Initialising ...
Initialising ...
Initialising ...
Initialising ...
Initialising ...
Initialising ...
Initialising ...
mixed oxidesVauchy, R.; Matsumoto, Taku; Hirooka, Shun; Uno, Hiroki*; Tamura, Tetsuya*; Arima, Tatsumi*; Inagaki, Yaohiro*; Idemitsu, Kazuya*; Nakamura, Hiroki; Machida, Masahiko; et al.
Journal of Nuclear Materials, 588, p.154786_1 - 154786_13, 2024/01
Times Cited Count：1 Percentile：72.91(Materials Science, Multidisciplinary)Matsumoto, Kazuya*; Idomura, Yasuhiro; Ina, Takuya*; Mayumi, Akie; Yamada, Susumu
Journal of Supercomputing, 75(12), p.8115 - 8146, 2019/12
Times Cited Count：2 Percentile：24.73(Computer Science, Hardware & Architecture)A communication-avoiding generalized minimum residual method (CA-GMRES) is implemented on a hybrid CPU-GPU cluster, targeted for the performance acceleration of iterative linear system solver in the gyrokinetic toroidal five-dimensional Eulerian code GT5D. In addition to the CA-GMRES, we implement and evaluate a modified variant of CA-GMRES (M-CA-GMRES) proposed in our previous study to reduce the amount of floating-point calculations. This study demonstrates that beneficial features of the CA-GMRES are in its minimum number of collective communications and its highly efficient calculations based on dense matrix-matrix operations. The performance evaluation is conducted on the Reedbush-L GPU cluster, which contains four NVIDIA Tesla P100 GPUs per compute node. The evaluation results show that the M-CA-GMRES is 1.09x, 1.22x and 1.50x faster than the CA-GMRES, the generalized conjugate residual method (GCR), and the GMRES, respectively, when 64 GPUs are used.
Nakata, Kotaro*; Hasegawa, Takuma*; Solomon, D. K.*; Miyakawa, Kazuya; Tomioka, Yuichi*; Ota, Tomoko*; Matsumoto, Takuya*; Hama, Katsuhiro; Iwatsuki, Teruki; Ono, Masahiko*; et al.
Applied Geochemistry, 104, p.60 - 70, 2019/05
Times Cited Count：9 Percentile：38.79(Geochemistry & Geophysics)no abstracts in English
Idomura, Yasuhiro; Ina, Takuya*; Mayumi, Akie; Yamada, Susumu; Matsumoto, Kazuya*; Asahi, Yuichi*; Imamura, Toshiyuki*
Proceedings of 8th Workshop on Latest Advances in Scalable Algorithms for Large-Scale Systems (ScalA 2017), p.7_1 - 7_8, 2017/11
A communication-avoiding generalized minimal residual (CA-GMRES) method is applied to the gyrokinetic toroidal five dimensional Eulerian code GT5D, and its performance is compared against the original code with a generalized conjugate residual (GCR) method on the JAEA ICEX (Haswell), the Plasma Simulator (FX100), and the Oakforest-PACS (KNL). The CA-GMRES method has 
higher arithmetic intensity than the GCR method, and thus, is suitable for future Exa-scale architectures with limited memory and network bandwidths. In the performance evaluation, it is shown that compared with the GCR solver, its computing kernels are accelerated by 
, and the cost of data reduction communication is reduced from 
to 
of the total cost at 1,280 nodes.
Am
)O
at high temperaturesMatsumoto, Taku; Arima, Tatsumi*; Inagaki, Yaohiro*; Idemitsu, Kazuya*; Kato, Masato; Morimoto, Kyoichi; Sunaoshi, Takeo*
Journal of Nuclear Science and Technology, 52(10), p.1296 - 1302, 2015/10
Times Cited Count：6 Percentile：45.92(Nuclear Science & Technology)The oxygen potentials of (PuAm)O were measured at 1873K, 1773K and 1473K by gas equilibrium method. It was shown that following the reduction of Am at the O/M ratio above 1.96, Pu was reduced at the O/M ratio below 1.96.
Matsumoto, Taku*; Arima, Tatsumi*; Inagaki, Yaohiro*; Idemitsu, Kazuya*; Kato, Masato; Uchida, Teppei
no journal, ,
no abstracts in English
Matsumoto, Taku; Arima, Tatsumi*; Inagaki, Yaohiro*; Idemitsu, Kazuya*; Kato, Masato; Morimoto, Kyoichi; Ogasawara, Masahiro*
no journal, ,
no abstracts in English
,Am)O
(x=0.03, 0.07)Matsumoto, Taku*; Arima, Tatsumi*; Inagaki, Yaohiro*; Idemitsu, Kazuya*; Kato, Masato; Morimoto, Kyoichi; Ogasawara, Masahiro*
no journal, ,
Development of minor actinide (MA) bearing MOX fuels has been advanced to reduce high level waste. In the development of such new type fuels, it is important to understand physical properties. The thermal conductivity is one of the important properties for development of the fuels, and effect of PuO and AmO content on thermal conductivity was investigated. A few thermal conductivity of PuO were reported previously, which differ widely in the measured data. The data of Am-containing MOX were also measured, but the effect of Am was unclear because of low Am content. In the present work, thermal conductivity of (Pu,Am)O was measured, and the effect of Am content was evaluated to contribute to analyze thermal conductivity of MA-bearing oxide fuels.
