Solutions for High Performance Computing

  1. Overview
  2. System Synthesis
  3. Processor Synthesis
  4. Virtual Platform

HPC Overview

High Performance Computing has been taking several avatars at different times based on the emerging requirements. A common requirement is the need for performing large number of computations and processing large amounts of data. A key problem is the amount of space and power needed to house dozens of racks of servers to achieve the desired performance levels.

Use Teraptor to model specialized instruction set architectures, stream architectures, network on chip and explore the system level architectures to identify the best architecture for optimally achieving the performance for a specific application or a general class of applications. Once you are satisfied with the system and processor architecture model, automatically synthesize your design using Teraptor VHLS engine.

HPC - Challenges

  • Continuous Advances in Technology
    • Impact On Entire Value Chain
    • Continuous Increase In Performance Requirements
    • Programmability
  • Need for New Products and Solutions
    • New Product Design is Expensive!
    • Entire Supply Chain is Stressed

Imperatives

  • Build General Purpose Platforms and Application Specific Solutions
  • Make it easier to innovate in an application specific manner
  • Automate development
  • Apply The Right Development Tools and Platforms
  • Keep Innovating

Solution - Progressive Compute ArchitectureTM

Use PCA to build petaflop grade Super computers, with significantly smaller number of nodes cutting down on power consumption

Teraptor enables the design of 'Progressive Super Computing Architectures' and reduce the number of data moving between arithmetic units within a single core.

Next Steps

Do Contact Us - to request a proposal or to learn more.