PSSC Labs Cluster Management System, CBeST to Migrate to Ubuntu in Light of Recent CentOS Stream Announcements

By marketing@site-a.com

When Red Hat®, Inc. announced in late 2020 that they would soon be releasing CentOS Stream, the tech community responded with an uproar of confusion and uncertainty. With many wondering about the future of their workloads and the expiration of their current operating system, solution providers like PSSC Labs have had to adapt to our clients’ needs. (Read more about CentOS Stream here).

In response to this operating system shift, we’re making a few changes to help. First, our CBeST Cluster Management Software Stack will now be compatible with Ubuntu. This will offer our clients a path out of their current CentOS/Red Hat environments without affecting their current scientific and computing applications. PSSC Labs will continue to support a truly open source cluster management environment with all of the same key features and benefits that have been present before.

Key features:

  • Operating System and BIOS performance optimizations tailored by CPU type (AMD®/Intel®)
  • Support for both Intel®, AMD®, NVIDIA® GPUs
  • NVIDIA® CUDA SDK software preinstalled
  • Multiple MPI platforms and versions installed by default
  • Environment Modules configuration. This will allow user environment customization to switch between different default MPI/application/compiler versions/environments.
  • Support for Cloud Bursting to Microsoft® Azure, Amazon® AWS and Google® Compute
  • Support for connecting to different parallel file systems including Gluster, Lustre, and Ceph
  • Support for Mellanox® InfiniBand up to 200 Gb/s and Intel® Omni-Path up to 100 Gb/s
  • Support for high speed ethernet standards 10/25/40/100 Gb/s.
  • Enhanced usability and management
  • Continued improvements on the build and use of slurm batch scheduler
  • Slurm customization, including support for GPUs and Quality of Service prioritizing
  • Slurm will allow support for bursting cloud deployment

Additionally, we offer various AI and HPC solutions that are compatible with several other operation system options, such as Microsoft® Windows® Server, Red Hat® Enterprise Linux, SUSE® Linux®, Ubuntu Linux, VMware® Linux, and more. Our HPC servers are designed to help you achieve your business objectives and overcome any potential roadblocks. For more information on our solutions, please contact us at 4sales@site-a.com or give us a call at (949) 380-7288.

Shearwater Technology: Q&A with Owner Andy Witzig

By marketing@site-a.com

PSSC Labs takes pride in their ability to help small and medium sized businesses accomplish their computational goals, which is exactly what they did for Shearwater Technology, a leading Illinois-based engineering firm. Shearwater both grew their business and became more customer responsive thanks to the implementation of an HPC Cluster solution from PSSC Labs. With 400% more core processing hours, $71,000+ in cost savings, and 0 days lost during installation, it’s no secret why Shearwater selected PSSC Labs as their High Performance Computing (HPC) vendor.

Below is a short Q&A with Andy Witzig, President of Shearwater Technology, that outlines what exactly Shearwater Technology was looking to achieve and how PSSC Labs was able to be a part of their growth.

Q:  Andy, what is Shearwater Technology and what is your role there?

A: Shearwater is a small engineering services company located in Central Illinois.  For the past decade, we have specialized in providing high-quality modeling and simulation support services to our customers in both government and commercial sectors.  We provide modeling solutions in the areas of high fidelity multi-physics, internal ballistic, external ballistics, impact, penetration, and blast. We also have deep experience modeling system performance using large-scale system effectiveness tools.

My role as President/CEO of Shearwater is multifaceted.  In addition to leading the company from an administrative and technical perspective, I am also a principle consultant and analyst. I work directly with our customers to solve their problems and ensure that the quality of our services is impeccable.  It’s also my job to ensure that Shearwater’s capabilities and resources are cutting edge and always growing. So, that means I’m regularly evaluating our computational and analytical resources to determine the key technologies in which we need to invest in order to continue providing the best service possible.

Q: What lead you to shop around for a high performance computing partner?

A: In the past few years, Shearwater had seen growing customer demand to process larger computational problems and we simply did not have the capacity to meet that demand.  We also realized that purchasing a HPC cluster would be a long-term, strategic investment that would change our business model and open up the possibility to service a broader range customers.

Q: What lead you to ultimately select the PSSC Labs HPC Cluster over other potential vendors?

A: First, as a small business owner I liked the idea of working with a smaller HPC supplier, like PSSC Labs, that would be more focused on product quality and customer service. Second, we had experience purchasing smaller HPC workstations and leasing time on commercial HPC clusters as a service, and we saw quickly that neither approach was going to be effective in the long term.  The HPC workstations could not scale to meet our needs, and leasing was much too expensive for the amount of processing we needed to perform.  In the end, it became clear that purchasing an HPC cluster from a trusted supplier was the best path forward.

