Forgot your password?
Document Actions
Problem Blast w/CT Scheme
|
CODE: ZEUS-3D version 3.4 PROBLEM: Blast -- the expansion of a hot sphere of plasma into an initially uniform magnetic medium. GEOMETRY: Cartesian XYZ GRID: 32 to 128 zones in each direction centered on the origin. The ratio of neighboring zone dimensions is 1.02. The smallest zones are near the origin. ALGORITHM: van Leer advection, original CT scheme to evolve magnetic fields PRECISION: Single precision on Crays (64-bits), DOUBLE PRECISION on others. DATA: In the table below, "tused" is the number of CPU seconds used by the master thread in computing the evolution (some system and ZEUS-3D overhead is excluded). The Zone-Cycles/sec is the number of mesh zones times the number of time steps divided by tused. Cray C90 Cray UNICOS machines have a hardware performance monitor (hpm), which gives the number of floating point operations per CPU second (FLOPS) performed by a given process. The FLOPS for other machines are determined from C90 FLOPS and the ratios of the Zone-Cycles/sec. All routines were compiled with: cft77 -ez
Processors tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Speedup/Processor
GRID: 32 x 32 x 32
1 1.2420 105532 113.01 1.00 1.00 1.00
GRID: 64 x 64 x 64
1 8.4265 217767 218.64 1.00 1.00 1.00
GRID:128 x 64 x 64
1 19.774 291647 298.86 1.00 1.00 1.00
GRID: 64 x 64 x128
1 24.031 239992 240.77 1.00 1.00 1.00
SGI Power Challenge
GRID: 32 x 32 x 32
Speedup/
Processors Wall Clock tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Processor
1 4.00 3.38 38829 41.58 0.37 1.00 1.00
2 2.00 1.95 67085 71.83 0.64 1.73 0.86
3 2.00 1.94 67503 72.28 0.64 1.74 0.58
4 3.00 1.49 87812 94.03 0.83 2.26 0.57
5 3.00 1.53 85914 91.99 0.81 2.21 0.44
6 3.00 1.37 95947 102.74 0.91 2.47 0.41
7 3.00 1.32 99675 106.73 0.94 2.57 0.37
8 3.00 1.22 107493 115.10 1.02 2.77 0.35
9 3.00 1.22 107429 115.03 1.02 2.77 0.31
10 3.00 1.13 115485 123.66 1.09 2.97 0.30
11 3.00 1.10 118821 127.23 1.13 3.06 0.28
12 3.00 1.10 119618 128.08 1.13 3.08 0.26
13 3.00 1.21 108669 116.36 1.03 2.80 0.22
14 3.00 1.09 120045 128.54 1.14 3.09 0.22
15 2.00 1.13 115480 123.65 1.09 2.97 0.20
16 3.00 1.13 115846 124.04 1.10 2.98 0.19
GRID: 64 x 64 x 64
Speedup/
Processors Wall Clock tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Processor
1 71.00 57.96 31661 31.84 0.15 1.00 1.00
2 42.00 27.49 66746 67.12 0.31 2.11 1.05
3 29.00 19.32 94991 95.53 0.44 3.00 1.00
4 27.00 13.36 137369 138.15 0.63 4.34 1.08
5 24.00 11.63 157751 158.64 0.72 4.98 1.00
6 23.00 10.02 183091 184.13 0.84 5.78 0.96
7 23.00 9.19 199587 200.72 0.92 6.30 0.90
8 22.00 8.29 221354 222.61 1.02 6.99 0.87
9 17.00 7.73 237332 238.68 1.09 7.50 0.83
10 18.00 7.66 239483 240.84 1.10 7.56 0.76
11 17.00 7.11 258007 259.47 1.18 8.15 0.74
12 17.00 7.44 246784 248.18 1.13 7.79 0.65
13 18.00 6.77 271215 272.75 1.25 8.57 0.66
14 21.00 7.05 260143 261.62 1.19 8.22 0.59
15 21.00 7.08 259359 260.83 1.19 8.19 0.55
16 21.00 7.10 258386 259.85 1.19 8.16 0.51
GRID: 128 x 64 x 64
Speedup/
Processors Wall Clock tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Processor
1 197.00 176.55 32666 33.49 0.11 1.00 1.00
2 106.00 88.53 65143 66.79 0.22 1.99 1.00
3 80.00 61.81 93307 95.66 0.32 2.86 0.95
