Aller Artix-7 FPGA Board with M.2 Interface Fits into a Laptop

A few days ago, we wrote about Nitefury M.2 card equipped with a Xilinx Artix-7 FPGA, and which you can connect to any laptop, board, a computer with a spare M.2 socket. It turns out Numato Lab has done something similar with the Aller board, specifically designed for development and integration of FPGA based accelerated features into other larger designs, and provided in a standard 2280 M.2 form factor M-key slot.

Aller FPGA board
Click to Enlarge

Aller FPGA M.2 card key features & specifications:

  • FPGA- Xilinx Artix-7 FPGA (XC7A100T-1FGG484C) with 101,440 Logic cells, ~126K Flip-flops, ~600KiB Block RAM, and 240 DSP slices
  • System Memory – 2Gbit DDR3 (MT41J128M16JT-125:KTR)
  • Storage – On-board 1Gb QSPI flash memory for FPGA configuration
  • Host Interface – 4 lane PCIe Gen1 (2.5GT/s) via M.2 Connector Interface, M-Key
  • Debugging – JTAG header for programming and debugging
  • Security – 1x Trusted Platform Module (AT97SC3205)
  • Misc – 100 MHZ CMOS oscillator, 1x RGB LED for custom use
  • Dimensions – 80 x 22 mm

The car leverages open source PCI Express IP cores such as litepcie to make the development of PCIe designs very cost-effective. Numato also provided an Aller PCI-Express x4 Test Project to quickly get started with the platform. You’ll find the source code here.

According to Rohit Sigh, Head of Development at the FPGA Division of Numato Lab, Aller was designed to help users with easier access to flexible and reconfigurable PCI Express infrastructure. With the decline of standard PCs and proliferation of M.2 connectors in Laptops, Servers and Desktop PCs, Aller breaks the barrier to PCI Express designs. Aller can also be used with standard PCI Express slots using M.2-to-PCI Express Adapter without any loss of functionality.

Beside allowing users to experiment with PCI Express based FPGA designs, Aller can also be used as hardware accelerator, for machine learning, and neural network engines, high hash-rate calculations for blockchain technologies, PCI Express kernel driver development training, digital signal processing, and development and testing of embedded processors.

The board is available for purchase at their online store for an introductory price of $399.95 per unit with volume discounts available on request.

Support CNX Software! Donate via PayPal or cryptocurrencies, become a Patron on Patreon, or buy review samples

Notify of
2 years ago

nvidia or amd should create graphic card in this form factor 🙂

6 months ago

All I want is a cheap amiga 4000 that runs morph os, web browsing, wifi and quake three and runs low power laptop and ssd and decent graphics chip set and could be made for less than 200 dollars.