Table 4 Performance of the hybrid MPI+OpenMP saCeSS and scalability analysis when the number of processors grows
P | Method | # proc. config. | Mean iter ±std | Mean evals ±std | Mean time ±std(s) | Speedup |
|---|---|---|---|---|---|---|
# (MPI x OpenMP) | ||||||
B1 | 10-eSS | 10 | 80 ±30 | 199214 ±44836 | 5378±1070 | - |
saCeSS | 10 (10 x 1) | 62 ±21 | 92122 ±35058 | 2753±955 | 1.95 | |
saCeSS | 10 (5 x 2) | 76 ±40 | 88186 ±46200 | 2762±1273 | 1.95 | |
20-eSS | 20 | 106 ±34 | 304972 ±88870 | 4524±942 | - | |
saCeSS | 20 (20 x 1) | 79 ±38 | 218871 ±107769 | 3125±1349 | 1.45 | |
saCeSS | 20 (10 x 2) | 86 ±43 | 147268 ±75752 | 2299±1199 | 1.97 | |
saCeSS | 20 (5 x 4) | 54 ±19 | 68907 ±27132 | 1288±380 | 3.51 | |
40-eSS | 40 | 79 ±21 | 455091 ±121875 | 3662±718 | - | |
saCeSS | 40 (40 x 1) | 51 ±19 | 274628 ±115644 | 2070±650 | 1.77 | |
saCeSS | 40 (20 x 2) | 64 ±27 | 205766 ±93753 | 1743±759 | 2.10 | |
saCeSS | 40 (10 x 4) | 58 ±30 | 111804 ±61212 | 1130±535 | 3.24 | |
saCeSS | 40 (5 x 8) | 57 ±29 | 82414 ±38357 | 1078±494 | 3.39 | |
B2 | 10-eSS | 10 | 450 ±167 | 1504503 ±541257 | 1914±714 | - |
saCeSS | 10 (10 x 1) | 846 ±982 | 1247699 ±1222378 | 1694±1677 | 1.13 | |
saCeSS | 10 (5 x 2) | 630 ±275 | 916166 ±357665 | 1298±700 | 1.47 | |
20-eSS | 20 | 355 ±130 | 2306298 ±851619 | 1465±546 | - | |
saCeSS | 20 (20 x 1) | 649 ±287 | 1901601 ±747379 | 1345±619 | 1.09 | |
saCeSS | 20 (10 x 2) | 904 ±646 | 1388211 ±854461 | 962±616 | 1.52 | |
saCeSS | 20 (5 x 4) | 609 ±244 | 925076 ±353594 | 734±320 | 2.00 | |
40-eSS | 40 | 297 ±108 | 3878818 ±1478105 | 1223±447 | - | |
saCeSS | 40 (40 x 1) | 571 ±542 | 3286363 ±2593258 | 1326±1764 | 0.92 | |
saCeSS | 40 (20 x 2) | 766 ±863 | 2185567 ±2228163 | 951±1468 | 1.29 | |
saCeSS | 40 (10 x 4) | 756 ±243 | 1309744 ±446419 | 506±180 | 2.42 | |
saCeSS | 40 (5 x 8) | 522 ±290 | 801661 ±380553 | 353±162 | 3.46 | |
B4 | 10-eSS | 10 | 99 ±121 | 2230089 ±2068300 | 750±692 | - |
saCeSS | 10 (10 x 1) | 35 ±24 | 1017956 ±728328 | 343±240 | 2.18 | |
saCeSS | 10 (5 x 2) | 177 ±269 | 1854102 ±2331070 | 916±1210 | 0.82 | |
20-eSS | 20 | 18 ±22 | 1206948 ±1177645 | 176±203 | - | |
saCeSS | 20 (20 x 1) | 11 ±8 | 819845 ±548341 | 111±93 | 1.58 | |
saCeSS | 20 (10 x 2) | 30 ±23 | 898459 ±657350 | 215±169 | 0.82 | |
saCeSS | 20 (5 x 4) | 294 ±692 | 2090044 ±3980316 | 736±1453 | 0.24 | |
40-eSS | 40 | 12 ±13 | 1811454 ±1608998 | 121±141 | - | |
saCeSS | 40 (40 x 1) | 9 ±5 | 1254545 ±603925 | 78±51 | 1.55 | |
saCeSS | 40 (20 x 2) | 26 ±26 | 1523817 ±1267964 | 159±160 | 0.76 | |
saCeSS | 40 (10 x 4) | 37 ±50 | 878008 ±1178953 | 168±238 | 0.72 | |
saCeSS | 40 (5 x 8) | 159 ±692 | 1666511 ±3980316 | 490±446 | 0.25 |