4 60.00 43.81 131636 134.95 0.45 4.03 1.01
5 54.00 35.89 160706 164.76 0.55 4.92 0.98
6 50.00 30.03 192061 196.90 0.66 5.88 0.98
7 44.00 27.38 210662 215.97 0.72 6.45 0.92
8 43.00 24.15 238783 244.80 0.82 7.31 0.91
9 41.00 22.89 251995 258.35 0.86 7.71 0.86
10 38.00 21.35 270101 276.91 0.93 8.27 0.83
11 37.00 19.51 295619 303.07 1.01 9.05 0.82
12 37.00 19.55 295020 302.46 1.01 9.03 0.75
13 42.00 18.16 317507 325.51 1.09 9.72 0.75
14 40.00 18.72 308041 315.81 1.06 9.43 0.67
15 36.00 18.30 315084 323.03 1.08 9.65 0.64
16 37.00 18.45 312567 320.45 1.07 9.57 0.60
GRID: 64 x 64 x 128
Speedup/
Processors Wall Clock tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Processor
1 205.00 183.83 31372 31.50 0.13 1.00 1.00
2 108.00 92.06 62643 62.91 0.26 2.00 1.00
3 82.00 63.28 91134 91.52 0.38 2.90 0.97
4 64.00 45.18 127656 128.19 0.53 4.07 1.02
5 56.00 37.68 153056 153.70 0.64 4.88 0.98
6 54.00 31.47 183259 184.03 0.76 5.84 0.97
7 46.00 28.45 202709 203.56 0.84 6.46 0.92
8 43.00 25.20 228881 229.84 0.95 7.30 0.91
9 41.00 23.05 250219 251.27 1.04 7.98 0.89
10 40.00 21.90 263352 264.46 1.10 8.39 0.84
11 41.00 20.27 284466 285.66 1.19 9.07 0.82
12 36.00 19.34 298157 299.41 1.24 9.50 0.79
13 37.00 17.78 324367 325.73 1.35 10.34 0.80
14 37.00 17.59 327853 329.23 1.37 10.45 0.75
15 39.00 18.08 318952 320.29 1.33 10.17 0.68
16 39.00 18.38 313845 315.16 1.31 10.00 0.63
SGI Challenge
GRID: 32 x 32 x 32
Speedup/
Processors Wall Clock tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Processor
1 17.00 15.65 8376 8.97 0.08 1.00 1.00
2 9.00 8.11 16162 17.31 0.15 1.93 0.96
3 9.00 7.43 17643 18.89 0.17 2.11 0.70
4 7.00 5.30 24730 26.48 0.23 2.95 0.74
5 7.00 5.51 23789 25.47 0.23 2.84 0.57
6 6.00 4.71 27802 29.77 0.26 3.32 0.55
7 5.00 4.38 29950 32.07 0.28 3.58 0.51
8 5.00 3.82 34305 36.73 0.33 4.10 0.51
9 5.00 3.77 34771 37.23 0.33 4.15 0.46
10 5.00 3.47 37768 40.44 0.36 4.51 0.45
11 5.00 3.15 41646 44.59 0.39 4.97 0.45
12 5.00 3.34 39185 41.96 0.37 4.68 0.39
GRID: 64 x 64 x 64
Speedup/
Processors Wall Clock tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Processor
1 205.00 194.79 9420 9.47 0.04 1.00 1.00
2 107.00 97.90 18744 18.85 0.09 1.99 0.99
3 85.00 75.23 24391 24.53 0.11 2.59 0.86
4 64.00 53.55 34265 34.46 0.16 3.64 0.91
5 58.00 48.49 37842 38.06 0.17 4.02 0.80
6 51.00 41.05 44707 44.96 0.21 4.75 0.79
7 46.00 36.42 50385 50.67 0.23 5.35 0.76
8 42.00 32.27 56872 57.19 0.26 6.04 0.75
9 40.00 30.12 60932 61.28 0.28 6.47 0.72
10 37.00 26.91 68202 68.59 0.31 7.24 0.72
11 35.00 25.10 73107 73.52 0.34 7.76 0.71
12 34.00 23.89 76825 77.26 0.35 8.15 0.68
GRID: 128 x 64 x 64
Speedup/
Processors Wall Clock tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Processor
1 585.00 571.28 10095 10.35 0.03 1.00 1.00
2 309.00 294.61 19576 20.07 0.07 1.94 0.97
3 235.00 218.89 26347 27.01 0.09 2.61 0.87
4 179.00 158.84 36307 37.22 0.12 3.60 0.90
5 158.00 138.63 41601 42.65 0.14 4.12 0.82
6 136.00 115.67 49859 51.12 0.17 4.94 0.82
7 124.00 103.34 55809 57.22 0.19 5.53 0.79
8 110.00 90.58 63671 65.28 0.22 6.31 0.79
9 102.00 82.81 69646 71.40 0.24 6.90 0.77
10 96.00 76.78 75113 77.01 0.26 7.44 0.74
11 90.00 70.62 81666 83.72 0.28 8.09 0.74
12 88.00 68.67 83983 86.10 0.29 8.32 0.69
GRID: 64 x 64 x 128
Speedup/
