feat(docs): elevate public presentation for interview and community impact

Unify README and GitHub Pages around an evidence-first narrative with mirrored
EN/ZH structure, dedicated highlights/interview/reference surfaces, and a
refined lightweight NVIDIA-inspired visual system. This makes project value,
validation boundaries, and technical lineage clearer for reviewers.

Also adds an OpenSpec change that codifies interview-readiness and traceable
reference expectations for project-presentation requirements.

Generated with [Devin](https://cli.devin.ai/docs)

Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>

Author: LessUp

README.md

@@ -8,14 +8,15 @@
 
 English | [简体中文](README.zh-CN.md)
 
-A compact CUDA SGEMM learning project that walks from a readable baseline kernel to Tensor Core WMMA, with cuBLAS verification and a CMake-first build.
+A CUDA SGEMM engineering notebook designed for both deep learning and interview presentation: from readable FP32 baselines to guarded Tensor Core WMMA, with cuBLAS-backed verification and explicit benchmark boundaries.
 
-## What makes it useful
+## Why this project stands out
 
-- **One optimization ladder**: naive -> tiled -> bank-conflict-free -> double-buffer -> Tensor Core.
-- **Comparable kernel interfaces**: every FP32 kernel uses the same `(A, B, C, M, K, N, stream)` launcher shape.
-- **Verification-first harness**: kernel output is checked against cuBLAS with separate tolerances for FP32 and Tensor Core paths.
-- **Learning-oriented docs**: GitHub Pages carries the full walkthrough instead of duplicating it in the README.
+- **Progressive kernel ladder**: naive -> tiled -> bank-conflict-free -> double-buffer -> Tensor Core.
+- **Evidence-first reporting**: performance claims are paired with correctness policy and scope labels.
+- **Comparable interfaces**: FP32 kernels share a unified `(A, B, C, M, K, N, stream)` launcher contract.
+- **Interview-ready narrative**: dedicated pages for project highlights, interview walkthrough, and references.
+- **Bilingual mirrored docs**: English and Chinese public pages stay aligned.
 
 ## Quick start
 
@@ -31,15 +32,26 @@ ctest --test-dir build
 
 Runtime tests and benchmarks require a CUDA-capable local machine. Hosted CI is limited to compile-time, formatting, repository-structure, OpenSpec, and Pages checks.
 
-## Start here
+## Start here (GitHub Pages)
 
 | Goal | Entry point |
 |------|-------------|
-| Use the project site | [GitHub Pages](https://lessup.github.io/sgemm-optimization/) |
-| Build and run once | [Getting Started](docs/getting-started.md) |
-| Follow the kernel ladder | [Learning Path](docs/learning-path.md) |
-| Inspect the source layout | [Architecture](docs/architecture.md) |
-| Read the normative specs | [Specifications](specs.md) |
+| Open English home | [Docs Home](https://lessup.github.io/sgemm-optimization/en/) |
+| Open Chinese home | [中文首页](https://lessup.github.io/sgemm-optimization/zh/) |
+| Build and run once | [Getting Started](https://lessup.github.io/sgemm-optimization/en/getting-started) |
+| Understand differentiation | [Project Highlights](https://lessup.github.io/sgemm-optimization/en/project-highlights) |
+| Prepare interview explanation | [Interview Playbook](https://lessup.github.io/sgemm-optimization/en/interview-playbook) |
+| Trace technical lineage | [References](https://lessup.github.io/sgemm-optimization/en/references) |
+| Read normative specs | [OpenSpec Specs](openspec/specs/) |
+
+## Validation boundary
+
+| Environment | What to trust |
+|-------------|---------------|
+| Hosted CI | Formatting, compile validity, OpenSpec structure, Pages buildability |
+| Local CUDA GPU | Runtime correctness verification and benchmark performance |
+
+This split is deliberate. CI keeps repository health; real GPU hardware validates runtime behavior and speed claims.
 
 ## Source map
 

README.zh-CN.md

@@ -8,14 +8,15 @@
 
 [English](README.md) | 简体中文
 
-这是一个紧凑的 CUDA SGEMM 学习项目：从最容易读懂的 baseline kernel 出发，逐步推进到 Tensor Core WMMA，并用 cuBLAS 做正确性对照。
+这是一个面向学习与面试展示的 CUDA SGEMM 工程化项目：从可读的 FP32 baseline kernel 演进到带保护回退的 Tensor Core WMMA，并通过 cuBLAS 对照建立可信验证。
 
-## 为什么值得看
+## 为什么它更有竞争力
 
 - **优化链条完整**：naive -> tiled -> bank-conflict-free -> double-buffer -> Tensor Core。
-- **接口保持一致**：FP32 kernel 都使用统一的 `(A, B, C, M, K, N, stream)` launcher 形态。
-- **验证先行**：所有 kernel 与 cuBLAS 对照，FP32 与 Tensor Core 使用不同容差。
-- **文档职责清晰**：README 只做仓库入口，完整学习路线放在 GitHub Pages。
+- **证据优先表达**：性能结论与正确性策略、测量范围一起呈现。
+- **接口保持一致**：FP32 kernel 使用统一 `(A, B, C, M, K, N, stream)` launcher 契约。
+- **面试友好叙事**：提供项目亮点、面试手册、参考文献三类高价值入口。
+- **中英文镜像文档**：公开页面结构保持一致，便于传播与复用。
 
