- Видео 162
- Просмотров 109 925
HPC Tech Shorts
США
Добавлен 27 янв 2021
This is the HPC engineering virtual water cooler in AWS, where you'll run into experts from a whole range of fields talking about the technology and services they're involved in building that customers use for HPC in the cloud.
We'll talk about new tech, new techniques, and the amazing open-source groups, partners, and customers we work with. You can find all of this, as well as blogs, tutorials, and workshops at day1hpc.com where we stash everything we create for the HPC community.
Feed our suggestion box for future shows, by telling us what you'd like to hear about. TechHpc (DMs are open), or email us at ask-hpc@amazon.com
We'll talk about new tech, new techniques, and the amazing open-source groups, partners, and customers we work with. You can find all of this, as well as blogs, tutorials, and workshops at day1hpc.com where we stash everything we create for the HPC community.
Feed our suggestion box for future shows, by telling us what you'd like to hear about. TechHpc (DMs are open), or email us at ask-hpc@amazon.com
Lustre performance tuning without the pain
Are you struggling to keep up with the storage demands of your HPC users? Got specific applications that are just ... even more needy than others?
Today we show you how to scale capacity, throughput, and metadata performance on the fly without disruption. This includes a show'n'tell about the new adjustable metadata IOPS features that launched just recently.
Whether you need high IOPS, sub-millisecond latencies, or hundreds of GB/s throughput, FSx for Lustre provides the flexibility to tailor your storage performance to match your compute requirements.
If you have ideas for technical topics you'd like to see us cover in a future show, let us know by emailing us at ask-hpc@amazon.com.
Chapters...
Today we show you how to scale capacity, throughput, and metadata performance on the fly without disruption. This includes a show'n'tell about the new adjustable metadata IOPS features that launched just recently.
Whether you need high IOPS, sub-millisecond latencies, or hundreds of GB/s throughput, FSx for Lustre provides the flexibility to tailor your storage performance to match your compute requirements.
If you have ideas for technical topics you'd like to see us cover in a future show, let us know by emailing us at ask-hpc@amazon.com.
Chapters...
Просмотров: 83
Видео
Behind the scenes: Large-scale analytics for financial market forecasting
Просмотров 12119 часов назад
Your pension fund is probably looked after by a company a lot like CFM. They use large scale, real-time analytics to help them shift their position on the financial markets to control the risk they - and their investors (like all of us) - are exposed to. It's critical stuff if you want to retire happily one day. Today, with the help of our HPC & AI partner, ANEO, we'll get a fascinating look be...
![Introducing AWS Parallel Computing Service [PCS-001]](http://i.ytimg.com/vi/BlgYbb6pdu0/mqdefault.jpg)
Introducing AWS Parallel Computing Service [PCS-001]
Просмотров 1 тыс.14 дней назад
Introducing AWS Parallel Computing Service, which is for customers who run high performance computing (HPC) workloads and build scientific and engineering models so they can accelerate their R&D and drive innovation. If you have ideas for technical topics you'd like to see us cover in a future show, let us know by emailing us at ask-hpc@amazon.com. Chapters 00:00 - Start 02:49 - Migrating HPC a...
![[PCS-011] Configuring EFA and multi-NIC instances for use in AWS PCS](http://i.ytimg.com/vi/5hkA-y4alB0/mqdefault.jpg)
[PCS-011] Configuring EFA and multi-NIC instances for use in AWS PCS
Просмотров 7814 дней назад
You need to do a few key things to get EFA enabled in AWS PCS clusters - and there's an extra step or two if you're using multi-rail EFA instances like the HPC7a or any accelerated instances from the P4 or P5 families. Matt shows us how. Introducing AWS Parallel Computing Service, which is for customers who run high performance computing (HPC) workloads and build scientific and engineering mode...
![[PCS-002] Your first cluster with AWS Parallel Computing Service](http://i.ytimg.com/vi/ciHU2fDzhSc/mqdefault.jpg)
[PCS-002] Your first cluster with AWS Parallel Computing Service
Просмотров 30414 дней назад
In 45 minutes you'll pick up all you need to know to build a complete HPC cluster on AWS. Imagine: an entire HPC cluster - soup to nuts, including fast network, high performance Lustre storage, and hundreds or even thousands of cores at your fingertips - in just a morning's work. This uses the new AWS Parallel Computing Service, which is for customers who run high performance computing (HPC) wo...
![[PCS-003] Launch templates, instance profiles, security groups, and AMIs in AWS PCS](http://i.ytimg.com/vi/YD3DdfKF2yk/mqdefault.jpg)
[PCS-003] Launch templates, instance profiles, security groups, and AMIs in AWS PCS
Просмотров 6714 дней назад
AWS Parallel Computing Service makes use of a lot of AWS technologies to enable you to abstract and tease out the layers of your complex software stacks and the environments they run in. Matt shows us how to extract the maximum benefits from using Launch templates, instance profiles, security groups , and AMIs to your advantage. Introducing AWS Parallel Computing Service, which is for customers...
![[PCS-004] Create and use a custom AMI for AWS Parallel Computing Service](http://i.ytimg.com/vi/3ysMkZrDlGI/mqdefault.jpg)
[PCS-004] Create and use a custom AMI for AWS Parallel Computing Service
Просмотров 7614 дней назад
Custom AMIs are powerful in PCS - and also quite necessary. Matt shows us the easiest way to create one, and lets on some tricks to make it easy - and extendable later on. Introducing AWS Parallel Computing Service, which is for customers who run high performance computing (HPC) workloads and build scientific and engineering models so they can accelerate their R&D and drive innovation. If you h...
