PME Optimisation data
 Share
The version of the browser you are using is no longer supported. Please upgrade to a supported browser.Dismiss

View only
 
ABCDEFGHIJKLMNOPQRSTUVWXYZ
1
gmx mdrun -ntmpi 4 -ntomp 10 -pme gpu -nb gpu -pin on -nsteps 1000 -v -npme 1 -notunepme
2
New Version, kernel time (us), v100
gerrit/gromacs3
3
0.650.961.5361224489619238476815363072
4
atoms9601533300060001200024000480009600019200038400076800015360003072000
5
Scatterfailed4.6124.5556.8989.26715.07225.81845.27980.346152.92293.27581.641133.52246.5
6
Gatherfailed5.2475.3985.6655.9326.73310.76217.57130.88660.984118.01229.82447.62878.08
7
Total9.8599.95312.56315.19921.80536.5862.85111.232213.904411.28811.461581.123124.58
8
time per atom
0.010269791670.0064924983690.0041876666670.0025331666670.0018170833330.0015241666670.0013093750.0011586666670.0011140833330.0010710416670.0010565885420.0010293750.001017115885
9
10
gmx mdrun -ntmpi 4 -ntomp 10 -pme gpu -nb gpu -pin on -nsteps 1000 -v -npme 1 -notunepme
11
Master - 0c26c550ed55e12b77954dd0e8c5d956421ae501, v100
git/gromacs
12
0.650.961.5361224489619238476815363072
13
atoms9601533300060001200024000480009600019200038400076800015360003072000
14
Scatter3.5034.2616.089.78316.97831.07656.474110.47196.04379.65796.981510.33003.12
15
Gather2.6352.8683.224.1116.059.40918.12433.92960.512113.66227.6449.61916.93
16
Total6.1387.1299.313.89423.02840.48574.598144.399256.552493.311024.581959.913920.05
17
time per atom
0.006393750.0046503587740.00310.0023156666670.0019190.0016868750.0015541250.001504156250.0013362083330.0012846614580.0013340885420.0012759830730.001276057943
18
19
16 Threads per atom
gerrit/gromacs2
20
gmx mdrun -ntmpi 4 -ntomp 10 -pme gpu -nb gpu -pin on -nsteps 1000 -v -npme 1 -notunepme
21
0.650.961.5361224489619238476815363072
22
atoms9601533300060001200024000480009600019200038400076800015360003072000
23
Scatter3.3744.0665.2868.12413.9725.57745.16882.56161.81306.53641.051167.42310
24
Gather3.4673.6764.1095.037.54812.04421.13836.3273.541140.29273.94539.971073.1
25
Total6.8417.7429.39513.15421.51837.62166.306118.88235.351446.82914.991707.373383.1
26
time per atom
0.0071260416670.0050502283110.0031316666670.0021923333330.0017931666670.0015675416670.0013813750.0012383333330.0012257864580.001163593750.0011913932290.001111569010.001101269531
27
28
16 threads per atom, save and reload
gerrit/gromacs1
29
gmx mdrun -ntmpi 4 -ntomp 10 -pme gpu -nb gpu -pin on -nsteps 1000 -v -npme 1 -notunepme
30
0.650.961.5361224489619238476815363072
31
atoms9601533300060001200024000480009600019200038400076800015360003072000
32
Scatter3.42444.2155.9889.81117.24531.22955.42106.65196.68379.29774.651488.72956.6
33
Gather2.57542.8413.2084.1796.2849.67718.30934.05662.334118.28232.49457.9921.88
34
Total5.99987.0569.19613.9923.52940.90673.729140.706259.014497.571007.141946.63878.48
35
time per atom
0.0062497916670.0046027397260.0030653333330.0023316666670.001960750.0017044166670.0015360208330.00146568750.001349031250.0012957552080.0013113802080.0012673177080.001262526042
36
37
gmx mdrun -ntmpi 2 -ntomp 10 -pme gpu -nb gpu -pin on -nsteps 1000 -v -npme 1 -notunepme
gerrit/gromacs3
38
New Version, kernel time (us), v100
39
0.650.961.5361224489619238476815363072
40
atoms9601533300060001200024000480009600019200038400076800015360003072000
41
Scatterfailed5.2665.446.9049.84913.45322.3140.63589.819171.6337.18674.481317.22622.3
42
Gatherfailed4.7184.8055.6555.8546.2868.9115.83532.94264.248127.55248.56484.59957.24
43
Total9.98410.24512.55915.70319.73931.2256.47122.761235.848464.73923.041801.793579.54
44
time per atom
0.01040.006682974560.0041863333330.0026171666670.0016449166670.0013008333330.0011764583330.0012787604170.0012283750.0012102343750.0012018750.0011730403650.001165214844
45
46
nvprof --concurrent-kernels off gmx mdrun -ntmpi 1 -ntomp 10 -pme gpu -nb gpu -pin on -nsteps 1000 -v  -notunepme
gerrit/gromacs3
47
New Version, kernel time (us), v100
48
0.650.961.5361224489619238476815363072
49
atoms9601533300060001200024000480009600019200038400076800015360003072000