 ## 快速开始
 
@@ -31,15 +32,26 @@ ctest --test-dir build
 
 运行时测试和 benchmark 需要本地 CUDA GPU。托管 CI 只覆盖编译、格式、仓库结构、OpenSpec 与 Pages 构建检查。
 
-## 从哪里开始
+## 推荐入口（GitHub Pages）
 
 | 目标 | 入口 |
 |------|------|
-| 打开项目站点 | [GitHub Pages](https://lessup.github.io/sgemm-optimization/zh/) |
-| 编译运行一次 | [快速上手](zh/docs/getting-started.md) |
-| 跟随优化路线 | [学习路径](zh/docs/learning-path.md) |
-| 查看源码结构 | [架构概览](zh/docs/architecture.md) |
-| 阅读稳定规范 | [规范索引](zh/specs.md) |
+| 打开中文首页 | [中文首页](https://lessup.github.io/sgemm-optimization/zh/) |
+| 打开英文首页 | [Docs Home](https://lessup.github.io/sgemm-optimization/en/) |
+| 编译运行一次 | [快速上手](https://lessup.github.io/sgemm-optimization/zh/getting-started) |
+| 了解项目差异化 | [项目亮点](https://lessup.github.io/sgemm-optimization/zh/project-highlights) |
+| 准备面试表达 | [面试手册](https://lessup.github.io/sgemm-optimization/zh/interview-playbook) |
+| 追溯技术来源 | [参考文献](https://lessup.github.io/sgemm-optimization/zh/references) |
+| 阅读规范源 | [OpenSpec 规范](openspec/specs/) |
+
+## 验证边界
+
+| 环境 | 可以信任什么 |
+|------|--------------|
+| 托管 CI | 格式、编译有效性、OpenSpec 结构、Pages 可构建性 |
+| 本地 CUDA GPU | 运行时正确性与 benchmark 性能 |
+
+这种拆分是刻意设计：CI 负责仓库健康，真实 GPU 负责运行时与性能结论。
 
