Why migrate to the Cloud?
Scalability – take advantage of elastic compute resources to efficiently manage large workloads or meet changing demand – on AWS (Amazon Web Services), Google Cloud or Microsoft Azure
Agility – tailor virtual machine (VM) sizes to your workloads: CPU, memory, networking, GPU accelerators can all be configured to match the your workload requirements
Unlimited Storage – scale different types of storage as you need it, at the locations of your choice
Predictable Cost – realize the benefits of the cloud without breaking the bank
… however …
Cloud HPC Migration Is Challenging
Moving existing HPC workflows to the cloud is non-trivial, and implementation mistakes can easily eliminate your performance and cost benefits. HPC leaders need support with getting started, choosing optimal hardware, managing costs, storage configuration, and tuning existing code to exploit the potential of the cloud.
The flexibility of the cloud allows the option to adapt both the application as well as the platform upon which it runs. Optimal resource selection is often not obvious and maybe counter-intuitive at times.
… and that is where we come …
Why choose SSG as your Cloud migration partner?
We can assist you with:
Choosing Your Cloud Provider – we work and partner with Google, AWS and Microsoft, SSG knows the strengths of the top Cloud Service Providers and can help you select the best for your needs
Navigating the Transition – SSG helps you understand which factors will limit performance and increase cost in the cloud (I/O, CPU, memory, data transfer)
Minimizing Cost – SSG optimizes both your applications and the cloud infrastructure they can run on. Tuning and modernizing existing code, choosing the right resources – all to reduce overall cost and/or increase performance.
Software Optimization and Code Modernization
- Is your application software suffering from poor performance today?
- Do you know if your application software is running at full potential on your current computing systems?
- Is your application software ready to exploit future computing technology?
The overwhelming proportion of performance from all mainstream current and future processors comes from parallel processing techniques. In addition, system architectures have increasingly complex memory hierarchies and performance characteristics. Software must be modified, or ‘modernized’ in order to exploit this hardware complexity to access the potential performance.
Software modernization is often a required step for an application to be able to effectively use GPUs (e.g., from NVIDIA or AMD),or Intel’s Xeon Phi, and other many-core processors. However, software modernization also delivers significant performance advantages on more traditional processors, such as Intel’s Xeon family.
SSG can help you improve the cost-effectiveness and performance of your numerical computing in science, engineering, finance, analytics and more.
Algorithm Design and Development
SSG designs and builds custom algorithms to meet your needs and solve your complex problems. SSG has been at the forefront of algorithm design and optimization for over 2 decades. Our experts collaborate with world leading research centers and academics to develop and maintain algorithms used in multitude of applications.
SSG’s custom algorithm development targets the configurations of your specific environment and your own problem types. Many aspects are considered including processor (CPU, coprocessor or GPU), memory configurations, communications networks and problem sizes, runtime windows, tolerances, areas of interest and much more.
Software Porting and Tuning
SSG can assess your code to identify the likely value of porting to GPU or manycore processors. Where performance or cost benefits are possible, SSG can port your code to GPUs or other manycore coprocessors using appropriate standards and languages - e.g. CUDA, OpenCL or OpenMP. We can also advise on appropriate alternative algorithms that may be better suited for use on manycore/coprocessor/GPU processors.