Computer Architecture
流水线、乱序执行、存储层次与分支预测。
About / 王文轩
Hi, I’m Wenxuan Wang, an undergraduate student majoring in Computer Science in the ACM Honors Class at Zhiyuan College, Shanghai Jiao Tong University.
Having completed my first year of study, I am particularly interested in the evolving relationship between humans and artificial intelligence—what each can do, how their roles may change, and how advances in AI may shape our future. I enjoy turning these broad questions into concrete problems and exploring possible answers through continuous learning and hands-on practice.
My current academic interests lie primarily in AI for Science and at the intersection of computer systems and chip design, particularly hardware–software co-design. I closely follow recent developments in these fields while continuing to build the mathematical, systems, and hardware foundations needed to explore them more deeply.
Outside academics, I am passionate about sports. I enjoy running, basketball, football, and tennis. I look forward to continuing to learn, embracing new challenges, and growing into a better version of myself.
01 Fields
流水线、乱序执行、存储层次与分支预测。
从数据搬运出发理解 GEMM、Softmax 与 Attention 优化。
搜索、博弈、强化学习,以及算法在真实系统中的落地。
索引、持久化、一致性,以及可靠的数据结构实现。
模块边界、测试、性能测量与长期可维护的软件。
把实现背后的理由整理成可复现、可讨论的文字。
02 Working principles
复杂系统里,先说清楚谁读取什么、谁能修改什么,再讨论执行顺序与优化。
性能结论必须带着同一输入、同一环境和同一计时方法;无法测量时就诚实留白。
当一个实现无法被清楚解释,它通常也还没有被真正理解。文档是设计的一部分。
03 Now
04 Education
ACM Honor Class · 2025 Cohort · 2025 — Present
大一完成程序设计与数据结构课程学习,并将兴趣继续延伸到体系结构、GPU 计算与 AI。
Have a question, an idea, or a system worth taking apart?