Q: What impact did the HPC cluster you purchased from PSSC Labs have on your business once implemented?

A: Within ten months of purchasing the cluster, we logged more than 650,000 core-hours of processing.  We estimated that it would have cost Shearwater $71,000 to process the same number of core-hours with a HPC leasing service.  We also estimated that we processed over four times the number of core-hours we would have logged with our previous HPC resources.  This means we were able to generate four times the data for our customers than in prior years. Another result of purchasing the cluster was that we were able to service more customers simultaneously which dramatically increased our revenue streams and overall income.

Q: What was the implementation process of the new cluster like? 

A: The implementation was really seamless because PSSC Labs manufactures products that are turnkey, custom solutions. This was important for the Shearwater team because, we really needed a HPC solution that we could install and put to work almost immediately. PSSC Labs sent their experts out to install the cluster for us and we were running workloads on it that same day. It was a really quick turnaround.

Q: Is there anything else you wanted to note about your experience working with PSSC Labs?

A: Yes, within less than a year of owning our PSSC cluster we started upgrading it by purchasing four new compute nodes. The nodes were safely shipped to Shearwater and we were able to install them quickly on our own. This is important because it represents the ability for Shearwater to continue expanding our HPC resource through periodic investments with immediate returns as our business grows.

I’m also thrilled with the customer support we’ve received from PSSC Labs. Their team is always very responsive and we’ve had really great, helpful interactions with them  It’s been such a positive experience, we’re already working on building our second PSSC cluster.

PSSC Labs is an American manufacturer of high performance computing (HPC) and AI solutions, here to help you take control with affordable, turnkey high-performance computing solutions. Find out how an on-premise HPC cluster can take your business to the next level. For more information, please contact us at 4sales@site-a.com or (949) 380-7288.

Chip Shortages Cause Challenges for IT Supply Chain

By marketing@site-a.com

If you’ve noticed it’s been more difficult than usual to get your hands on the IT parts you need, you’re not the only one. Growing demand and manufacturing hiccups seem to be to blame for the significant and growing strain on IT supply chains and forecasts indicate these issues could increase well into 2021.

CPUs, GPUs and memory chips from tech giants like AMD, NVIDIA, and Intel have all announced recent shortages for their popular components that could challenge organizations into 2021. Add to this existing chaos the implications of COVID-19 and it’s easy to see why manufacturers are having such a difficult time getting their technology products to market. Technology leaders point to a core shortage of the chemicals needed to manufacture key technology components, including memory chips as the real issue, in addition to a growing over-ordering out of an abundance of caution in uncertain times. “A lot of companies are over-ordering because of a concern of availability of parts. Once you have the illusion of short supply, everyone orders more.” (“You’re not imagining things, there is a serious chip shortage”, NetworkWorld).

So what does the current chip shortage mean for companies looking to purchase AI and HPC solutions within the year? It means that many organizations should look at the purchase cycle realistically. It’s important for those in the market for new technology solutions to see the reality that AI and HPC component delays will likely result in pushed back delivery timeframes. This means if you have a business objective that requires an upgrade of computing power, the quicker you act, the better. Meet with vendors and discuss your computing needs so that you can get ahold of the high performance computing solutions you need with minimal wait times.

At PSSC Labs, we’re always available to discuss your unique business objectives to make sure you get the perfect, custom solution for your work. If you have questions about our solutions or how we can help you, please reach out to us anytime at 4sales@site-a.com or give us at call at 949-380-7288.

The PowerServe Uniti HGX: An Artificial Intelligence Powerhouse

By marketing@site-a.com

LAKE FOREST, CALIF., FEBRUARY 1, 2021 – PSSC Labs, a developer of artificial intelligence (AI), high performance computing (HPC) and big data computing solutions, today announced the launch of PowerServe Uniti HGX, the newest product in their AI/HPC Sever product family.

This unique platform is built to reach new heights in the artificial intelligence (AI) and machine learning (ML) performance spaces, using up to 8 A100 GPUs connected via SXM4 Form Factor. Processor support includes 2 x Intel® Xeon® Scalable Processors or AMD EPYCTM processors with a maximum 128 total cores/256 cores with hyperthreading. With memory support of 32 x DDR4 memory Dimm slots, up to 4 TB memory capacity, 3200 / 2933 / 2666 / 2400 MHz support, and 8 channel memory architecture, this system is nothing short of extraordinary. Computer scientists will experience a faster time to solution, allowing massive data sets and exploding models to be crunched without the complexity of a full cluster. This is the first server that we’ve seen that can truly replace the power of an entire cluster.

