Intelligent Computing Framework


Benchmark data on application and hardware performance across CSPs and available hardware types and configurations


Infrastructure capacity and reliability data across CSP regions and hardware types


Pricing data across CSPs, hardware types, and service levels


Environmental impact across available global data center infrastructure


Rescale provides users with a data-driven framework to make informed decisions on architectural choices that optimize any workload across a range of objectives such as increasing workload speed, decreasing workload cost, reducing carbon footprint, and improving business continuity. With these intelligence metrics, CIOs and IT / HPC managers can gain insight into the full economic impact of their decisions. Having visibility and control over these factors enable organizations to define policies and best-practices that help them accelerate time to answer, minimize SW/HW costs, or simply get more out of their existing budgets. 

This unique and holistic framework, known as the Rescale Intelligent Computing Framework, consists of four key intelligence metrics:

  • Performance Intelligence: A composite score of an architecture’s computation performance that factors in common workload benchmarks at various scales to help organizations choose the best hardware for each workload. Optimizing for performance plays a critical role in optimizing time to solution, job cost, and the ability to tackle larger problems. See the figure below for an example where a specific cloud provider provides optimal performance and value while also offering competitive capacity.
  • Maturity Intelligence: A hardware rating on an architecture and its infrastructure provider based on its up-time track record, capacity, and volume of core hours under operation. Optimizing for maturity helps organizations ensure consistent infrastructure service delivery and business continuity. See the figure below for an example where a specific cloud provider provides optimal maturity while also offering competitive performance.
  • Value Intelligence: A ratio of the performance score divided by price of the architecture type, which can vary across geographies, regions, and cloud providers. Optimizing for value can reduce the costs from computing, licensing, and overall cost of computing operations. See the figure below for an example where a specific cloud provider provides optimal value while also offering competitive performance.

Taking a closer look at how value can be affected by price fluctuations across regions in addition to cloud providers, the charts below how the on-demand infrastructure price for each instance type can vary dramatically. Other infrastructure service levels such as spot market and reserved instances further complicate comparisons.

  • Sustainability Intelligence: A designation that indicates when an architecture is optimized for reduced environmental impact e.g. reduced carbon emissions due to using electricity from renewable energy sources. This data point is the latest addition to the Rescale Intelligent Computing Framework and will feature more benchmarks coming soon as we make more benchmarks available.

Related: See Intelligent Resource Automation to learn more about how Rescale enables administrators to prioritize how jobs are run in order to save costs and maximize the efficiency of their workloads.