Download CUDA Fortran for Scientists and Engineers. Best Practices by Gregory Ruetsch PDF

By Gregory Ruetsch

CUDA Fortran for Scientists and Engineers indicates how high-performance software builders can leverage the ability of GPUs utilizing Fortran, the well-known language of clinical computing and supercomputer functionality benchmarking. The authors presume no previous parallel computing event, and canopy the fundamentals in addition to top practices for effective GPU computing utilizing CUDA Fortran.

To assist you upload CUDA Fortran to current Fortran codes, the ebook explains find out how to comprehend the objective GPU structure, determine computationally in depth components of the code, and adjust the code to regulate the knowledge and parallelism and optimize functionality. All of this can be performed in Fortran, with no need to rewrite in one other language. each one suggestion is illustrated with genuine examples so that you can instantly overview the functionality of your code in comparison.

  • Leverage the ability of GPU computing with PGI's CUDA Fortran compiler
  • Gain insights from individuals of the CUDA Fortran language improvement team
  • Includes multi-GPU programming in CUDA Fortran, overlaying either peer-to-peer and message passing interface (MPI) approaches
  • Includes complete resource code for all of the examples and a number of other case reports
  • Download resource code and slides from the book's significant other website

Show description

Read Online or Download CUDA Fortran for Scientists and Engineers. Best Practices for Efficient CUDA Fortran Programming PDF

Similar design & architecture books

Cisco Press MPLS and VPN Architectures

A realistic consultant to knowing, designing, and deploying MPLS and MPLS-enabled VPNs In-depth research of the Multiprotocol Label Switching (MPLS) structure particular dialogue of the mechanisms and lines that represent the structure learn the way MPLS scales to aid tens of hundreds of thousands of VPNs vast case reports advisor you thru the layout and deployment of real-world MPLS/VPN networks Configuration examples and guidance help in configuring MPLS on Cisco® units layout and implementation concepts assist you construct numerous VPN topologies Multiprotocol Label Switching (MPLS) is an leading edge strategy for high-performance packet forwarding.

Assignment Problems in Parallel and Distributed Computing

This ebook has been written for practitioners, researchers and stu­ dents within the fields of parallel and disbursed computing. Its aim is to supply designated insurance of the purposes of graph theoretic tech­ niques to the issues of matching assets and standards in multi­ ple desktops.

Cloud Computing. Theory and Practice

Cloud Computing: idea and perform offers scholars and IT execs with an in-depth research of the cloud from the floor up. starting with a dialogue of parallel computing and architectures and dispensed structures, the booklet turns to modern cloud infrastructures, how they're being deployed at prime businesses resembling Amazon, Google and Apple, and the way they are often utilized in fields resembling healthcare, banking and technology.

Microservices, IoT, and Azure: Leveraging DevOps and Microservice Architecture to deliver SaaS Solutions

This publication offers functional information for adopting a excessive speed, non-stop supply approach to create trustworthy, scalable, Software-as-a-Service (SaaS) options which are designed and outfitted utilizing a microservice structure, deployed to the Azure cloud, and controlled via automation. Microservices, IoT, and Azure deals software program builders, architects, and operations engineers' step by step instructions for construction SaaS applications—applications which are to be had 24x7, paintings on any machine, scale elastically, and are resilient to change--through code, script, workouts, and a operating reference implementation.

Extra info for CUDA Fortran for Scientists and Engineers. Best Practices for Efficient CUDA Fortran Programming

Example text

5 Execution Configuration . . . . . . . . . . 1 Thread-Level Parallelism . . . . . . . . . 1 Shared Memory . . . . . . . . . 2 Instruction-Level Parallelism . . . . . . . . CUDA Fortran for Scientists and Engineers. 00003-1 © 2014 Elsevier Inc. All rights reserved. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6 Instruction Optimization . . . 1 Device Intrinsics . . . . 1 Directed Rounding . 2 C Intrinsics . . . 3 Fast Math Intrinsics . 2 Compiler Options . . . . 3 Divergent Warps . . . . 7 Kernel Loop Directives .

107 108 108 108 108 108 109 110 113 113 114 In the previous chapter we discussed how we can use timing information to determine the limiting factor of kernel execution. Many science and engineering codes turn out to be bandwidth bound, which is why we devote the majority of this relatively long chapter to memory optimization. CUDA-enabled devices have many different memory types, and to program effectively, we need to use these memory types efficiently. Data transfers can be broken down in to two main categories: data transfers between host and device memories, and data transfers between different memories on the device.

Download PDF sample

Rated 4.02 of 5 – based on 35 votes