For information regarding pricing and configuration details, please visit www.site-a.com or give PSSC Labs a call a (949) 380-7288.

About PSSC Labs

For technology powered visionaries with a passion for challenging the status quo, PSSC Labs is the answer for hand-crafted HPC and Big Data computing solutions that deliver relentless performance with the absolute lowest total cost of ownership. We are true innovators offering high performance computing solutions to solve the world’s most demanding problems. For 25+ years, organizations of all sizes and from a variety of sectors rely on PSSC Labs’ computing systems. We are proud to support many departments within the United States government, Fortune 500 companies, as well as small and medium-sized businesses.

All products are designed and built at the company’s headquarters in Lake Forest, California.

CentOS 8 to CentOS Stream: What It Means for the HPC Community

By marketing@site-a.com

This past December, Red Hat®, Inc. announced the upcoming release of CentOS Stream scheduled to take place in late 2021. With the community around the current versions of CentOS being as vast as it is, this announcement unsurprisingly caused a lot of uproar and confusion as you can see on the CentOS blog (https://blog.centos.org/2020/12/future-is-centos-stream/). Many have been left to wonder about the future of their operating system. Will CentOS Stream be a replacement for CentOS 8 entirely? Or merely an evolution of it, allowing users to choose the operating system version they continue to run their workloads on? The PSSC Labs Team gathered together to study this news in detail with the hope of clearing up any confusion that might have arisen.

In September of 2019, Chris Wright, VP and CTO at Red Hat said of CentOS Stream, “CentOS Stream is an upstream development platformed designed for CentOS community members, Red Hat partners, ecosystem developers, and many other groups to more quickly and easily see what’s coming next in the Red Hat Enterprise Linux (RHEL) ecosystem and to help shape the product.” In essence, CentOS Stream is rolling out in something of a pilot phase. Red Hat is encouraging their partners and developers to not only begin engagement in CentOS Stream but to actually start utilizing it to test solutions to their own unique business challenges. This gives users the opportunity to interact within the operating system while its official version is still being developed. The idea is that Red Hat can identify how its users utilize and interact with the system and then use that information to further build out the official version. Red Hat’s intended goal is to ensure that CentOS Stream is designed to work the way users want it to, but practically what does this mean for you and your team?

What does this mean for CentOS 8?

For now, it seems that CentOS 8 will remain an option for users that wish to continue their workloads on this version of the operating system throughout the year. But as stated by Red Hat themselves, that time will come to an end very shortly. “Updated to the CentOS Linux 8 distribution continue until the end of 2021; users can choose to switch over directly to CentOS Stream 8” where updates distribution will continue through the full Red Hat Enterprise Linux (RHEL) support phase.

Consider Your Options

Rather than migrate to CentOS Stream, consider the possibility of migrating to other versions of Linux, such as Ubuntu. Ubuntu is a Linux distribution based on Debian and mostly composed of free and open-source software. It’s also released in three editions: Desktop, Server, and Core for Internet of Things devices and robots. Ubuntu has a long history of success when it comes to supporting computationally intensive applications, meaning our clients in the Life Science, Design and Engineering, and other spaces will still have a supported operating system with Ubuntu.

PSSC Labs Supports You

For PSSC Labs clients that choose to migrate to Ubuntu, we are releasing a new version of our CBeST Cluster Management Software Stack. CBeST is our full featured cluster management software, compromised of kernel optimizations, custom scripts and open source tools that are elegantly integrated, optimized and customized for your specific PowerWulf HPC Cluster hardware and applications. It’s offered at no additional cost to our cluster clients.

No matter the future of your current operating system, PSSC Labs is here to support our clients however they need. Please feel free to contact us with any questions.

A Checklist for Building AI Workstations

By marketing@site-a.com

Artificial Intelligence (AI) seems to have become the buzzword of the last decade and not without good reason. Some of the most notable technological advancements over the last 10 years can be attributed to emerging, robust (review link comment to Wikipedia)– many that have settled into our daily lives without us even noticing their arrival.

