parlab

Parallel-computing exercises from Stanford CS149 and CMU 15-418/618, adapted for local Linux and NVIDIA development.

Contents

Directory	Source	Focus
asst1	CS149	CPU threads, SIMD/ISPC, performance analysis
asst2	CS149	Task systems and dependency scheduling
asst3	CS149	CUDA SAXPY, prefix sum, circle rendering
cmu-asst4	15-418/618, Fall 2025	MPI wire routing
asst5-kernels	CS149	GPU kernel optimization

Notes

• asst1 adds missing standard C++ headers.
• asst3 uses -arch=native; rebuild when moving to a different GPU.
• Local environments and generated KMeans data, logs, and plots stay untracked.

Unofficial and unaffiliated with either course. Enrolled students should use the official starters linked above and follow their course's academic-integrity policy.