Rescale - Industry Solutions
From conception to production, the design of consumer products requires both a variety of development resources as well as an adaptable environment in which to innovate. High performance computing hardware and software is needed to run materials analyses on bicycle frames, fatigue analyses on tennis shoes, and vortex modeling of a spinning soccer ball. Consumer product companies worldwide are accessing the immense compute capabilities of Rescale's platforms. Here is how one bicycle manufacturer optimized performance using the ScaleX Pro platform.
STAR-CCM+ volume rendering of drafting bicycle rider simulation
"With Rescale's ScaleX Pro platform, we can build complex STAR-CCM+ and HEEDS simulations, and run analyses much faster than we can internally - drastically improving our simulation process." - Chad Manuell, Director of Engineering, Trek Bicycle"
For leading bicycle designer, Trek, the mission has always been simple: build the best bikes in the world. For complex coupled analyses, Trek turns to Rescale to:
- Run coupled, complex Star-CCM+ / HEEDS optimization analyses leveraging existing software licenses for reduced cost
- Significantly reduce simulation runtime by using the latest HPC resources available in Rescale's 30+ advanced, global data centers
- Fully explore the experiment design space to make informed decisions about drafting techniques related to competitive bicycling
From chemical kinetics for combustion analysis to intake and exhaust simulations to piston deformation, automotive engineers run an expansive range of computer-aided engineering (CAE) simulations on Rescale's platforms. With the ability to instantly scale the hardware regardless of the workload size, automotive companies are able to stay on the leading edge without the leading edge costs. Below is how an engine manufacturer customer deployed Rescale's ScaleX Enterprise.
Four-stroke engine model simulation by Pinnacle Engines
"Running jobs on Rescale allows us to reduce capital expenses while still running the simulations needed to design industry-leading engines." - Monty Cleeves, Pinnacle Engines CTO
Pinnacle Engines is a game-changing engine company developing and commercializing an ultra-efficient engine architecture. To instantly burst beyond their internal resources, Pinnacle Engines leverages Rescale's cloud simulation platform to:
- Instantly execute 240+ customized HPC simulations in parallel to analyze an internal combustion engine
- Reduce the time to results by 80% by running multiple jobs in parallel
- Achieve greater than 73% cost savings compared to the cost of expanding internal resources
Flow trajectories in the front pipe of the exhaust system showing Adblue droplet size and ammonia distribution on substrates
"Our higher productivity will enable us to handle more projects. The Rescale platform makes us more efficient and more attractive to our customers, which then brings us more business and more development projects." - Kasper Steen Andersen, CAE Manager, Dinex
Dinex A/S is a multinational designer and manufacturer of exhaust and emissions systems for trucks, buses, and off-road equipment. Dinex used Rescale’s ScaleX to design an exhaust treatment system for a large Chinese truck manufacturer. Using Rescale, Dinex achieved:
- 4x speedup in simulation runtime
- 25% reduction in time-to-market, reducing project time from 18 months to 13.5 months
- Higher project throughput and higher revenue from additional projects
- More productive simulation engineers from elimination of simulation bottleneck
- Increased marketability to customers from their use of industry-leading technologies
Aerospace engineering requires significant compute resources to perform massive scale structural, fluid dynamics, thermal, and electrical analyses. Rescale's customers leverage the latest technology to analyze aerospace applications, ranging from complex turbine analyses to analyzing the structural integrity of wing designs to modelling propulsion flow. See how one of Rescale's aerospace clients leverages Rescale's ScaleX Pro platform for ITAR sensitive simulations. ScaleX Pro platform for ITAR sensitive applications.
Fluid-structure interaction model of large telescope array using XFlow
"Thanks to Rescale's HPC cloud platform, we are now able to run very complex multi-physics analysis with unprecedented high fidelity models. What takes several hours on Rescale's cluster would take a week on our workgroup analysis server." - William Villers, TEN TECH LLC Director of Engineering
As an ITAR certified company, TEN TECH LLC frequently services the highly confidential and secure sector of military and government organizations. Using Rescale's industry-leading security environment, TEN TECH LLC can:
- Execute flow analysis using on-demand XFlow simulation software and Rescale's security-leading cloud environment
- Reduce runtime by over 90% and quickly scale ITAR HPC resources to suit varying simulation need
- Instantly access over 1TB SSD disk storage using a high performance hardware configuration
Scientists are running compute intensive genetic engineering analyses, bio-sequencing simulations and more, leveraging Rescale's simulation platforms. Using optimized file transfer technology, scientists can transfer the large quantities of data needed to perform molecular docking and quantum chemistry analyses to the Rescale's platform. ScaleX Pro and ScaleX Enterprise feature built-in collaboration tools to increase productivity and allow scientists to better isolate genetic traits, develop improved antibiotics, and predict fire flow patterns. Here is how a biotechnology researcher used Rescale's cloud simulation platform.
