COM SCI 254A: Advanced Graphics Processing Unit Microarchitecture for Artificial Intelligence and Graphics Rendering
Exploration of design principles, optimization strategies, and emerging trends in graphics processing unit (GPU) microarchitecture. Emphasis on their critical role in artificial intelligence, machine learning, and graphics rendering. Through a blend of theoretical analysis, hands-on projects, and cutting-edge research discussions, study dissects the architectural innovations driving modern GPUs, such as Nvidia's tensor cores and ray tracing. Leverages the open-source Vortex GPU platform for practical experimentation, enabling students to implement and evaluate microarchitectural concepts in Verilog. Topics include compute/graphics pipelines, memory hierarchies, parallelism, compiler optimizations, and hardware/software codesign. Students propose and execute a research project, supported by curated readings from patents, seminal papers, and industry white papers.
4 units, lecture, discussion
Loading...
| BLAISE TINE | Spring 2026 | 16 | 75.0% (12) |
|---|
A-range % and grade counts cover only each instructor’s latest recorded term, shown in the "Latest Data From" column, combining all sections for the course. A-range % is the share of recorded letter grades (A+ through F) that are A+, A, or A−. Parentheses show the A-range count over the total letter-grade count. P/NP and incompletes are excluded. Records may be incomplete.
- Available Academic Terms
- Spring 2026