Skip to main content
grasp.study
About
Explore
Pricing
Log in
Try it for free
Back to course
👋
Welcome to a Grasp lesson
Create a free account
to keep studying and create your own study path.
High-Performance AI Computing
·
Module 3
Modern GPU Architecture
GPU Architecture: SMs, CUDA Cores, and Warp Schedulers
Describe the high-level architecture of a modern GPU, including Streaming Multiprocessors (SMs), CUDA cores, and warp schedulers.
2
Understanding GPU Thread Hierarchy
Explain the GPU thread hierarchy: threads, warps, thread blocks, and grids.
3
GPU Memory Hierarchy Explained
Describe the GPU memory hierarchy: registers, shared memory, L1/L2 caches, and global memory (HBM).
4
GPU Memory Hierarchy: Bandwidth & Latency Comparison
Compare the bandwidth and latency characteristics of each level of the GPU memory hierarchy.
5
Warp Scheduling and Memory Latency Hiding
Explain how concurrent execution of warps on an SM is used to hide memory latency.
6
SM Occupancy: Latency Hiding & Resource Limits
Define SM occupancy and explain its relationship to latency hiding and resource limitations (registers, shared memory).
Previous module
CPU and Parallelism Fundamentals
Next module
CUDA Programming Fundamentals