{"id":1199,"date":"2021-11-16T08:32:37","date_gmt":"2021-11-16T08:32:37","guid":{"rendered":"https:\/\/salarydistribution.com\/machine-learning\/2021\/11\/16\/worlds-fastest-supercomputers-changing-fast\/"},"modified":"2021-11-16T08:32:37","modified_gmt":"2021-11-16T08:32:37","slug":"worlds-fastest-supercomputers-changing-fast","status":"publish","type":"post","link":"https:\/\/salarydistribution.com\/machine-learning\/2021\/11\/16\/worlds-fastest-supercomputers-changing-fast\/","title":{"rendered":"World\u2019s Fastest Supercomputers Changing Fast"},"content":{"rendered":"
\n

Modern computing workloads \u2014 including scientific simulations, visualization, data analytics, and machine learning \u2014 are pushing supercomputing centers, cloud providers and enterprises to rethink their computing architecture.<\/p>\n

The processor or the network or the software optimizations alone can\u2019t address the latest needs of researchers, engineers and data scientists. Instead, the data center is the new unit of computing, and organizations have to look at the full technology stack.<\/p>\n

The latest rankings of the world\u2019s most powerful systems show continued momentum for this full-stack approach in the latest generation of supercomputers.<\/p>\n

NVIDIA technologies accelerate over 70 percent, or 355, of the systems on the TOP500 list<\/a> released at the SC21 high performance computing conference this week, including over 90 percent of all new systems. That\u2019s up from 342 systems, or 68 percent, of the machines on the TOP500 list released in June<\/a>.<\/p>\n

NVIDIA also continues to have a strong presence on the Green500 list of the most energy-efficient systems, powering 23 of the top 25 systems on the list, unchanged from June. On average, NVIDIA GPU-powered systems deliver 3.5x higher power efficiency than non-GPU systems on the list.<\/p>\n

Highlighting the emergence of a new generation of cloud-native systems, Microsoft\u2019s GPU-accelerated Azure supercomputer<\/a> ranked 10th on the list, the first top 10 showing for a cloud-based system.<\/p>\n

AI is revolutionizing scientific computing.\u00a0 The number of research papers leveraging HPC and machine learning has skyrocketed in recent years; growing from roughly 600 ML + HPC papers submitted in 2018 to nearly 5,000 in 2020.<\/p>\n

\"\"<\/p>\n

The ongoing convergence of HPC and AI workloads is also underscored by new benchmarks such as HPL-AI and MLPerf HPC.<\/p>\n

HPL-AI is an emerging benchmark of converged HPC and AI workloads that uses mixed-precision math \u2014 the basis of deep learning and many scientific and commercial jobs \u2014 while still delivering the full accuracy of double-precision math, which is the standard\u00a0 measuring stick for traditional HPC benchmarks.<\/p>\n

And MLPerf HPC addresses a style of computing that speeds and augments simulations on supercomputers with AI, with the benchmark measuring performance on three key workloads for HPC centers: astrophysics (Cosmoflow), weather (Deepcam) and molecular dynamics (Opencatalyst).<\/p>\n

NVIDIA addresses the full stack with GPU-accelerated processing, smart networking, GPU-optimized applications, and libraries that support the convergence of AI and HPC. This approach has supercharged workloads and enabled scientific breakthroughs.<\/p>\n

Let\u2019s look more closely at how NVIDIA is supercharging supercomputers.<\/p>\n

Accelerated Computing<\/b><\/h2>\n

The combined power of the GPU\u2019s parallel processing capabilities and over 2,500 GPU-optimized applications allows users to speed up their HPC jobs, in many cases from weeks to hours.<\/p>\n

We\u2019re constantly optimizing the CUDA-X libraries<\/a> and the GPU-accelerated applications<\/a>, so it\u2019s not unusual for users to see an x-factor performance gain on the same GPU architecture.<\/p>\n

As a result, the performance of the most widely used scientific applications \u2014 which we call the \u201cgolden suite\u201d \u2014 has improved 16x over the past six years, with more advances on the way.<\/p>\n

<\/p>\n

\"\"<\/p>\n

<\/a>

16x performance on top HPC, AI and ML apps from full-stack innovation.**<\/figcaption><\/figure>\n

And to help users quickly take advantage of higher performance, we offer the latest versions of the AI and HPC software through containers from the NGC catalog<\/a>. Users simply pull and run the application on their supercomputer, in the data center or the cloud.<\/p>\n

Convergence of HPC and AI\u00a0<\/b><\/h2>\n

The infusion of AI in HPC helps researchers speed up their simulations while achieving the accuracy they\u2019d get with the traditional simulation approach.<\/p>\n

That\u2019s why an increasing number of researchers are taking advantage of AI to speed up their discoveries.<\/p>\n

That includes four of the finalists for this year\u2019s Gordon Bell prize<\/a>, the most prestigious award in supercomputing. Organizations are racing to build exascale AI computers to support this new model, which combines HPC and AI.<\/p>\n

That strength is underscored by relatively new benchmarks, such as HPL-AI and MLPerf HPC, highlighting the ongoing convergence of HPC and AI workloads.<\/p>\n

To fuel this trend, last week NVIDIA announced a broad range of advanced new libraries and software development kits<\/a> for HPC.<\/p>\n

Graphs \u2014 a key data structure in modern data science \u2014 can now be projected into deep-neural network frameworks with Deep Graph Library, or DGL, a new Python package.<\/p>\n

NVIDIA Modulus<\/a> builds and trains physics-informed machine learning models that can learn and obey the laws of physics.<\/p>\n

And NVIDIA introduced three new libraries:<\/p>\n