Products

We are a boutique systems integrator specializing in preparation for and execution of heterogeneous high performance cloud computing resources. Our product offering is flexible enough to engage your team consistently through ongoing support contracts or intermittently by serving offerings as needed. Every client and project is different. Talk to us about developing an offering that fits your needs.

Service Models

Self-Service

Usage Based Billing

We allow you access to our existing HPC Cloud cluster and are here to support when your team needs us. Our HPCaaS cluster can be configured to fit your needs using available documentation on top or mixing some Professional Services from our team when you need our help.

Onboarding Services

Phased Project Billing

We work with you to get up and running on the cloud and set you free. You can always reach out if you need anything and let us assist immediately or develop a monthly or annual management agreement that works for your team. Collaboration Engineering is a key component for project success and considered critical from the earliest phases of project development. If we do not have support from the organization we simply cannot move forward. This component is not broken out as a professional service offering as it integrates with our entire practice.

Initial phases include discovery and orientation to familiarize ourselves with your needs. This solution development work informs optimization within a deployment timeline. Discovery exposes a project to all parties so that we can collaborate on an approach. Orientation allows us to position ourselves as a part of your team so that we are able to stay on track and course correct when necessary. Deployment Timeline development provides a big picture for stakeholders while also provisioning goals and objectives for development and production teams.

We employ a number of resources in our toolkit that help us understand your needs. These include hardware, software, compute and human resources. Once discovery is complete we are able to architect, engineer, and deploy a system that is prepared for bench marking and tuning.

A System Architecture phase meets the expected challenges of the entire ecosystem that potentially effect the project. We dream with you to better understand where you want to be and suggest how to get there. There are overlaps between architecture, engineering and programming that offer opportunities to expose a development workflow and lifecycle that fits the needs of both stakeholder and development teams.

System Engineering phases develop as we progress through Architecting. Systems can be originally designed to house legacy products, with a road-map towards staged engineering milestones that allow for continuous operations. Avoiding workflow or operational interruptions is a primary goal of our designs. Deployment becomes a seamless operation that happens without disruption.

Staging deployment roll-out phases with a scaled approach helps mitigate issues that can arise when we start to implement change. We understand the cultural shifts that might occur when change is introduced. By following a phased project timeline we can ensure execution will be as smooth as possible.

Professional Service

Fluid Numerics provides consulting and software development resources that assist our clients in utilizing state-of-the-art high performance computing, scientific computing, and cloud computing technologies. We specialize in a variety of competencies including hardware and software infrastructure, scientific application development and performance tuning, and cloud system design, engineering, and operations.

We recommend these services to anyone who doesn't know where to start except that they need an Accelerated or High Performance Computing Cloud Solution. We have a team ready to plan something that will fit your team's needs from pilot through to go live. Our services incorporate proven change management strategies that help us make technology transition and transfer as seamless as possible. Now, let's say you've started down the HPC path and you're needing some "professional" help.

We have the experience necessary to assist you in many collateral and proprietary HPC and Cloud environments, allowing us to help you plan, maintain or recover an ecosystem that exists or needs to. Clients who are already up and running on the Cloud and interested in HPC or need support with an HPC Cloud resource are a perfect fit for these types of consultations. If this is you we’re here to help as issues arise.

More details are available about the following services below:

Managed Service or CHPCaaS (Cloud High Performance Computing as a Service)

Contract Billing

What does managed look like? We do it all for you, except pay the bill. Ongoing management and optimization of your HPC Cloud workloads.

Our fully transparent team integrates with your workflow to introduce and/or administrate your HPC Cloud services. We offer on-call as well as monthly and annual service level agreements. This approach includes treating each case and it’s unique needs individually. We pride ourselves on providing services that meet and exceed our client expectations for the term of their service agreement and beyond. Through provisioning training services and empowering our clients with proper system documentation we are able to disengage when a client is ready to handle their system independently and gladly do so.

Of course if you’d like to mix and match services we are certainly open to you doing so. Contact us to discuss your project with us further.

Our Process

Engagements begin with discovery, where we gain opportunities to learn about our client’s goals and objectives. Through discovery phase analysis, key performance indicators (KPIs) are established in order to measure these goals and objectives. During the discovery phase we also start to understand and develop a roadmap for additional work. Using information gathered in discovery, we are able to conduct benchmarking exercises throughout future product development phases that inform a view of the project that correlates with goals and objectives.

Post-discovery, we enable work execution in close collaboration with client stakeholders and relevant staff through an academic and educational approach. Ultimately, we want our clients to be able to independently and internally support our contributions after we complete projects. We are capable of providing ongoing/managed support whenever necessary.

Service Level Agreements

24/7

Allowable Contracts

Annual, Bundled + Bolt-Ons

Benefits

24/7 support. FMR within 24 hours.

Provides access to virtual system administration, Cloud-HPC Consulting, and Gcloud engineering. Additional hourly services are “bolt-on services” offered at a discounted hourly rate.


9-5

Allowable Contracts

Annual, Bundled + Bolt-Ons

Benefits

Business hours support. FMR within 2 business days.

Provides access to virtual system administration, Cloud-HPC Consulting, and Gcloud engineering. Additional hourly services are “bolt-on services” offered at a discounted hourly rate.