 ## 源码地图
 
@@ -48,8 +60,7 @@ src/kernels/   CUDA SGEMM kernel 实现
 src/utils/     CUDA RAII、验证与 benchmark 工具
 src/main.cu    benchmark CLI
 tests/         基于 cuBLAS 的 Google Test 覆盖
-docs/          英文学习文档与 Pages 内容
-zh/docs/       中文学习文档与 Pages 内容
+docs/          中英文 Pages 文档（含 /en 与 /zh）
 openspec/      稳定 specs 与变更工作流
 ```
 

docs/.vitepress/config.ts

@@ -11,8 +11,13 @@ const base = rawBase
 
 export default withMermaid(defineConfig({
   base,
-  title: 'SGEMM Optimization',
-  description: 'Bilingual CUDA SGEMM optimization tutorial and reference implementation',
+  title: 'SGEMM Optimization Lab',
+  description: 'Interview-grade CUDA SGEMM engineering notebook from naive FP32 to guarded Tensor Core WMMA',
+
+  head: [
+    ['meta', { name: 'theme-color', content: '#76b900' }],
+    ['meta', { property: 'og:type', content: 'website' }],
+  ],
 
   ignoreDeadLinks: [
     // External links that VitePress can't verify at build time
@@ -24,29 +29,34 @@ export default withMermaid(defineConfig({
       label: 'English',
       lang: 'en-US',
       link: '/en/',
-      title: 'SGEMM Optimization',
-      description: 'CUDA SGEMM optimization tutorial from naive kernels to Tensor Core WMMA',
+      title: 'SGEMM Optimization Lab',
+      description: 'CUDA SGEMM project with benchmark discipline, interview storytelling, and research references',
       themeConfig: {
         nav: [
-          { text: 'Guide', link: '/en/getting-started', activeMatch: '/en/' },
-          { text: 'Architecture', link: '/en/architecture' },
-          { text: 'Learning Path', link: '/en/learning-path' },
+          { text: 'Home', link: '/en/', activeMatch: '^/en/$' },
+          { text: 'Quick Start', link: '/en/getting-started' },
+          { text: 'Highlights', link: '/en/project-highlights' },
+          { text: 'Interview', link: '/en/interview-playbook' },
           { text: 'Kernels', link: '/en/kernel-naive', activeMatch: '/en/kernel-' },
-          { text: 'Reference', link: '/en/cuda-memory-cheatsheet', activeMatch: '/en/' },
+          { text: 'Benchmark', link: '/en/benchmark-results' },
+          { text: 'References', link: '/en/references' },
         ],
         sidebar: {
           '/en/': [
             {
-              text: 'Getting Started',
+              text: 'Project Brief',
               items: [
-                { text: 'Introduction', link: '/en/getting-started' },
-                { text: 'Learning Path', link: '/en/learning-path' },
+                { text: 'Home', link: '/en/' },
+                { text: 'Getting Started', link: '/en/getting-started' },
+                { text: 'Project Highlights', link: '/en/project-highlights' },
+                { text: 'Interview Playbook', link: '/en/interview-playbook' },
                 { text: 'Architecture', link: '/en/architecture' },
               ],
             },
             {
-              text: 'Kernel Optimizations',
+              text: 'Kernel Ladder',
               items: [
+                { text: 'Learning Path', link: '/en/learning-path' },
                 { text: 'Naive Kernel', link: '/en/kernel-naive' },
                 { text: 'Tiled Kernel', link: '/en/kernel-tiled' },
                 { text: 'Bank Conflict Free', link: '/en/kernel-bank-free' },
@@ -55,12 +65,18 @@ export default withMermaid(defineConfig({
               ],
             },
             {
-              text: 'Reference',
+              text: 'Performance & Ops',
               items: [
-                { text: 'CUDA Memory Cheatsheet', link: '/en/cuda-memory-cheatsheet' },
+                { text: 'Benchmark Results', link: '/en/benchmark-results' },
                 { text: 'Optimization Playbook', link: '/en/optimization-playbook' },
                 { text: 'Performance Casebook', link: '/en/performance-casebook' },
-                { text: 'Benchmark Results', link: '/en/benchmark-results' },
+                { text: 'CUDA Memory Cheatsheet', link: '/en/cuda-memory-cheatsheet' },
+              ],
+            },
+            {
+              text: 'Research References',
+              items: [
+                { text: 'Papers & Repositories', link: '/en/references' },
               ],
             },
           ],
@@ -71,29 +87,34 @@ export default withMermaid(defineConfig({
       label: '简体中文',
       lang: 'zh-CN',
       link: '/zh/',
-      title: 'SGEMM 优化',
-      description: 'CUDA SGEMM 优化教程，从朴素内核到 Tensor Core WMMA',
+      title: 'SGEMM 优化实验室',
+      description: '面向面试展示与社区传播的 CUDA SGEMM 项目文档：从朴素内核到 Tensor Core WMMA',
       themeConfig: {
         nav: [
-          { text: '指南', link: '/zh/getting-started', activeMatch: '/zh/' },
-          { text: '架构', link: '/zh/architecture' },
-          { text: '学习路径', link: '/zh/learning-path' },
-          { text: '内核', link: '/zh/kernel-naive', activeMatch: '/zh/kernel-' },
-          { text: '参考', link: '/zh/cuda-memory-cheatsheet', activeMatch: '/zh/' },
+          { text: '首页', link: '/zh/', activeMatch: '^/zh/$' },
+          { text: '快速开始', link: '/zh/getting-started' },
+          { text: '项目亮点', link: '/zh/project-highlights' },
+          { text: '面试手册', link: '/zh/interview-playbook' },
+          { text: '内核阶梯', link: '/zh/kernel-naive', activeMatch: '/zh/kernel-' },
+          { text: '基准验证', link: '/zh/benchmark-results' },
+          { text: '参考文献', link: '/zh/references' },
         ],
         sidebar: {
           '/zh/': [
             {
-              text: '入门指南',
+              text: '项目总览',
               items: [
-                { text: '入门介绍', link: '/zh/getting-started' },
-                { text: '学习路径', link: '/zh/learning-path' },
+                { text: '首页', link: '/zh/' },
+                { text: '快速上手', link: '/zh/getting-started' },
+                { text: '项目亮点', link: '/zh/project-highlights' },
+                { text: '面试手册', link: '/zh/interview-playbook' },
                 { text: '架构概述', link: '/zh/architecture' },
               ],
             },
             {
-              text: '内核优化',
+              text: '内核阶梯',
               items: [
+                { text: '学习路径', link: '/zh/learning-path' },
                 { text: '朴素内核', link: '/zh/kernel-naive' },
                 { text: '分块内核', link: '/zh/kernel-tiled' },
                 { text: '消除 Bank Conflict', link: '/zh/kernel-bank-free' },
@@ -102,12 +123,18 @@ export default withMermaid(defineConfig({
               ],
             },
             {
-              text: '参考资料',
+              text: '性能与工程',
               items: [
-                { text: 'CUDA 内存速查表', link: '/zh/cuda-memory-cheatsheet' },
+                { text: 'Benchmark 结果', link: '/zh/benchmark-results' },
                 { text: '优化手册', link: '/zh/optimization-playbook' },
                 { text: '性能案例集', link: '/zh/performance-casebook' },
-                { text: '基准测试结果', link: '/zh/benchmark-results' },
+                { text: 'CUDA 内存速查表', link: '/zh/cuda-memory-cheatsheet' },
+              ],
+            },
+            {
+              text: '论文与引用',
+              items: [
+                { text: '论文与仓库索引', link: '/zh/references' },
               ],
             },
           ],