Size = 15000
A single multi-threaded MKL DGEMM
#threads = 1  Compute Speed = 26.47 GFLOP/sec
#threads = 2  Compute Speed = 52.02 GFLOP/sec
#threads = 3  Compute Speed = 76.34 GFLOP/sec
#threads = 4  Compute Speed = 101.35 GFLOP/sec
#threads = 5  Compute Speed = 124.04 GFLOP/sec
#threads = 6  Compute Speed = 146.68 GFLOP/sec
#threads = 7  Compute Speed = 165.92 GFLOP/sec
#threads = 8  Compute Speed = 181.84 GFLOP/sec
#threads = 9  Compute Speed = 190.65 GFLOP/sec
#threads = 10  Compute Speed = 228.97 GFLOP/sec
#threads = 11  Compute Speed = 227.48 GFLOP/sec
#threads = 12  Compute Speed = 252.35 GFLOP/sec
#threads = 13  Compute Speed = 264.65 GFLOP/sec
#threads = 14  Compute Speed = 298.56 GFLOP/sec
#threads = 15  Compute Speed = 297.89 GFLOP/sec
#threads = 16  Compute Speed = 327.80 GFLOP/sec
#threads = 17  Compute Speed = 330.35 GFLOP/sec
#threads = 18  Compute Speed = 320.16 GFLOP/sec
#threads = 19  Compute Speed = 300.97 GFLOP/sec
#threads = 20  Compute Speed = 402.21 GFLOP/sec
#threads = 21  Compute Speed = 312.86 GFLOP/sec
#threads = 22  Compute Speed = 278.97 GFLOP/sec
#threads = 23  Compute Speed = 281.52 GFLOP/sec
#threads = 24  Compute Speed = 467.05 GFLOP/sec
Multiple MKL DGEMMs called in parallel
#threads = 1  Total Compute Speed = 26.12 GFLOP/sec
#threads = 2  Total Compute Speed = 52.23 GFLOP/sec
#threads = 3  Total Compute Speed = 75.72 GFLOP/sec
#threads = 4  Total Compute Speed = 100.84 GFLOP/sec
#threads = 5  Total Compute Speed = 123.53 GFLOP/sec
#threads = 6  Total Compute Speed = 148.67 GFLOP/sec
#threads = 7  Total Compute Speed = 168.51 GFLOP/sec
#threads = 8  Total Compute Speed = 192.64 GFLOP/sec
#threads = 9  Total Compute Speed = 210.70 GFLOP/sec
#threads = 10  Total Compute Speed = 231.83 GFLOP/sec
#threads = 11  Total Compute Speed = 249.04 GFLOP/sec
#threads = 12  Total Compute Speed = 272.09 GFLOP/sec
#threads = 13  Total Compute Speed = 288.00 GFLOP/sec
#threads = 14  Total Compute Speed = 311.34 GFLOP/sec
#threads = 15  Total Compute Speed = 315.72 GFLOP/sec
#threads = 16  Total Compute Speed = 333.78 GFLOP/sec
#threads = 17  Total Compute Speed = 330.56 GFLOP/sec
#threads = 18  Total Compute Speed = 346.65 GFLOP/sec
#threads = 19  Total Compute Speed = 349.41 GFLOP/sec
#threads = 20  Total Compute Speed = 330.06 GFLOP/sec
#threads = 21  Total Compute Speed = 327.15 GFLOP/sec
#threads = 22  Total Compute Speed = 311.93 GFLOP/sec
#threads = 23  Total Compute Speed = 318.43 GFLOP/sec
#threads = 24  Total Compute Speed = 324.82 GFLOP/sec
Multiple MKL DGEMMs called in nested parallel
#threads = 2  Total Compute Speed = 51.53 GFLOP/sec
#threads = 4  Total Compute Speed = 100.82 GFLOP/sec
#threads = 6  Total Compute Speed = 147.74 GFLOP/sec
#threads = 8  Total Compute Speed = 191.17 GFLOP/sec
#threads = 10  Total Compute Speed = 232.04 GFLOP/sec
#threads = 12  Total Compute Speed = 272.44 GFLOP/sec
#threads = 14  Total Compute Speed = 314.00 GFLOP/sec
#threads = 16  Total Compute Speed = 336.65 GFLOP/sec
#threads = 18  Total Compute Speed = 348.58 GFLOP/sec
#threads = 20  Total Compute Speed = 341.34 GFLOP/sec
#threads = 22  Total Compute Speed = 317.43 GFLOP/sec
#threads = 24  Total Compute Speed = 320.58 GFLOP/sec