Last year, Oracle announced that it was the first major cloud provider to make NVIDIA A100 Tensor Core GPUs generally available in the cloud. The bare metal shape is accelerating innovation for enterprises around the world by enabling large artificial intelligence (AI) training workloads to scale to more than 500 NVIDIA A100 tensor Core GPUs in a single high-speed, low-latency cluster. This year, Oracle’s collaboration with NVIDIA continues to grow with Oracle becoming a sponsor at NVIDIA GTC.
GTC is a global AI conference taking place online November 8–11. Oracle is showcasing its work with NVIDIA through two sponsored sessions available for viewing on the GTC website.
Data Science service batch jobs with GPU acceleration (presented by Oracle Cloud Infrastructure) by Lyudmil Pelov
Time is one of the most important metrics in business. A successful business is one that can maximize its efficiency, which is exactly what the Oracle Data Science Platform team’s new feature, Jobs, is designed to do. With Jobs, processing tasks have never been easier. In this session, Lyudmil dives deep into the features and usage of Jobs. He then runs through a short demo on how to access jobs and demonstrates why it’s necessary for maximizing workload efficiency.
Speeding up genomic analysis time to solution by 20-fold using NVIDIA Clara Parabricks and Oracle for Research by David Chen, Clifford Patterson, and Gloria Lee
Traditionally, the Genome Analysis Toolkit (GATK) has been the go-to tool used for genomic sequencing analysis, running on CPUs. While GATK can achieve high accuracy, it’s still imperfect in memory management and running efficiency, which can lead to a high per job cost. The NVIDIA Clara Parabricks application framework offers the perfect solution to this approach, running on GPUs instead of CPUs to produce the same results as GATK with lower costs and faster speeds. In this presentation, David, Clifford, and Gloria discuss recent work in testing and deploying NVIDIA Clara Parabricks and compare these results to an identical analysis done on CPU servers.
Hope to see you there!
Find our sessions on the GTC webpage and register for free to access these sessions and our virtual booth November 8–11.