How Long to Read Scalable Parallel Programming for High Performance Seismic Simulation on Petascale Heterogeneous Supercomputers

How Long Does it Take to Read Scalable Parallel Programming for High Performance Seismic Simulation on Petascale Heterogeneous Supercomputers?

It takes the average reader 1 hour and 51 minutes to read Scalable Parallel Programming for High Performance Seismic Simulation on Petascale Heterogeneous Supercomputers by Jun Zhou

Assuming a reading speed of 250 words per minute. Learn more

Description

The 1994 Northridge earthquake in Los Angeles, California, killed 57 people, injured over 8,700 and caused an estimated $20 billion in damage. Petascale simulations are needed in California and elsewhere to provide society with a better understanding of the rupture and wave dynamics of the largest earthquakes at shaking frequencies required to engineer safe structures. As the heterogeneous supercomputing infrastructures are becoming more common, numerical developments in earthquake system research are particularly challenged by the dependence on the accelerator elements to enable "the Big One" simulations with higher frequency and finer resolution. Reducing time to solution and power consumption are two primary focus area today for the enabling technology of fault rupture dynamics and seismic wave propagation in realistic 3D models of the crust's heterogeneous structure. This dissertation presents scalable parallel programming techniques for high performance seismic simulation running on petascale heterogeneous supercomputers. A real world earthquake simulation code, AWP-ODC, one of the most advanced earthquake codes to date, was chosen as the base code in this research, and the testbed is based on Titan at Oak Ridge National Laboratory, the world's largest hetergeneous supercomputer. The research work is primarily related to architecture study, computation performance tuning and software system scalability. An earthquake simulation workflow has also been developed to support the efficient production sets of simulations. The highlights of the technical development are an aggressive performance optimization focusing on data locality and a notable data communication model that hides the data communication latency. This development results in the optimal computation efficiency and throughput for the 13-point stencil code on heterogeneous systems, which can be extended to general high-order stencil codes. Started from scratch, the hybrid CPU/GPU version of AWP-ODC code is ready now for real world petascale earthquake simulations. This GPU-based code has demonstrated excellent weak scaling up to the full Titan scale and achieved 2.3 PetaFLOPs sustained computation performance in single precision. The production simulation demonstrated the first 0-10Hz deterministic rough fault simulation. Using the accelerated AWP-ODC, Southern California Earthquake Center (SCEC) has recently created the physics-based probablistic seismic hazard analysis model of the Los Angeles region, CyberShake 14.2, as of the time of the dissertation writing. The tensor-valued wavefield code based on this GPU research has dramatically reduced time-to-solution, making a statewide hazard model a goal reachable with existing heterogeneous supercomputers.

How long is Scalable Parallel Programming for High Performance Seismic Simulation on Petascale Heterogeneous Supercomputers?

Scalable Parallel Programming for High Performance Seismic Simulation on Petascale Heterogeneous Supercomputers by Jun Zhou is 111 pages long, and a total of 27,861 words.

This makes it 37% the length of the average book. It also has 34% more words than the average book.

How Long Does it Take to Read Scalable Parallel Programming for High Performance Seismic Simulation on Petascale Heterogeneous Supercomputers Aloud?

The average oral reading speed is 183 words per minute. This means it takes 2 hours and 32 minutes to read Scalable Parallel Programming for High Performance Seismic Simulation on Petascale Heterogeneous Supercomputers aloud.

What Reading Level is Scalable Parallel Programming for High Performance Seismic Simulation on Petascale Heterogeneous Supercomputers?

Scalable Parallel Programming for High Performance Seismic Simulation on Petascale Heterogeneous Supercomputers is suitable for students ages 10 and up.

Note that there may be other factors that effect this rating besides length that are not factored in on this page. This may include things like complex language or sensitive topics not suitable for students of certain ages.

When deciding what to show young students always use your best judgement and consult a professional.

Where Can I Buy Scalable Parallel Programming for High Performance Seismic Simulation on Petascale Heterogeneous Supercomputers?

Scalable Parallel Programming for High Performance Seismic Simulation on Petascale Heterogeneous Supercomputers by Jun Zhou is sold by several retailers and bookshops. However, Read Time works with Amazon to provide an easier way to purchase books.

To buy Scalable Parallel Programming for High Performance Seismic Simulation on Petascale Heterogeneous Supercomputers by Jun Zhou on Amazon click the button below.