昨日の続きで Scale 28 の問題を Cuda 5.0 と 5.5 で解いたときの違いを調べてます。少しだけ 5.5 の方が性能が良くなってます。
Scale 28
Cuda 5.0 : median_TEPS: 3.04957e+09
Cuda 5.5 : median_TEPS: 3.05506e+09
◯最適化問題(SDP)用 GPU 計算クラスタ
Intel Xeon + 4 GPU マシン（４台）
CPU：Xeon X5690(3.46GHz,6コア)×2
メモリ：192GB(16GB×12)
HDD：SATA500GB×2(システム、システムバックアップ)
NIC : GbE x 1 & Inifiniband(FDR) x 1
GPGPU：Tesla C2075(C2070)×4
OS：CentOS 6.3 for x86_64
============= Result ==============
SCALE: 28
edgefactor: 16
NBFS: 64
graph_generation: 55.5277900696
num_mpi_processes: 16
construction_time: 188.005583048
redistribution_time: 24.3721840382
min_time: 1.33681
firstquartile_time: 1.38533
median_time: 1.40584
thirdquartile_time: 1.4272
max_time: 1.48541
mean_time: 1.40586
stddev_time: 0.0331768
min_nedge: 4294927670
firstquartile_nedge: 4294927670
median_nedge: 4294927670
thirdquartile_nedge: 4294927670
max_nedge: 4294927670
mean_nedge: 4294927670
stddev_nedge: 0
min_TEPS: 2.89141e+09
firstquartile_TEPS: 3.00933e+09
median_TEPS: 3.05506e+09
thirdquartile_TEPS: 3.10029e+09
max_TEPS: 3.21282e+09
harmonic_mean_TEPS: 3.05501e+09
harmonic_stddev_TEPS: 9.08311e+06
min_validate: 36.0176
firstquartile_validate: 36.9006
median_validate: 37.2257
thirdquartile_validate: 37.6069
max_validate: 38.5538
mean_validate: 37.2293
stddev_validate: 0.521994
