GPU and ML systems 0/0

GPU and ML systems

GPU and TPU architecture

How GPUs and TPUs run work: streaming multiprocessors and warps, the memory hierarchy, tensor cores and systolic arrays, and the roofline that compares compute with memory bandwidth. Kernel and serving performance follow from these limits.

0 / 9 solved

Questions

0/9 solved
Question Tags Difficulty Success
1. How Much Traffic Tiling Removes
shared-memorytiling +3
Medium —
2. Room for More
occupancywarps +1
Medium —
4. Straight Line Traffic
memory-coalescinghbm
Hard —
6. Grid of Multipliers
tpusystolic-array +1
Hard —
7. Stride Bank Collision
gpushared-memory +1
Hard —
8. Chasing the Ridge Point
rooflinearithmetic-intensity +1
Hard —
9. Why the Extra Dimension
tpuici-topology +1
Insane —