fem2d Benchmark: Native vs MPI
Configuration:
MPI ranks: 4
Julia threads: 2
Date: 2025-12-17 22:31:42
L
n (grid points)
Native (s)
MPI (s)
Speedup
1
14
0.024
0.808
0.03x
2
56
0.038
0.994
0.04x
3
224
0.097
1.376
0.07x
4
896
0.671
4.684
0.14x
5
3584
3.203
12.922
0.25x
6
14336
33.246
101.618
0.33x
Speedup > 1 means MPI is faster than Native