![[PCS-015] Bring your own login nodes for AWS Parallel Computing Service](http://i.ytimg.com/vi/Az0wLO7B_Ns/mqdefault.jpg)
[PCS-015] Bring your own login nodes for AWS Parallel Computing Service
Просмотров 12814 дней назад
You can build login nodes that connect to AWS PCS clusters - even if they're not part of the cluster! This is super helpful if you want to give each one of your users access to their own personal login node - to keep their CPU and I/O usage from impacting each other. You can also use this to cook visualization nodes with amazing remote viz tech like DCV. Introducing AWS Parallel Computing Servi...
Why Gencove uses Graviton for most of their work - it's all about the science
Просмотров 14828 дней назад
Fascinated by the behind-the-scenes science that brings steak to your plate? Today we discuss how Gencove works at the intersection of genomics and data science to solve hard problems. Some generate new techniques - like gleaning new insights out of old data - and some are just more practical and involve putting food on your plate. The team at Gencove long ago ported most of their workloads to ...
Putting visual login nodes on your ParallelCluster the easy way
Просмотров 177Месяц назад
Looking for an easy way to offer your ParallelCluster environment to your end users through a visual interface? Look no further! We're excited to show you how to seamlessly connect ParallelCluster with Research & Engineering Studio (RES). This powerful combination gives you near instant access to your HPC environment right from within RES's interactive desktops portal. You can visualize results...
DCV Access Console - a walk thru
Просмотров 2272 месяца назад
Managing remote desktops can be a headache as your organization scales. Keeping track of users, hosts, and session configurations only gets more complex over time. In this video, see how DCV Access Console solves these problems. Andrew gives a full tour of the easy-to-use UI that lets end users instantly spin up Windows or Linux desktops with a single click. Admins get powerful tools to manage ...
Introducing DCV's new Access Console
Просмотров 1882 месяца назад
DCV Access Console is an exciting new solution that allows companies to provide secure, remote access to graphical workloads for their employees. For companies struggling to scale remote work capabilities while maintaining security, DCV Access Manager is a game-changer. It enables administrators to easily create and configure virtual workspaces tailored to each employee's needs - from Windows o...
Building sustainable HPC
Просмотров 1982 месяца назад
You could be like Baker Hughes, and reduce your carbon footprint for your HPC cluster by 99%. How? Watch this talk by Francesco Ruffino, one of our global experts on HPC who has a passion for making HPC more efficient - and kinder to the planet. We filmed this on site at ISC'24 where the theme was "reinventing HPC". The cloud is a perfect place to do just that. If you have ideas for technical t...
Robotec.ai is powering their simulations using a new Batch feature
Просмотров 1672 месяца назад
Intrigued by how advanced robotics and simulation can optimize industrial operations? Today we offer a behind-the-scenes look at how Robotec.ai leverages Batch's new multi-container features to simulate autonomous vehicle environments in mines, on farms, and in warehouse. Watch as robots autonomously navigate underground tunnels, calculating optimal routes and tweaking productivity. Dominik Jar...
Agent based models let you simulate ... basically everything.
Просмотров 1,6 тыс.3 месяца назад
Fascinated by how complex systems like economies, supply chains, and biological networks function? Want to understand how to model and optimize these systems using the power of simulation, graph theory, and maybe even AI? Today's Tech Short features Justin, the founder of Simudyne, a company using a novel graph computational approach to create massive agent-based simulations of entire economies...
Nextflow, Seqera containers, and Arm support for bioinformatics
Просмотров 2583 месяца назад
Nextflow, Seqera containers, and Arm support for bioinformatics
Announcing Seqera Containers - supported by AWS
Просмотров 3993 месяца назад
Announcing Seqera Containers - supported by AWS
How accurate (and fast) are Surrogate ML models in engineering design?
Просмотров 3383 месяца назад
How accurate (and fast) are Surrogate ML models in engineering design?
AI and surrogate machine learning models might change engineering design .. very soon.
Просмотров 3024 месяца назад
AI and surrogate machine learning models might change engineering design .. very soon.
Building TByte/s Lustre without pain or remorse
Просмотров 3444 месяца назад
Building TByte/s Lustre without pain or remorse
Tunnelling SSH into your cloud instances, without exposing them to the internet
Просмотров 2184 месяца назад
Tunnelling SSH into your cloud instances, without exposing them to the internet
Cooking real clusters from the HPC Recipe Library
Просмотров 1975 месяцев назад
Cooking real clusters from the HPC Recipe Library
Moving your /home off the cluster in ParallelCluster
Просмотров 2145 месяцев назад
Moving your /home off the cluster in ParallelCluster
A walk through Research and Engineering Studio on AWS
Просмотров 9805 месяцев назад
A walk through Research and Engineering Studio on AWS
Overview of services built for HPC customers @ AWS
Просмотров 5175 месяцев назад
Overview of services built for HPC customers @ AWS
Performance tools Pt3 - RPATH, RUNPATH, and the loader
Просмотров 1556 месяцев назад
Performance tools Pt3 - RPATH, RUNPATH, and the loader
Performance tools Pt2 - static vs dynamic linking
Просмотров 1416 месяцев назад
Performance tools Pt2 - static vs dynamic linking
Performance tools Pt 1 - binutils and friends
Просмотров 3037 месяцев назад
Performance tools Pt 1 - binutils and friends
In search of performance, details matter - behind the scenes in ParallelCluster
Просмотров 2387 месяцев назад
In search of performance, details matter - behind the scenes in ParallelCluster
What you need to know about AWS Batch on Fargate
Просмотров 2,2 тыс.7 месяцев назад
What you need to know about AWS Batch on Fargate