Matsumoto, Taku; Arima, Tatsumi*; Inagaki, Yaohiro*; Idemitsu, Kazuya*; Kato, Masato; Morimoto, Kyoichi; Tamura, Tetsuya*
no journal, ,
no abstracts in English
Am
)OMatsumoto, Taku; Arima, Tatsumi*; Inagaki, Yaohiro*; Idemitsu, Kazuya*; Kato, Masato; Morimoto, Kyoichi; Sunaoshi, Takeo*
no journal, ,
no abstracts in English
Matsumoto, Taku; Arima, Tatsumi*; Inagaki, Yaohiro*; Idemitsu, Kazuya*; Morimoto, Kyoichi; Kato, Masato; Uno, Hiroki*; Tamura, Tetsuya*
no journal, ,
The diffusion couples of (U
Pu
Am
)O and UO were annealed at 1873K in an atmosphere of 
G
that correspond to O/M = 2.00 or 1.96, and the concentration profile of each actinide elements around the interface was measured with EPMA. It was clearly found that the diffusion coefficient at O/M = 2.00 was about two orders of magnitude greater than that at O/M = 1.96. In addition, it was observed that grain boundary diffusion is so large compared to bulk one for O/M = 2.00. On the other hand, for O/M = 1.96, grain boundary and bulk diffusions of these actinides were not clearly observed.
Matsumoto, Taku; Arima, Tatsumi*; Inagaki, Yaohiro*; Idemitsu, Kazuya*; Kato, Masato; Morimoto, Kyoichi; Uno, Hiroki*; Tamura, Tetsuya*
no journal, ,
O/M ratio dependence of inter-diffusion coefficient of U and Pu at 1873 K was evaluated from MOX and UO diffusion couple. The diffusion coefficient at O/M = 2.00 was about two orders of magnitude greater than that at O/M = 1.96. In addition, it was observed that grain boundary diffusion is so large compared to bulk one for O/M = 2.00. On the other hand, GB and bulk diffusions were not clearly observed for O/M = 1.96.
Shinohara, Takenao; Kai, Tetsuya; Oikawa, Kenichi; Segawa, Mariko; Harada, Masahide; Nakatani, Takeshi; Oi, Motoki; Arai, Masatoshi; Aizawa, Kazuya; Kamiyama, Takashi*; et al.
no journal, ,
no abstracts in English
Matsumoto, Taku; Kato, Masato; Morimoto, Kyoichi; Arima, Tatsumi*; Inagaki, Yaohiro*; Idemitsu, Kazuya*; Ogasawara, Masahiro*; Sunaoshi, Takeo*
no journal, ,
no abstracts in English
Shinohara, Takenao; Kai, Tetsuya; Oikawa, Kenichi; Nakatani, Takeshi; Segawa, Mariko; Harada, Masahide; Hiroi, Kosuke; Su, Y.; Aizawa, Kazuya; Hayashida, Hirotoshi*; et al.
no journal, ,
no abstracts in English
Matsumoto, Kazuya; Asahi, Yuichi*; Ina, Takuya; Idomura, Yasuhiro
no journal, ,
We present the implementation and performance evaluation results of the plasma physics simulation code called GT5D on a GPU cluster. In this study, an iterative matrix solver, which is identified as a performance bottleneck in the code, is tuned on the GPU. The measured performance is compared with attainable performance calculated by the roofline model. Additionally, we show the implementation with direction communications between GPUs for utilizing many GPUs.
Matsumoto, Kazuya*; Idomura, Yasuhiro; Ina, Takuya*; Mayumi, Akie; Yamada, Susumu
no journal, ,
Communication avoiding (CA) Krylov methods are promising solutions for communication bottlenecks on supercomputers based on many core processors or accelerators. In this work, we implemented the CA-GMRES method on a GPU cluster, the HA-PACS, and evaluated its performance on a non-symmetric matrix solver from a nuclear CFD code. The result shows that the CA-GMRES method is significantly faster than the conventional Krylov methods such as the GMRES method and the GCR method.
Idomura, Yasuhiro; Ina, Takuya*; Mayumi, Akie; Yamada, Susumu; Matsumoto, Kazuya*; Asahi, Yuichi*; Imamura, Toshiyuki*
no journal, ,
We propose a modified communication-avoiding generalized minimal residual (CA-GMRES) method, which reduces both computation and memory access by 30% with keeping the same CA property as the original CA-GMRES method. These numerical properties, less communication and computation with higher arithmetic intensity, are promising features for future exascale machines with limited memory and network bandwidths. The modified CA-GMRES method is applied to a large scale non-symmetric matrix in an implicit solver of the gyrokinetic toroidal five dimensional Eulerian code GT5D, and its performance is estimated on the Oakforest-PACS (KNL). The numerical experiment shows that compared with the generalized conjugate residual method, computing kernels are accelerated by 1.5x, and the cost of data reduction communication is reduced from 12.5% to 1% of the total cost at 1,280 nodes.
Ozaki, Hirokazu*; Yoshimura, Kazuya; Katayose, Yuji*; Matsumoto, Takumi*; Asaoka, Yoshihiro*; Hayashi, Seiji*
no journal, ,
no abstracts in English
Ozaki, Hirokazu*; Hayashi, Seiji*; Yoshimura, Kazuya; Katayose, Yuji*; Matsumoto, Takumi*; Asaoka, Yoshihiro*
no journal, ,
no abstracts in English
