|
|
|
|
|
|
|
|
|
|
|
| Partitioned Memory
Parallel Programming Library (PMLIB) |
|
|
|
| Based on BSP-RAMP:
Partitioned Memory Parallel Programming Framework |
|
|
|
| |
|
|
Prof.
Subodh Kumar |
Prof.
Sorav Bansal |
|
|
|
|
|
|
| Tarun Beri |
|
|
|
| Indian Institute of
Technology, Delhi |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
N-Body Simulation |
|
|
|
|
|
|
| Experimental
Configuration |
|
7 |
|
| Four 64-bit Ubuntu
Linux 8.04.2 Eight Core Machines [Intel Xeon CPU E5450 3.00 GHz with 16 GB
physical memory]; |
|
3 |
|
| Two machines have
one Tesla C1060 GPU Card each |
|
5 |
|
| Watch Dog/Kernel
Execution Timeout Disabled; mpiexec with 4 processes on different machines; |
|
|
|
| CUDA Version 3.1; OpenMPI Version 1.78;
OpenMP Version 3.0; gcc Version 4.2.4 |
|
|
|
| Equal task
partitioning among all processing elements |
|
|
|
|
| 100 iterations |
|
| N-Bodies |
Computation
Time (in seconds) |
|
| |
Serial Task |
1 host CPU Task |
1 host GPU Task |
1 host Task
CPU+GPU |
Cluster CPU Task |
Cluster GPU Task |
Cluster Task
CPU+GPU |
|
| 1000 |
2.16 |
0.40 |
24.20 |
24.41 |
0.39 |
12.23 |
12.25 |
|
| 5000 |
54.60 |
7.20 |
24.20 |
30.18 |
2.93 |
12.61 |
13.20 |
|
| 10000 |
217.00 |
28.40 |
24.40 |
48.16 |
9.60 |
13.07 |
17.35 |
|
| 20000 |
872.00 |
110.80 |
24.70 |
126.44 |
32.67 |
13.94 |
46.67 |
|
| 30000 |
1942.00 |
248.90 |
25.20 |
255.03 |
68.50 |
14.83 |
92.35 |
|
| 50000 |
5457.00 |
689.90 |
26.40 |
669.11 |
182.31 |
16.70 |
200.62 |
|
| 100000 |
2197864.00 |
2751.00 |
29.80 |
2607.66 |
739.27 |
21.92 |
694.71 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| Comparison of
Serial and PMLIB Tasks |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| Comparison of PMLIB
Tasks |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| Comparison of
Serial Task, Fastest PMLIB Task and Slowest PMLIB Task |
|
| N-Bodies |
Computation
Time (in seconds) |
|
| |
Serial Task |
1 host CPU Task |
1 host GPU Task |
1 host Task
CPU+GPU |
Cluster CPU Task |
Cluster GPU Task |
Cluster Task
CPU+GPU |
Fastest PMLIB Task |
Slowest PMLIB Task |
|
| 1000 |
2.160 |
0.400 |
24.20 |
24.41 |
0.39 |
12.23 |
12.25 |
0.393 |
24.41 |
|
| 5000 |
54.60 |
7.20 |
24.20 |
30.18 |
2.93 |
12.61 |
13.20 |
2.93 |
30.18 |
|
| 10000 |
217.00 |
28.40 |
24.40 |
48.16 |
9.60 |
13.07 |
17.35 |
9.60 |
48.16 |
|
| 20000 |
872.00 |
110.80 |
24.70 |
126.44 |
32.67 |
13.94 |
46.67 |
13.94 |
126.44 |
|
| 30000 |
1942.00 |
248.90 |
25.20 |
255.03 |
68.50 |
14.83 |
92.35 |
14.83 |
255.03 |
|
| 50000 |
5457.00 |
689.90 |
26.40 |
669.11 |
182.31 |
16.70 |
200.62 |
16.70 |
689.90 |
|
| 100000 |
2197864.00 |
2751.00 |
29.80 |
2607.66 |
739.27 |
21.92 |
694.71 |
21.92 |
2751.00 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Fastest LPC Task |
Slowest LPC Task |
|
|
|
|
|
|
|
|
|
|
| N-Bodies |
Computation
Time (in seconds) |
% Speedup of PMLIB Task over Serial Task |
|
|
|
|
|
| |
Serial Task |
Fastest PMLIB Task |
Slowest PMLIB Task |
Fastest PMLIB Task |
Slowest PMLIB Task |
|
|
|
| 1000 |
2.160 |
0.393 |
24.41 |
549.62 |
8.85 |
|
| 5000 |
54.60 |
2.93 |
30.18 |
1864.12 |
180.91 |
|
| 10000 |
217.00 |
9.60 |
48.16 |
2260.18 |
450.59 |
|
| 20000 |
872.00 |
13.94 |
126.44 |
6256.73 |
689.67 |
|
Max PMLIB Speedup over Serial Task |
|
| 30000 |
1942.00 |
14.83 |
255.03 |
13092.43 |
761.47 |
|
100253.8x |
|
| 50000 |
5457.00 |
16.70 |
689.90 |
32676.65 |
790.98 |
|
| 100000 |
2197864.00 |
21.92 |
2751.00 |
10025379.74 |
79893.28 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
PMLIB Tasks
slower than Serial Task |
PMLIB Tasks
faster than Serial Task |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
|
|
|
|
|
|
|
|
|
|
|
