Smoothed Particle Hydrodynamics with GRAPE and Parallel Virtual Machine

We have developed Remote-GRAPE, a subroutine library to use the special purpose computer GRAPE3A. The GRAPE-3A can e ciently calculate gravitational force between particles, and construct neighbor lists. All other calculations are performed on the host workstation (WS) which is directly connected to GRAPE. The use of GRAPE for Smoothed Particle Hydrodynamics (GRAPE-SPH) can in principle greatly speed up the calculations on WS. However the current bottleneck of GRAPE-SPH is that its performance is limited by the speed of the host WS. To solve this problem, we implement Remote-GRAPE; it allows us to run applications which use GRAPE-3A hardware on the signi cantly faster computers than the physical host WS. Thus, we can take advantage of the fast computers even though they can not physically be connected to GRAPE. The Remote-GRAPE system is implemented on the Parallel Virtual Machine (PVM). The detail of implementation is described. We analyze the performance of Remote-GRAPE and obtain the following results. 1) When RemoteGRAPE is used to calculate only gravitational forces, the overhead due to the network decrease according to the number of particles (20 % 40 % of total time). 2) To calculate gravity and neighbor lists, the overhead due to the network does not occupy a large fraction of total time but only 20 30 %, because the computation required on the slave machine (see text) is very large due to the property of GRAPE system. 3) We also compare the performance of Remote-GRAPE with the tree method. The tree method requires more time for higher degree of clustering, while the required time for Remote-GRAPE does not much depend on it. We then analyze the performance of GRAPE-SPH with Remote-GRAPE. Its performance is about 4 times faster than GRAPE-SPH with usual GRAPE for our con guration. Using Remote-GRAPE, we can calculate the GRAPE part and other parts in parallel. This parallel method leads to further speed up of our SPH code. We estimate how the performance of Remote-GRAPE depends on the con guration. We also show that the performance of Remote-GRAPE can be further improved.