"I was very impressed with the runtime performance. This massive productivity gain helped my research efforts tremendously." - Eric Lee, master's candidate, University of Kansas
University of Kansas
A researcher in the area of biotechnology at the University of Kansas with a focus on nanofluidics requires large computational resources to handle the massive amount of data needed for his nanowetting analyses. Using the open source molecular dynamics software LAAMPS, researcher Eric Lee was able to:
- Drastically reduce runtime from 4 days down to less than 12 hours
- Receive quick technical support for simulation setup from Rescale's industry-expert technical team
- Customize HPC resources to his specific molecular simulation needs
Oil & Gas
Scalability and flexibility. These are just two features of Rescale's simulation platforms that help petroleum engineers define reservoir simulation workflows and geologists create analysis workflows to help identify hydrocarbon traps. Engineers use ScaleX Enterprise and ScaleX Pro to perform simulations on drillings, reservoir geomechanics, and deepwater systems, among others. Using Rescale simulation workflows, engineers can consider all possible options and calculate adjustment sensitivities made to highly complex systems, ensuring optimal field production. Below is how one energy client utilizes Rescale's platforms.
"I was skeptical about using the cloud, however, without Rescale's simulation platforms, we never would have met our project deadline. I look forward to using Rescale's service in the future." - Senior petroleum engineer
In the area of oilfield services, a major global oil and gas supplier needed to increase hardware capacity in a short period of time. With offices and engineers stationed all around the world, they needed a flexible system to handle complex computational fluid dynamics (CFD) simulations for reservoir modeling while easily managing data across international teams. This oilfield service company used Rescale's platforms to:
- Instantly scale out HPC hardware by 8X for a single simulation
- Reduce simulation runtime by over 75% for CFD analyses
- Store 20+TB of simulation data in Rescale's secure storage system
While engineering simulations require HPC resources, it can't be overlooked that the computing hardware itself was designed and engineered as well. Electrical engineers are utilizing Rescale's simulation platforms to develop powerful and robust printed-circuit boards, integrated circuits, cooling equipment, and power storage, among others. With natively integrated design of experiment and optimization tools, electrical engineers can run complex thermal analyses or minimize noise margins easily from any computer. Here is how a large electronics enterprise optimizes simulations with Rescale's platforms.
"Before ScaleX Pro, even if we had a sudden increase, we were limited to the software licenses we had previously purchased. With Rescale's platform, we can now access extra licenses at any time." - Manager, Electrical Power
An enterprise company in the technology space works with Rescale to help optimize and improve the time to market on communications equipment. With the flexibility to scale both hardware and software at any time in the design cycle, electrical engineers are able to:
- Quickly access both open source and commercial software tools through one platform
- Reduce capital and operating expenses by 30% over a 1-year period
- Share simulations with colleagues and simultaneously clone jobs to make fast design iterations
Enterprise companies are not the only organizations requiring expandable compute capacity for research. Academic institutions often experience lengthy internal queues and limited resources that are not necessarily optimized to handle engineering and science simulation workloads. Through Rescale, researchers, students, and professors have access to the latest computing technology. Flexible pricing tiers and the ability to use existing software licenses provide academic members with a cloud simulation platform that is capable of handling all types of needs.
"We are, therefore, very happy with the present collaboration with Rescale, which helped us tremendously by providing a system that worked perfectly for us, right out of the box." - Stijn Pennings, Team Manager
The team from Delft University in The Netherlands has been competing for over 10 years in the Formula Student Competition, and, in 2014, placed 1st in the Overall Class 1 Division. They use the ScaleX Pro platform to significantly improve racecar design decisions, as well as:
- Achieve 6x reduction in turnaround time by taking simulations that used to run for 24+ hours and reducing them down to less than 4 hours
- Run multiple jobs in parallel without having to wait in lengthy queues that they were experiencing internally
- Optimize simulation configurations during the final design stages by running custom HPC hardware
Machine learning R&D, especially in the field of Deep Learning, attempts to mimic the neural networks of the brain to complete human-level tasks such as image recognition and natural language processing. Deep Learning is extremely compute-intensive and requires state-of-the-art graphics processors. Research is typically constrained by expensive specialty hardware. Rescale's platform supports popular deep learning software packages and provides unlimited, pay-as-you-go access to the latest GPUs so organizations can stay on the cutting edge of R&D. An intuitive, web-based workflow and excellent customer support make setup and execution easy.
Rescale is optimized for machine learning on the cloud with:
- HPC hardware: K80 GPUs, Skylake processors, InfiniBand interconnect, 1TB+ RAM
- State-of-the-art machine learning software: All the latest versions of popular libraries, including TensorFlow, Keras, Caffe, Torch, Chainer, Nolearn, and Pylearn. All configured to run on GPUs out-of-the-box.
- Flexible interactive and batch training interfaces: Interactive modeling and visualization in Jupyter notebooks, automated model batch training via Rescale's API
- Tools for large datasets: File transfer tools for efficiently uploading large training datasets, fast and easy access to many public datasets
Setting up your algorithms and configuring hardware is quick and easy with Rescale. Just upload your input file, select your software, select your hardware, and you're up and running in minutes.
Learn how to run Torch and TensorFlow on the cloud with Rescale:
View TensorFlow Tutorial
View Torch Tutorial