Platz(Juni 2017)
RechnerHersteller
Betreiber
Land
CPU-Cores
GPU-Cores
RmaxPFlops
EnergieeffizienzGFlops/W
HPCGTFlops
1 (1)
Sunway TaihuLight (NRCPC)
Nat. Supercomp. Center in Wuxi
China
40.960 × 260 ShenWei 26010, 1,45 GHz
–
93,016
6,05
371,16
2 (2)
Tianhe-2 (NUDT)
Nat. Supercomp. Center in Guangzhou
China
32.000 × 12 Xeon E5-2692v2, 2,2 GHz
48.000 × 57 Xeon Phi
33,862
1,9
580,11
3 (3)
Piz Daint (Cray)
Centro Svizzero di Calculo Scientifico (CSCS)
Schweiz
5320 × 12 Xeon E5-2690v3, 2,6 GHz (+2862 × 18 Xeon E5-2695v4)
5320 × 56 Nvidia Tesla P100
19,59
10,4
470
4 (69)
Gyoukou (Exascaler)
Agency for Marine-Earth Science and Technology
Japan
1250 × 16 Xeon D-1571 1,3 GHz
10.000 × 1984 Pezy SC2 (700 MHz)
19,16
14,17
k. A.
5 (4)
Titan (Cray)
Oak Ridge National Lab (ORNL)
USA
18.688 × 16 Opteron 6274, 2,2 GHz
18.688 × 14 Nvidia Tesla K20x
17,59
2,14
322,32
6 (5)
Sequoia (IBM)
Lawrence Livermore National (LLNL)
USA
98.304 × 16 Bluegene/Q, 1,6 GHz
–
17,173
2,18
330,37
7 (10 upgrade)
Trinity (Cray)
DOE/NNSA/LANL/SNL
USA
18.8.61 × 16 Xeon E5-2698v3, 2,3 GHz + 14144 × 68 Xeon Phi 7250
–
14,137
3,678
546,12
8 (6)
Cori (Cray)
NERSC
USA
9152 × 68 Xeon Phi 7250
–
14,014
3,56
355,44
9 (7)
Oakforest-PACS (Fujitsu)
JCAHPC
Japan
8208 × 68 Xeon Phi 7250
–
13,554
4,99
385,48
10 (8)
K Computer (Fujitsu)
RIKEN AICS
Japan
88.128 × 8 SPARC64 VIIIfx, 2 GHz
–
10,51
0,83
602,74