Arria 10

"Intel® Arria® 10 FPGAs deliver more than a speed grade faster core performance and up to a 20% f_MAX advantage compared to the competition, using publicly-available OpenCore® designs.^† Arria 10 FPGAs and SoCs are up to 40 percent lower power than previous generation FPGAs and SoCs and feature the industry’s only hard floating-point digital signal processing (DSP) blocks with speeds up to 1.5 tera floating-point operations per second (TFLOPS)^†.

• One speed grade faster than competing FPGAs and SoCs - View OpenCore Benchmark
• 1.5 TFLOPS of DSP performance with IEEE 754 compliant hard floating point blocks
• 1.5 GHz dual-core ARM*-based CPU - industry's only 20 nm SoC
• Highest performance 2,400 Mbps DDR4 SDRAM memory interface
• The industry’s only midrange FPGA with 25.78 Gbps transceivers
• 96 transceiver lanes deliver 3.3 Tbps of serial bandwidth

Additional features include:

• Fast compile times and advanced design environment
• Best-in-class IP cores including 100G Ethernet, 100G Interlaken, Hybrid Memory Cube, and PCIe Gen3, with 2X performance and lower latency compared to previous generations
• C-based design entry using the Intel^® FPGA SDK for OpenCL™, offering a design environment that is easy to implement on FPGAs
• System-level design environment with Intel Qsys system integration tool
• DSP Builder for Intel FPGAs–A model-based DSP environment within the MATLAB*/Simulink* environment

The Arria 10 FPGAs and SoCs are ideal for a broad array of applications such as communications, data center, military, broadcast, automotive, and other end markets."

A development kit runs about $4500.

https://www.altera.com/products/fpga/arria-series/arria-10/overview.html

https://www.altera.com/products/boards_and_kits/dev-kits/altera/kit-a10-gx-fpga.html

APPLICATIONS

An OpenCL(TM) Deep Learning Accelerator on Arria 10 - http://hgpu.org/?p=16915