Lattice2010

Session

Parallel 53: Algorithms and machines

18 Jun 2010, 14:30

Villasimius, Sardinia

192. AuroraScience

Luigi Scorzato (ECT*)

18/06/2010, 14:30

Algorithms and machines

I will report on the status of the AuroraScience project ( web.infn.it/aurorascience )

86. Solving the Dirac equation on QPACE

Andrea Nobile (uni regensburg)

18/06/2010, 14:50

Algorithms and machines

I present the implementation of a parallel solver currently used on QPACE for N_f=2 flavours of Clover fermions. I choose the mixed-precision Schwarz preconditioned FGCR algorithm in order to satisfy network bandwidth and latency constraints, to make efficient use of the multicore parallelism and on-chip memory, and to achieve flexibility in the choice of lattice sizes. I present benchmarks...

335. apeNET+: a 3D toroidal network enabling petaFLOPS scale Lattice QCD simulations on commodity clusters

Roberto AMMENDOLA (INFN Roma Tor Vergata)

18/06/2010, 15:10

Algorithms and machines

apeNET is a seven-years old project aimed at the acceleration of numerical simulations, mainly Lattice QCD, on commodity clusters of O(1000) nodes using a custom-designed 3D toroidal interconnect. In this paper we report on apeNET+, the new generation of our network adapters supporting wire speeds up to 34 Gbit/s per link, PCIe x8 gen 2, improved hardware RDMA support and enhanced...

321. Attaining multi-teraflops performance for QCD on GPUs

Ronald Babich (Boston University)

18/06/2010, 15:30

Algorithms and machines

In recent years, graphics processing units (GPUs) have gained prominence as general-purpose computing devices that offer an order of magnitude improvement in price/performance for a range of applications, including lattice gauge theory. In this contribution, an optimized Dirac solver for Wilson and clover-improved Wilson fermions is described. We discuss strategies for achieving high...

198. QUDA Programming for Staggered Quarks

Steven Gottlieb Gottlieb (Indiana University)

18/06/2010, 15:50

Algorithms and machines

We have been extending the QUDA GPU code developed at Boston University to include the case of improved staggered quarks. Improved staggered quarks such as asqtad and HISQ require both first and third nearest neighbor terms in the Dirac operator. We call the corresponding links fatlinks and longlinks. The fatlinks are not unitary and staggered phases are included in the links, so link...