Fast application acceleration on FPGAs
Modern cloud providers and enterprise data centers are upgrading to support the growth of cognitive era applications that are key to business strategic initiatives. Whether it is improving portfolio risk modeling at a financial institution or an AI algorithm for image analysis, the large growth in data that feeds modern applications makes parallel computing with a mix of specialized hardware such as high-core CPUs, GPUs, and FPGAs essential.
Hardware developers with the expertise to deploy applications across heterogeneous platforms are scarce as the industry has seen a steep rise in the number of talented software developers and data scientist. To enable acceleration to software developers, Merlin™ Compiler provides a familiar software flow with an automated C/C++ to FPGA. Merlin bridges the gap between Cognitive Era applications and specialized hardware enabling software developers to enhance their application performance on heterogeneous platforms.
Merlin Speed-Up on Xilinx FPGA
vs. 16-Core CPU
Merlin Compiler takes an automated approach to FPGA acceleration. Current acceleration flows require partitioning and rewriting algorithms in OpenCL or lower level hardware description languages. Merlin Compiler provides an “OpenMP” approach from C/C++ to FPGA resulting in up to 10X productivity advantage.
Merlin Compiler is available on public and private clouds and on-prem on your choice of Intel and Xilinx FPGA platforms.
Falcon’s Software Flow for
- Fine-grained data-path optimizations
- Coarse-grained dataflow generation
- Memory access optimizations
Get started without
- No FPGA expertise required
- Auto HW/SW interface generation
- Pure C/C++ flow
- ML powered optimization
- Auto-generated pragmas
- Auto source to source optimization