50
Scatter7.6257.8368.4139.87314.18725.20748.44886.074157.33297.66584.461148.32278.8
51
Gather6.5356.427.5047.2058.62211.86820.20835.41966.827130.19270.62538.181059
52
Total14.1614.25615.91717.07822.80937.07568.656121.493224.157427.85855.081686.483337.8
53
time per atom
0.014750.0092994129160.0053056666670.0028463333330.001900750.0015447916670.0014303333330.0012655520830.0011674843750.0011141927080.0011133854170.001097968750.001086523438
54
55
nvprof --concurrent-kernels off gmx mdrun -ntmpi 1 -ntomp 10 -pme gpu -nb gpu -pin on -nsteps 1000 -v -notunepme
git/gromacs
56
Master - 0c26c550ed55e12b77954dd0e8c5d956421ae501, v100
57
0.650.961.5361224489619238476815363072
58
atoms9601533300060001200024000480009600019200038400076800015360003072000
59
Scatter5.3936.3827.2339.96916.68434.24958.011109.29197.84387.2766.261517.42998.8
60
Gather4.6425.0944.7975.5077.83211.25819.59334.86561.565116.6232.24471.59947.06
61
Total10.03511.47612.0315.47624.51645.50777.604144.155259.405503.8998.51988.993945.86
62
time per atom
0.0104531250.0074859752120.004010.0025793333330.0020430.0018961250.001616750.0015016145830.0013510677080.0013119791670.0013001302080.0012949153650.001284459635
63
64
16 Threads per atom
gerrit/gromacs2
65
nvprof --concurrent-kernels off gmx mdrun -ntmpi 1 -ntomp 10 -pme gpu -nb gpu -pin on -nsteps 1000 -v -notunepme
66
0.650.961.5361224489619238476815363072
67
atoms9601533300060001200024000480009600019200038400076800015360003072000
68
Scatter5.6256.1666.4938.64113.94825.97848.70191.101158.93305.93598.831161.92305.7
69
Gather5.3936.1215.5886.2368.8612.49121.92939.35871.735138.49277.06551.961088
70
Total11.01812.28712.08114.87722.80838.46970.63130.459230.665444.42875.891713.863393.7
71
time per atom
0.011477083330.0080150032620.0040270.00247950.0019006666670.0016028750.0014714583330.0013589479170.0012013802080.001157343750.0011404817710.0011157942710.001104720052
72
73
16 threads per atom, save and reload
74
nvprof --concurrent-kernels off gmx mdrun -ntmpi 1 -ntomp 10 -pme gpu -nb gpu -pin on -nsteps 1000 -v -notunepme
75
0.650.961.5361224489619238476815363072
76
atoms9601533300060001200024000480009600019200038400076800015360003072000
77
Scatter5.246.3477.159.94916.63334.17557.649109.53195.99384.38759.111497.22971.6
78
Gather4.5915.0544.7615.4828.07910.97219.99935.52762.789119.58236.89479.11961.13
79
Total9.83111.40111.91115.43124.71245.14777.648145.057258.779503.969961976.313932.73
80
time per atom
0.0102406250.0074370515330.0039703333330.0025718333330.0020593333330.0018811250.0016176666670.0015110104170.0013478072920.0013123958330.0012968750.0012866601560.001280185547
81
82
2019.2
83
gmx mdrun -ntmpi 4 -ntomp 10 -pme gpu -nb gpu -pin on -nsteps 1000 -v -npme 1 -notunepme
84
0.650.961.5361224489619238476815363072
85
atoms9601533300060001200024000480009600019200038400076800015360003072000
86
Scatter3.4774.1996.0829.81916.05827.27851.869109.28192.81380.97798.21505.72998.8
87
Gather2.5842.8633.1214.0895.7528.28217.19733.75459.949115.13227.95448.82915.48
88
Total6.0617.0629.20313.90821.8135.5669.066143.034252.759496.11026.151954.523914.28
89
time per atom
0.0063135416670.004606653620.0030676666670.0023180.00181750.0014816666670.0014388750.00148993750.0013164531250.0012919270830.0013361328130.0012724739580.001274179688
90
91
2019.2
92
gmx mdrun -ntmpi 2 -ntomp 10 -pme gpu -nb gpu -pin on -nsteps 1000 -v -npme 1 -notunepme
93
0.650.961.5361224489619238476815363072
94
atoms9601533300060001200024000480009600019200038400076800015360003072000
95
Scatter3.514.4846.0629.3715.84927.08751.579114.99203.38398.97864.081624.73487.7
96
Gather2.632.8873.2313.9055.7558.27317.14834.52261.491118.05238.3467.7991
97
Total6.147.3719.29313.27521.60435.3668.727149.512264.871517.021102.382092.44478.7
98
time per atom
0.0063958333330.0048082191780.0030976666670.00221250.0018003333330.0014733333330.00143181250.0015574166670.0013795364580.001346406250.0014353906250.0013622395830.001457910156
99
100
2019.2
Loading...