Main

AMD + Ansys – Together We Advance

In this insightful video, we dive deep into the groundbreaking partnership between Ansys and AMD, highlighting how their collaboration is pushing the boundaries of high-performance computing (HPC) and driving technological advancement. Discover the synergy between Ansys's predictive simulation capabilities and AMD's high-performance engines, and how together they are setting new benchmarks in computational efficiency and performance optimization. From achieving a remarkable 26% performance improvement in Ansys Mechanical using AMD EPYC processors to leveraging AMD GPUs on the Azure Cloud for superior performance in motorsports applications, this partnership exemplifies innovation. The video also explores the advantages of AMD 3D V-Cache technology and its perfect fit for high-performance computing workloads, as demonstrated in the HBv4 Azure instance. Moreover, learn about the collaborative efforts with Honda Motorsports to harness exciting applications in fluids and mechanical simulations, and the role of Ansys's RedHawk-SC in ensuring design correctness before silicon investment. The discussion further delves into the transformative potential of AI/ML technology in achieving super-linear speedups in network training on GPUs, showcasing early successes in CAE tools and other applications. This partnership not only delivers enhanced performance but also ensures energy-efficient solutions for customers, reflecting a shared commitment to excellence and innovation. Join us as we explore the fruits of this collaboration, offering a glimpse into the future of high-performance computing and its impact on human advancement. 00:00 Introduction to Ansys and AMD Partnership 00:10 Power of Predictive Simulation for Advancement 00:20 Complementary Missions of Ansys and AMD 00:33 Performance Improvements in Ansys Mechanical and HFSS 00:56 Joint Work with Honda Motorsports and Azure Cloud 01:14 AMD 3D V-Cache Technology and High-Performance Computing 01:45 Ansys's RedHawk-SC for Silicon Design Verification 02:02 Impact of AI/ML Technology on High-Performance Computing 02:24 Benefits of the Ansys-AMD Partnership for Customers For more information, visit https://www.ansys.com/partner-ecosystem/high-performance-computing-partners/amd.

Ansys

7 days ago

it's great to talk about the partnership between NIS and and AMD PR you know partnership is one of my favorite topics we help our customers through the power of predictive simulation right to to really do uh human advancement AMD mission is completely complimentary uh to the ancest mission because we're the engines of high performance and we're able to really drive an optimization that really all HBC C customers will benefit from with this joint work we have been able to take aners mechanical an
d get 26% higher performance using the Epic processors the same thing with hfss and some current work that is going on with with ls Dino these are the kind of improvements that often don't come until you have a new generation of a processor but this was done by partnership and optimizing uh the software and the stack together I want to highlight some work that we are doing jointly with our joint customer we are working with Honda motor supports to take exciting applications in fluids and mechani
cal and leveraging the AMD gpus on the zure cloud right to to get really really fantastic performance so we have the AMD 3D vcash so it's AMD epic it's a server chip x86 server with a 3D cache stacked right on top of the CPU so it turns out that that architecture is ideal for high performance compute workload and uh the example I'll give you is the HB V4 uh Azure instance that has this AMD 3D V cache and so I think we've been able to very closely partner with you and Azure that can make that hig
h performance configuration broadly available across the world with uh red hoc uh SC it's a fundamental part of our what I call our sign off we rely on aners to make sure that we're correct by construction by Design before we actually invest in building that silicon in fact we we are recognizing that a IML is one of the most destructive Technologies in our area all our prior work has been to take our use numerical methods approaches IM IML provides the opportunity to get sort of super linear spe
ed UPS as you train those neural networks on using the gpus I mean you can get fantastic speed ups and we have got early results with some of our tools like cfd and so on the partnership uh delivers more performance uh and energy efficient performance for our mutual customers so it's been an excellent partnership where you have shared with us the performance demands or any bottlenecks that you have in your high performance application and we have tuned it and built it right in our optimized comp
ilers and [Music] libraries

Comments