20hr

Allowable Contracts

Quarterly, Bolt-Ons, Fixed Deliverable

Benefits

Part time services

Hourly

Allowable Contracts

Month-to-Month, Fixed Deliverable


Pricing

We believe that every cloud application has unique needs and there is no “one size fits all” pricing and contract model to address everyone’s needs. Like our Cloud and HPC solutions, our pricing schedules and contracting arrangements are customizable to fit your application’s exact needs. Pricing schedules range over “hourly a la carte”, minimum spend consulting retainers, longer term “discovery-development-support” packages or a hybrid that makes sense for you..

Professional Service Details

Project Discovery

Traditional HPC procurements can take months to years to get off the ground. Cloud platforms make it possible to develop a flexible system continuously at scale to utilize resources at maximum capacity. They also provide access to a variety of hardware including various CPU models, GPUs, and TPUs. The cloud paradigm allows you to rapidly experiment with parallelizing and scaling your HPC application. With cloud, you pay only for the compute resources you use. Sure, all of these attributes are beneficial for your HPC development activities, but, how do you get started ?

Project discovery is the start of all of our project engagements. In some cases, clients already have an idea of their cloud needs but need our help getting the project off the ground. You provide us your specific and detailed needs and we develop, package and deploy a Cloud-Native or Cloud-Hybrid solution to support them. You may be thinking about entering the HPC world for your specific domain or industry or you are looking to expand your compute resources with cloud infrastructure. Engage us to discover, document and deliver a solution optimized for your application.

HPC/Cloud-HPC Consulting

HPC and Cloud-HPC Consulting is a service available at any time as an hourly bolt on to any service package or as a stand-alone service. Cloud-HPC consulting is a service that provides subject matter expertise for addressing concrete system and software design, engineering, and operations questions. From the moment you engage us for a consultation, we carefully curate feedback that is consistent with scientific and industry best practices. This service fits best for functional teams looking to make a further reach into new and unfamiliar technology and software.

Virtual System Administration

Virtual System Administration provides you access to expert support staff to monitor, operate, and maintain the custom cloud-based systems delivered by Fluid Numerics. In addition, we provide a support ticket service integration into slack for your team to quickly submit support tickets and engage our team. We firmly believe in a separation of duties. You should be able to focus on your scientific and HPC passions, while we keep systems up and running for you.

HPC-Cloud Architecting & Engineering

HPC-Cloud Architecting & Engineering services engage an HPC Specialist and HPC-Cloud Developer to develop and implement a fully customized Cloud-Native or Cloud-Hybrid system. With this service, we will engage with you and your team transparently and productively to design and deliver a cloud based HPC platform tailored to your applications and workflows. Let us help you reduce the time-to-science.

Software Parallelization

Software parallelization is often a crucial step in meeting performance goals for an application. We help you develop the appropriate parallelization strategy for your application and identify the best suited standards, libraries, and APIs to implement this parallelism. Once a strategy is developed, we pair up with your team to rapidly develop a minimum viable product. Our software development strategy includes an educational component as we work closely with your team. As with all of our services, our goal is to enable your team to support the software we help develop.

GPU Acceleration

GPU Acceleration can help reduce time-to-science, energy costs, and cloud spending for many compute intensive scientific applications. The GPU Acceleration service gives you access to an expert General Purpose GPU software developer accelerate your application. This service is typically engaged once you have a porting strategy and an estimate of GPU acceleration payoffs. Once a strategy is developed, we pair up with your team to rapidly develop a minimum viable product. Our software development strategy includes an educational component as we work closely with your team. As with all of our services, our goal is to enable your team to support the software we help develop.

CPU Profiling and Performance Tuning

CPU profiling focused and performance tuning services are available to help you better understand your CPU-only applications bottlenecks. Complex scientific applications, both serial and parallel, often under-utilize compute hardware and can result in overspending on cloud platforms or prevent you from meeting your performance goals. Our profiling strategy identifies hotspots, potential causes for under-performing applications, and the theoretical peak performance on target hardware. Profiling provides the necessary information to engage in performance tuning. Performance tuning activities involve code refactoring aimed at enhancing application performance while maintaining correctness.

GPU Profiling and Performance Tuning

GPU profiling focused and performance tuning services are available to help you better understand your GPU-accelerated applications bottlenecks. Complex scientific applications, both serial and parallel, often under-utilize compute hardware and can result in overspending on cloud platforms or prevent you from meeting your performance goals. Common struggles on GPU accelerated platform come from GPU and CPU memory management issues in addition to memory handling on many-core architectures. Our profiling strategy identifies hotspots, potential causes for under-performing applications, and the theoretical peak performance on target hardware. Profiling provides the necessary information to engage in performance tuning. Performance tuning activities involve code refactoring aimed at enhancing application performance while maintaining correctness.

Application Porting and Benchmarking

Application porting and benchmarking is an application “lift and shift” service where we help migrate your application build and run system to new platforms. Typically, this activity begins after a thorough discovery of compiler and software dependency requirements. We will deploy your existing, validated, and verified applications to new platforms on your behalf. Additionally, once the application is migrated, we can execute a canned benchmark with your application. We complete this activity with a summary of the porting strategy and the benchmark results. This allows to quickly find out whether or not a platform migration will be worth the effort for your team.