Arm has released an exciting new addition to the DesignStart family, available on the cloud! The vibrant DesignStart community of academic and commercial system developers can now access a new prototyping platform on the Amazon Web Services (AWS) cloud, enabling the development of software, and, soon, hardware design as well, for next-generation IoT devices.
The first ‘DesignStart FPGA on Cloud’ product, an Arm Cortex-M33 subsystem, on the Amazon Elastic Compute Cloud’s FPGA instances (Amazon EC2 F1), provides a seamless user experience and makes software development for Cortex-M33 easier, faster and more accessible. This is the first release, which we plan to enhance with hardware accelerator addition capabilities and transform it from a software development platform to a complete hardware/software design environment.
The DesignStart program has enjoyed enormous success since its launch a decade ago, and has become synonymous with being the easiest path for developing, prototyping and taping out an Arm-based embedded system-on-chip (SoC). With Arm Cortex-M0 and Cortex-M3 processors at the heart of a complete Arm-based system, and the recent announcement of the FPGA-optimized Cortex-M1 and Cortex-M3 CPU soft IP on Xilinx FPGAs, DesignStart gives academia and start-ups everything they need to get started with SoC development built on a proven foundation. With more than 3,000 DesignStart CPU prototyping downloads in the last 12 months alone, Arm is now taking an extra step to unlock exciting new use cases by providing developers access to one of the latest Cortex-M processors, the Cortex-M33 processor, in a cloud-based FPGA system available on AWS Marketplace. Developers can have instant access, and the ability to design IoT solutions that span hardware, embedded and cloud software environment. Companies and academia benefit from the unprecedented scale to design and simulate their entire IoT software and/or chips in the cloud, accelerating development and deployment of complete IoT solutions.
Cortex-M0 and Cortex-M3 adoption in DesignStart
Cloud services are one of the fastest growing technology sectors. According to Gartner, worldwide cloud services revenue is set to almost double in just four years, from $153 billion in 2017 to $302 billion in 2021. This rapid cloud growth is bringing a ripple effect in multiple industries, including System-on-Chip (SoC) designs. Amazon Web Services (AWS) is the first public cloud provider to introduce FPGA as a service in their cloud offerings. FPGA can be used to make SoC prototyping easier, faster and without upfront infrastructure costs. To take advantage of this new service, Arm is partnering with AWS to bring developers DesignStart FPGA on Cloud. This is a new cloud-based platform that will pave the way for hardware/software co-design and FPGA-based SoC prototyping, using the Amazon EC2 F1 instances. F1 instances consist of a host CPU connected to up to eight Xilinx VU9P FPGAs. They are easy to program and come with an FPGA developer AMI that includes Xilinx’s Vivado tools, at no additional software charge, that allows developers to compile, simulate and debug hardware accelerators on the cloud.
Using AWS infrastructure provides an easy to use, FPGA-based software development platform demonstrating realistic performance for Arm processors, even before they become widely available on physical development boards. What’s more, Arm charges no access fees to system developers for any of the Arm deliverables, so users will only be subject to the standard AWS EC2 F1 cloud usage fees. Research and education users can benefit from Amazon’s programs ‘AWS Cloud Credits for Research’ and ‘AWS Educate’, which provide credits that enable educators and researchers to work in the cloud, innovating more quickly and at lower cost.
Cortex-M33 is the first Arm processor to be made available on the Cloud, and for good reason. As one of Arm’s latest M-class processors, it is fully optimized for next-generation embedded and IoT devices, with security being built in to the heart of the CPU. It is very area efficient, making it the ideal option for low-cost embedded devices, while it deploys Armv8-M architecture and 32-bit instruction set with floating point and DSP capabilities for complex applications.
Cortex-M33 unlocks a range of interesting new capabilities for designers, including Machine Learning inference on the edge. Machine Learning applications are moving to edge IoT devices to ensure the best user experience, greatest responsiveness and optimum use of network resources. Cortex-M33’s DSP capabilities make it the right processor to efficiently run optimised ML IoT applications. Security is also a significant priority when designing and implementing the vision of a connected, smart world. Cortex-M33 fully supports Arm TrustZone, enabling software and hardware isolation across the CPU and SoC system. New Cortex-M33 based device families have recently been announced, so any software prototyping work that takes place on the Cloud can soon be ported to new physical platforms too.
FPGA on Cloud includes the digital subsystem around the Cortex-M33, Corelink SSE-200. This reference system, CoreLink SSE-200, is also end-to-end TrustZone-enabled, built with the Arm Platform Security Architecture (PSA) guidelines in mind. CoreLink SSE-200 consists of two Cortex-M33 instances, allowing for additional processing or even physical separation between secure and non-secure applications, system interconnect and peripherals. Security isolation is an important aspect of PSA (a framework for securing one trillion connected devices), which notes the importance of secure system partitioning. Such separation can be achieved using a single processor with TrustZone, or two processors if preferred. By using SSE-200, developers have a great starting point, with many of the security basics taken care of.
It’s easy to get started – simply click the link below, or visit AWS Marketplace and search for 'DesignStart FPGA on Cloud'. This solution includes all the software and FPGA images that you need to load to an Amazon EC2 F1 instance in order to get started. With AWS Marketplace developers can start prototyping their software for next generation IoT devices within a few minutes!
As for hardware developers, you may not be able to modify the first available FPGA image, but there’s good news coming soon. Early next year, Arm will be releasing an extendable version of CoreLink SSE-200 on Amazon EC2 F1 instances. The entire design consumes about 30% of a single FPGA on F1, leaving developers with about 70% of available custom logic area to add their own hardware functionality and create a truly custom SoC design. This will allow developers to add their own hardware peripherals and accelerators to the CoreLink SSE-200 system, simulate them, create their own FPGA image of the complete system and prototype this on AWS FPGAs.
Access DesignStart FPGA on Cloud