Bitfusion announced that Amazon Web Services (AWS) customers can deploy deep learning workspaces and inference on AWS F1 FPGA instances with Bitfusion Flex.
Bitfusion Flex now includes everything one needs to manage AI infrastructure as well as develop, train and deploy deep learning applications across CPUs, GPUs and FPGAs. Developers and data scientists can develop and train their deep learning models on AWS directly from their laptop as well as deploy inference to Xilinx FPGAs on AWS and scale them without knowing the details of FPGA programming or AWS infrastructure provisioning.
Application performance demands have increasingly been outpacing Moore’s law in a variety of fields, particularly deep learning and AI.,” said Subbu Rama, co-founder and CEO of Bitfusion. “Re-configurable chips like FPGAs, can morph to the application at hand, providing immense speedup compared traditional compute architectures. With Bitfusion Flex, AI developers can now take advantage of CPUs, GPUs for development, training of deep learning applications and F1 instances on AWS powered by Xilinx FPGAs for inference deployment right from their laptop, without having to deal with any infrastructure details.”
Bitfusion Flex operates via an intuitive command line interface (CLI) directly from a user laptop. It only takes a few simple steps along with the user’s AWS credentials to get the CLI up and running and to create FPGA workspaces or to deploy inference on AWS F1 FPGA instances.
The momentum behind FPGA acceleration in the cloud is prevalent across many important applications like data analytics, genomics, video processing, and machine learning,” said Andy Walsh, senior director of strategic market development at Xilinx. “Bitfusion Flex will play an important role in enabling developers to infuse machine learning capabilities into cloud-based applications using Xilinx FPGA acceleration.”
Sign up for the free insideBIGDATA newsletter.
Speak Your Mind