copilot Chat

Inspired by Game Developer Feedback

This feature came directly from feedback during our Private and Public Previews. Game developers working with large C++ codebases told us that full-solution builds were an adoption blocker. In many cases, you told us that you were walking away mid-conversation and switching tabs because the trace took too long.

We kept hearing the same request: “Let me analyze a single project.”

With our project-specific analysis update, you will get the same analysis, optimizations, and validated results, all scoped to the project you choose. The agent’s intelligence has not changed. Project-specific builds scope the build step, not the analysis model. You will continue to see bottleneck identification for expensive headers, long function generation times, and costly template instantiations.

Getting Started with Project-Specific Builds

There are three ways to start a project-specific build performance session in Visual Studio. All entry points will feed into the same analysis pipeline.

Right-Click a Project in Solution Explorer

Right-click any project in Solution Explorer and select Run Build Insights > Improve Build Performance. This will open a Copilot Chat session pre-filled with the project you selected, so you will not need to type anything.

Screenshot of right clicking on the Lyra sample project and navigating to Run Build Insights > Improve Build Performance

Context Menu Entry

After selecting one of your projects in the Solution Explorer, the Build Menu will have a new entry. Select Build > Run Build Insights on Selection > Improve Build Performance to open the GitHub Copilot Chat window. This will start a session to optimize the build performance of your selected project.

Screenshot of context menu Build > Run Build Insights on Selection > Improve Build Performance

Type Directly in Copilot Chat

Open the GitHub Copilot Chat pane, select the @BuildPerfCpp responder from the agent combo box. selecting BuildPerfCpp from the agent combo box in GitHub Copilot ChatAfter selecting the responder, type:

Help me improve build performance for <MyProject>

Replace MyProject with the name of the MSBuild project or CMake target you want to optimize.

example prompt of "help me improve my build for the Lyra project"

Other Ways to Access GitHub Copilot Build Performance for Windows

Project-specific builds are the newest way to access the agent, but you can still analyze your full solution through the existing entry points.

From The Build Menu

Select Build > Run Build Insights on Solution > Improve Build Performance to open the GitHub Copilot Chat window. This will start a session to optimize the build performance of your entire solution.

From The Solution Explorer Context Menu

Right-click the solution node in Solution Explorer and select Improve Build Performance on Solution. This will open a Copilot Chat session scoped to your full solution.

From The Build Insights View

If you already have an. etl trace file open from Build Insights, click the Improve button from the diagnostics session view. This reuses data from the existing trace results and will not start a new analysis until the current changes are processed.

These entry points work well when you want to analyze your entire build. Use project-specific builds when you want to focus on one project.

A Few Things to Keep in Mind

  • The agent scopes your build but not the trace collection. vcperf and Build Insights work the same way as before.
  • If your selected project has deep dependency chains, the build may still take time. The feature reduces scope, not dependency depth.
  • For CMake, all selected targets must share a common root project. This is a limitation of the underlying CMake APIs.
  • Per-file builds are not currently supported.

Feedback Wanted!

This feature exists because of your feedback, and we want to keep improving. Whether you work on a large game engine, an enterprise monorepo, or a smaller project, we would love to hear how project-specific builds work for you. Is there a scenario we have not covered? A workflow that could be smoother? We encourage you to share your thoughts with us by filling out this survey, commenting below, through Help > Send Feedback in Visual Studio, or on X (@VisualC). Thank you for your continued support!

Learn more: Documentation for GitHub Copilot build performance for Windows | Microsoft Learn

Category

Topics

Author

David Li

I am a video game developer focused PM on the Visual C++ Team. Talk to me about video games!