Processors Wall Clock tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Processor
1 588.00 574.83 10033 10.08 0.04 1.00 1.00
2 314.00 295.75 19500 19.58 0.08 1.94 0.97
3 236.00 221.74 26009 26.12 0.11 2.59 0.86
4 178.00 158.23 36447 36.60 0.15 3.63 0.91
5 157.00 137.74 41870 42.05 0.17 4.17 0.83
6 134.00 115.11 50101 50.31 0.21 4.99 0.83
7 121.00 101.26 56954 57.19 0.24 5.68 0.81
8 108.00 88.90 64872 65.14 0.27 6.47 0.81
9 100.00 81.25 70980 71.28 0.30 7.07 0.79
10 89.00 73.17 78815 79.15 0.33 7.86 0.79
11 87.00 68.15 84629 84.98 0.35 8.44 0.77
12 82.00 63.29 91116 91.50 0.38 9.08 0.76
SGI Indigo 2 Extreme
Processors tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Speedup/Processor
GRID: 32 x 32 x 32
1 17.090 7669 7.91 0.07 1.00 1.00
GRID: 64 x 64 x 64
1 201.83 9091 8.76 0.04 1.00 1.00
Convex Exemplar SPP-1000
GRID: 32 x 32 x 32
Speedup/
Processors Wall Clock tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Processor
1 18.48 13.66 9594 10.27 .09 1.00 1.00
2 10.60 6.54 20047 21.47 .19 2.09 1.04
3 6.99 4.82 27168 29.09 .26 2.83 .94
4 5.91 3.57 36762 39.36 .35 3.83 .96
5 6.12 3.73 35101 37.59 .33 3.66 .73
6 5.17 2.88 45557 48.78 .43 4.75 .79
7 5.40 3.21 40777 43.66 .39 4.25 .61
8 5.25 3.00 43717 46.81 .41 4.56 .57
GRID: 64 x 64 x 64
Speedup/
Processors Wall Clock tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Processor
1 213.90 200.13 9169 9.22 .04 1.00 1.00
2 124.01 101.36 18103 18.21 .08 1.97 .99
3 91.58 71.80 25557 25.70 .12 2.79 .93
4 73.26 53.28 34441 34.64 .16 3.76 .94
5 66.30 46.17 39748 39.97 .18 4.34 .87
6 58.31 37.81 48529 48.80 .22 5.29 .88
7 55.63 34.10 53805 54.11 .25 5.87 .84
8 51.17 30.32 60530 60.87 .28 6.60 .83
GRID: 128 x 64 x 64
Speedup/
Processors Wall Clock tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Processor
1 636.14 601.41 9589 9.83 .03 1.00 1.00
2 345.62 305.53 18876 19.35 .06 1.97 .98
3 254.59 215.05 26818 27.49 .09 2.80 .93
4 202.70 159.71 36111 37.02 .12 3.77 .94
5 173.86 136.37 42290 43.36 .15 4.41 .88
6 152.31 115.18 50070 51.33 .17 5.22 .87
7 142.38 104.95 54950 56.34 .19 5.73 .82
8 130.32 89.23 64631 66.26 .22 6.74 .84
GRID: 64 x 64 x 128
Speedup/
Processors Wall Clock tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Processor
1 673.70 630.38 9149 9.19 .04 1.00 1.00
2 360.55 319.57 18047 18.12 .08 1.97 .99
3 263.08 222.71 25895 26.00 .11 2.83 .94
4 207.31 168.46 34236 34.38 .14 3.74 .94
5 179.45 140.11 41162 41.33 .17 4.50 .90
6 159.48 121.19 47586 47.79 .20 5.20 .87
7 146.25 106.74 54032 54.26 .23 5.91 .84
8 134.77 93.27 61831 62.09 .26 6.76 .84
HP 715/80
GRID: 32 x 32 x 32
Processors tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Speedup/Processor
1 12.030 10895 11.30 0.10 1.00 1.00
Convex C3880
Processors tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Speedup/Processor
GRID: 32 x 32 x 32
1 5.4688 23967 24.86 0.22 1.00 1.00
GRID: 64 x 64 x 64
1 66.547 27574 26.28 0.12 1.00 1.00
GRID:128 x 64 x 64
1 265.11 21754 20.93 0.07 1.00 1.00
GRID: 64 x 64 x128
1 218.98 26336 24.10 0.10 1.00 1.00
IBM RS/6000 Model 530 (128MB)
GRID: 32 x 32 x 32
Processors tused(s) Zone-Cycles/sec MFLOPS C90s Speedup Speedup/Processor
1 55.17 2376 2.54 0.02 1.00 1.00
Back to ZEUS-3D Main |
The Computational Astrophysics Data Analysis Center