Toshihiro Kato

GPU accelerated computing–from hype to mainstream, the rebirth of vector computing

Acceleration technologies, in particular GPUs and Cell, are receiving considerable attention in modern-day HPC. Compared to classic accelerators and traditional CPUs, these devices not only exhibit higher compute density, but also sport significant memory bandwidth and vector-like capabilities to stream data at bandwidth of 100 GB/s or more. The latter qualifies such accelerators as a rebirth of vector computing. With large-scale deployments of GPUs such as Tokyo Techs TSUBAME 1.2 supercomputer facilitating 680 GPUs in a 100-Teraflops scale supercomputer, we can demonstrate that, even under a massively parallel setting, GPUs can scale both in dense linear algebra codes as well as vector-oriented CFD codes. In both cases, however, careful algorithmic developments, especially latency hiding, are important to maximize their performance.

Numerical Simulations of Large-Scale Sediment Transport Caused by the 2011 Tohoku Earthquake Tsunami in Hirota Bay, Southern Sanriku Coast

A numerical sediment transport model (STM) was used to investigate coastal geomorphic changes that resulted from the 2011 Tohoku earthquake tsunami in Rikuzentakata City and Hirota Bay on the southern Sanriku Coast of Japan. The simulation was verified using observed inundation processes and heights, measured topographic changes and sediment deposition. Aerial video footage recorded by the Iwate Prefectural Police was also used. The results show that the numerical model was able to predict the spatial distribution and volume of erosion and deposition in Hirota Bay, as well as sediment transport processes. The effects of sediment transport on tsuneimi inundation were also investigated. Numerical results revealed that the majority of the sand dunes were eroded by the first wave, especially during the strong return flow of the receding wave. Large flows and sand dune erosions can occur elsewhere if tsunamis inundate a plain with a limited shore-normal width. These events could cause large-scale morphological changes comparable to those that occurred in Rikuzentakata City.

Large-scale, high-speed tsunami prediction for the Great Nankai Trough Earthquake on the K computer

We improved the tsunami simulation code JAGURS, which is a paralleled version of URSGA code for a large-scale, high-speed tsunami prediction in the Nankai trough, Japan. We optimized the loop kernel for velocity update and intergrid communication on a three-dimensional torus network. Linear scaling was achieved up to the full system capability of the K computer (82,944 nodes) in a strong scaling test that used 100 billion finite-difference grid points. The measured performance on the K computer was 1.2 petaflops (11.5% of peak speed). Intergrid communication was optimized for a three-nested-grid model consisting of 0.68 billion grid points. Grid spacing in the area with the finest grid (180 km × 120 km) was about 5 m. We successfully implemented a large-scale tsunami simulation using this model that ran in about 30% of real time. We believe that this is the fastest tsunami prediction achieved to date with such a large-scale model. Our code can provide high-resolution tsunami prediction for broad regions within a reasonable time to assist emergency rescue and relief operations during future devastating tsunamis comparable to the 2004 Sumatra, 2010 Chile, and 2011 Tohoku tsunamis.

Parallel-algorithm extension for tsunami and earthquake-cycle simulators for massively parallel execution on the K computer

This article presents a case study on the extension of parallel algorithms in tsunami and earthquake-cycle simulators for massively parallel execution on the K computer. We use two target applications: a tsunami-simulation program, “JAGURS,” and an earthquake-cycle program, “RSGDX.” Our optimization strategy for collective communication is to split the Message Passing Interface (MPI) communicator and perform multistage localized communication to minimize the communication frequency, transferred data size, and network congestion. Moreover, in the case of severe load imbalances, we apply cyclic distribution and extend the axes for parallelization. For each application, we conduct a performance evaluation with massively parallel execution on the K computer. It is shown that our optimized code enables JAGURS to attain a 21.8× speedup for collective communication and a 7.9× speedup for the time-step loop on 8748 nodes (69,984 cores). RSGDX attains a 4.25× speedup for collective communication and an 18.7× speedup for the time-step loop on 8192 nodes (65,536 cores).