Add challenge 75: Sparse Matrix-Dense Matrix Multiplication (SpMM) (Medium)

[claude]

Summary

• Adds challenge 75: Sparse Matrix-Dense Matrix Multiplication (SpMM) at medium difficulty
• Given a sparse matrix A (M×N, ~65% zeros) and a dense matrix B (N×K), compute dense output C = A × B (M×K)
• Extends challenge 18 (SpMV) to the matrix-matrix case, requiring solvers to think about 2D work distribution, shared memory tiling for B, and load-imbalance across sparse rows
• Validated: all functional and performance tests pass on NVIDIA Tesla T4

Test plan

• generate_example_test matches HTML example (3×4 × 4×2 → 3×2 output)
• generate_functional_test has 9 cases covering edge cases (1×1, K=1), zero matrix, identity, powers-of-2, non-powers-of-2, negatives, and realistic sizes
• generate_performance_test uses M=4,096, N=2,048, K=512 (~35% density), fits 5× in 16 GB
• All 6 starter files present and lint-clean
• pre-commit run --all-files passes
• CUDA solution validated via run_challenge.py --action submit — all tests passed

🤖 Generated with Claude Code