thunderx2Here we present the results of our benchmarking of the ARM cluster at Dirac. The benchmarks we are running are our HECBioSim benchmark suite here. Our benchmarks aren't too heavily tuned as these are designed to be used by our community to work out how much HPC time to ask for on this resource, so are set at a level that would be reasonable for any normal biomolecular MD simulation. More information on the benchmarks themselves can be found at the above link.

The ARM cluster at Dirac is hosted by Leicester University and is a Tier2 national machine in the UK. The cluster is made up of 4096 cores of ARMs Marvel ThunderX2 platform each node has 64 cores (2 x 32 core) and are linked by a Mellanox Bluefield interconnect.

 

Gromacs 2020.2 Single Precision

 

nodes cores 20k atoms 61k atoms 465k atoms 1.4M atoms 3M atoms
ns/day ns/day ns/day ns/day ns/day
1 64 74.0 25.0  3.1 1.35 0.633
2 128  109.3  42.8  6.2  2.64  1.255
3 192  178.8  58.6  9.5  3.77  1.867
4 256  211.4  74.9  11.9  5.11  2.453
5 320  242.3  89.4  14.3  6.24  2.928
6 384  250.2  100.0  17.4  6.88  3.458
7 448 145.8  130.4  20.5  7.76  3.844
8 512  216.7  100.5  21.3  9.35  4.275
10 640  failed  141.5  19.5  9.89  5.318
12 768 failed  153.5  23.0  11.27  6.285
14 896 failed  failed 25.1  12.49  7.24
16 1024    failed  26.5  15.19  7.945
18 1152      30.3  16.17  7.244
20 1280      32.1  16.18  9.673
22 1408      27.3  18.35 8.333
24 1536      45.9  18.39  9.037
28 1792     37.8 21.55 10.054
32 2048     36.8 14.85 11.177
36 2304     failed    
40 2560     failed    
48 3072     failed    
64 4096          

 

LAMMPS 07.8.2019

 

nodes cores 20k atoms 61k atoms 465k atoms 1.4M atoms 3M atoms
ns/day ns/day ns/day ns/day ns/day
1 64 19.4 6.3  0.80 0.26 0.13
2 128  31.6  11.0  1.52  0.55  0.23
3 192  37.3  15.2      
4 256  39.1  17.9  2.79  1.08  0.51
5 320  37.7  18.6      
6 384  34.3  19.1  3.85 1.47  0.73
7 448 37.3  18.7      
8 512  43.4  21.3  5.01  1.86  0.91
10 640  47.1  25.8  6.10  2.24  1.15
12 768 51.6  26.3  7.04  2.65  1.32
14 896 43.7  27.3      
16 960    7.71  3.15  1.64
18 1280      9.24  3.84  1.90
20 1600      10.64  4.46  2.34
22 1920      11.21  5.11 2.77
24 2240      11.85  5.67 3.16
28 2880     12.83 6.70 3.68
32 3840     13.14 7.54 4.22

 

NAMD 2.13

 

nodes cores 20k atoms 61k atoms 465k atoms 1.4M atoms 3M atoms
ns/day ns/day ns/day ns/day ns/day
1 64 24.0 9.1 1.28 0.45 0.17
2 128  37.8  16.2  2.34  0.79  0.35
3 192  53.4  18.4      
4 256  68.3 15.7  3.10  1.34  0.72
5 320  77.1 17.8      
6 384  88.3  26.2  5.23 1.61  1.01
7 448 76.2  29.3      
8 512  81.0  28.4  6.25  2.52 1.21
10 640  77.4 28.1  7.71  2.86  1.41
12 768 76.1  33.9  10.12  3.51  2.06
14 896 59.6  40.5      
16 960  66.8  30.2  12.32  4.23  2.51
18 1280      13.48  5.51  3.10
20 1600     6.69  6.45  3.53
22 1920      13.38  7.59 3.74
24 2240      10.83  9.14 4.51
28 2880     19.78 11.01 5.50
32 3840     20.44 12.52 6.99