DOES Artificial Intelligence Need Special Computers ? Quo Vadis Hardware for AI?

14.30 - Slovak Time zone,
06.30 am Los Angeles time zone
8.30pm - Beijing Time zone
9.30pm - Tokyo Time zone

language : English

„SOA in HARDWARE for Artificial Intelligence“
(The State of the ART in HARDWARE for AI)

Presentation of leading-edge companies
Huawei versus NVIDIA for AI present and Future.

Program :

1. Opening
2. NVIDIA representative - 30 minutes (Ralph Hinsche, NVIDIA)
3. Huawei representative - 30 minutes (Goran Licanin, Huawei)
4. Q&A to the topic
   ----

NVIDIA Title : “Extreme Scaling of AI : Supercharging Machine Learning and Deep Learning"

Abstract :
In this presentation we will demonstrate the extreme scaling capabilities within a single DGX A100 node with 5 Petaflop/s (Tensor Core operations). Following that we highlight the mature and advanced DL and ML ecosystem supporting DGX. We conclude with architectural blueprints for massive multi-node scaling to further expand overall performance and provide an inside view on our own SaturnV AI-Supercluster.

Bio :
Ralph Hinsche is currently responsible at NVIDIA as a Business Development Manager for GPU-Computing (TESLA) and Deep Learning (DGX) within the DACH and Eastern Europe region since 2014. With more than 30 years of HPC experience Ralph started as a student in 1987 at Parsytec (Transputer, OCCAM) in Aachen/Germany. This was followed by head of department activities at several SUN Microsystems partners with the focus on HPC and he contributed to a national development project (parallel computer GIGAmachine). In 1996 he became a sales engineer at EUREM with a focus on "Wide Area Automation" (Distributed Intelligence). In his former position, he was Key Account Manager at circular for nearly 10 years mainly in the field of HPC including close cooperation with SUN Microsystems and DELL.

------

HUAWEI Title : Extremely scalable architecture and solution for Deep Learning computing with NPUs

Abstract :
This presentation will deliver the short introduction to Huawei’s NPU (Neural Processor Unit) with specialized Da Vinci architecture for neural networks as well as the extremely scalable solution of computing platform for acceleration of deep learning. Huawei will demonstrate full-stack DL (Deep Learning) platform with the Ultimate Development and User Cases of Device-Edge-Cloud Collaboration from Atlas 200 module to Atlas 900 cluster which is still one of the most high-performance clusters in its class. And all this information will be illustrated by market examples of extreme achievements in the area of image processing and NLP.

Bio :
Goran Licanin is working for more than 20 years in the ICT domain, from engineer to consultant for Fortune 500 in IT strategy, IT infrastructure, digitization, software development and data science. Last five years, he has been working in HUAWEI CEE&Nordic regional office, helping companies in digital transformation and adoption of 5G, AI, IoT solutions in the Carrier BG. He has a master’s degree in computer science, and he got his specialization in Strategy and innovations at MIT in the USA. Today he is in the position of Industry and Ecosystem Development Director for Cloud&AI in CEE&Nordic
please reserve your time

-----

Join Zoom Meeting
https://us02web.zoom.us/j/82761546700

Meeting ID: 827 6154 6700
Passcode: DEEP