This is why it’s not uncommon for firms to leverage AI-focused workstations as a critical part of their AI workflows. AI workstations can handle all of the core AI processes that are needed for a variety of industries including engineering, manufacturing, medical, energy, higher education institutional research, earth sciences and more. The sky is truly the limit with workstations empowered by artificial intelligence and  , but this is only true when those AI workstations are developed with a thorough understanding of your business and vertical market. Firms deploying AI workstations need vendors that understand their unique workflow and business demands. The artificial intelligence partner needs to be able to digest this information and then utilize it to build a custom AI system that is comprehensively built for performance, speed, and reliability, and ideally one that is also delivered application-ready.

Even with the right artificial intelligence vendor, it’s usually best for the client to work through a checklist of their own to ensure that the system being built on their behalf will have the right specifications required to be successful in their unique situation.  The PSSC Labs AI team have compiled the following four important considerations that should be focused upon when working with a AI partner, like PSSC Labs, to develop the perfect AI workstation.

Your Checklist for Building AI Workstations

  1. Prepare Today for the Future of AI

Those working in the tech industry would say that the – see link comments is going to be all about its ability to support a complex and diverse business infrastructure with near full autonomy. In reality, this vision is still years away – but that doesn’t mean that firms aren’t making large strides towards that reality. AI practitioners should take a forward-thinking, focused approach to how they will utilize and implement their AI applications.  The PSSC Labs team has some excellent resources that can help support you and your organization with this process so email us today at sales@site-a.com to receive our latest AI Workstation future planning guide.

  1. Consider Existing and Future Artificial Intelligence Infrastructure Needs

AI workflows are best supported when they are built to be integrated with existing or future high-performance computing infrastructure. Keeping this in process in mind when building and deploying AI technologies using workstations that will be able to fully assist you and your organization with the focus of streamlining future transitions. As COVID-19 continues to disrupt the way that firms operate, these transitions between on premise systems, private cloud and public cloud are sure to come.

  1. Focus on Security

With COVID-19, we saw employees transferring their workstations into their at-home setup at recording breaking numbers. This shift in the way firms are working has meant that security has never been more important, particularly if the firm possesses customer or other sensitive data.

Most AI decision makers would agree that security, performance, and cost are the most important features to consider when choosing an AI system as demonstrated by a Gartner report on AI security published in 2020. We recommend that users make security capabilities a key priority. Afterwards, use cases can be assessed to balance between speed and cost. With the right AI/HPC partner like PSSC Labs, you shouldn’t have to sacrifice performance or security simply due to budget.

  1. Carefully Select the Right Components for Your AI Needs

The right components will make all the difference in the livelihood of your system. As the brains of the system, CPU selection continues to be one of the most important elements in your AI workstation. The Intel Xeon® Scalable Processor delivers truly breakthrough performance for handling big data tasks such as real-time analytics, virtualized infrastructure and high-performance computing.

In addition to advanced architecture, Intel Xeon® processors feature a rich suite of platform and security innovations for enhanced application performance, including Intel AVX-512, Intel Mesh Architecture, Intel QuickAssist, Intel Optane SSDs and Intel Omni-Path Fabric.

The AMD EPYC Rome Processor, based on the Zen 2 microarchitecture is another viable option. AMD touts this product as offering a very high performance per dollar in the marketplace, with significantly improved instructions per cycle over the first-generation processor, as well as more cores and better thermal efficiency. Additionally, the processor features a higher base frequency and a higher boost frequency than the previous AMD top-of-the-line processor.

While CPU selection is significant, one should also be mindful in selecting the right GPU option. One of our favorite GPU options is the NVIDIA A100, an all-around powerhouse built for speed and performance. The A100 is designed for scientific computing, graphics, and data analytics in data centers and touted as the “best data center GPU ever made” by NVIDIA CEO Jensen Huang.

With the NVIDIA A100, the right amount of computing power, memory, and scalability is delivered to tackle massive workloads. There are more than 54 billion transistors and it’s the world’s largest 7nm processor. The A100 can also efficiently scale to thousands of GPUs or, with NVIDIA Multi-Instance GPU (MIG) technology, be partitioned into seven GPU instances to accelerate workloads of all sizes. Read up on other GPUs to consider.

PSSC Labs takes a unique approach to developing custom Artificial Intelligence workstation solutions for organizations looking to deploy AI workstations. Our experts take your unique vision and business needs and then design a custom system architected with the latest CPU technology for your AI workflow. With a California-based support team of experienced AI engineers we are here for you and your organization. To discuss your situation in more detail and to receive a free quote for your customized AI workstation solution please call us today at 949-380-7288 or email us at 4sales@site-a.com.  We can not only provide you with a custom quote for your AI workstation but also discuss other possible workstation options you may wish to consider.