1 / 8

Reviving Moore’s Law: GPU as a Service (GaaS) in Modern Computing

In the era of cloud computing, the original promise of Moore’s Law has shifted from faster CPU speeds to increased parallel processing capabilities. This presentation from the AWS Hackathon Berlin 2012 explores the concept of GPU as a Service (GaaS), demonstrating how startups can leverage software to unify multiple GPUs into a single, powerful processing unit. By showcasing benchmarks like vector addition and matrix multiplication, we illustrate the potential of utilizing cloud-based GPU resources for computational efficiency and scalable performance.

etenia
Télécharger la présentation

Reviving Moore’s Law: GPU as a Service (GaaS) in Modern Computing

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. “Till about 2002 you could safely misinterpret Moore’s law as promising that clock speeds would double every 18 months. Actually what it says is that circuit densities will double every 18 months.” “It would be great if a startup could give us something of the old Moore's Law back, by writing software that could make a large number of CPUs look to the developer like one very fast CPU” “Not any more. Intel can no longer give us faster CPUs, just more of them.” (Paul Graham) GPU as a Service (GaaS) - AWS Hackathon Berlin 2012

  2. PC Android GaaS API EC2 (cg1.4xlarge) Ruby on Rails NVidia CUDA GPU 1 GPU 2 GPU as a Service (GaaS) - AWS Hackathon Berlin 2012

  3. Android PC GaaS API EC2 (m1.medium) Ruby on Rails Job Queue (Amazon SQS) EC2 (cg1.4xlarge) EC2 (cg1.4xlarge) Low-level service Low-level service NVidia CUDA NVidia CUDA GPU 1 GPU 1 GPU 1 GPU 1 GPU as a Service (GaaS) - AWS Hackathon Berlin 2012

  4. Live Demo! GPU as a Service (GaaS) - AWS Hackathon Berlin 2012

  5. Benchmark #1: Vector addition CPU total: n * kcpu EC2 total: n * (knet+ kcuda) kcuda<< kcpu<< knet:-( Network: n * knet CPU comp: n * kcpu CUDA comp: n * kcuda GPU as a Service (GaaS) - AWS Hackathon Berlin 2012

  6. Benchmark #2: Matrix multiplication Network: n2 * knet CPU comp: n2,8 * kcpu CUDA comp: n2,8 * kcuda CPU total: n2,8 * kcpu EC2 total: n2 * knet + n2,8 * kcuda kcuda<< 1 & n2,8 >> n2:-) GPU as a Service (GaaS) - AWS Hackathon Berlin 2012

  7. GaaS up! http://doiop.com/gaas GPU as a Service (GaaS) - AWS Hackathon Berlin 2012

More Related