
FPGA Acceleration of Convolutional Neural Networks (CNNs)
White Paper FPGA Acceleration of Convolutional Neural Networks Overview Convolutional Neural Networks (CNNs) have been shown to be extremely effective at complex image recognition problems.
Obsolete Product Notice:
This is an obsolete product and is no longer available for purchase. Contact BittWare for a recommended newer product.
The Terabox 200DE enables world-class FPGA acceleration to be deployed in the more challenging, harsh environments demanded of edge applications. The 200DE is NEBS compliant, 2U short-depth server with an extended operating temperature range based on the new DELL PowerEdge™ XE2420. It can be populated with a range of BittWare accelerator products featuring the latest FPGAs from Achronix, Intel and Xilinx to support the increasingly complex demands of highly specialized workloads.
Customers can purchase the TeraBox 200DE with FPGA accelerators pre-integrated directly from Dell Technologies via the Dell OEM program. The system includes Dell certification – a single comprehensive warranty covers server and BittWare FPGA accelerator cards. Dell can supply and support custom configurations.
For system management, BittWare’s FPGA cards are equipped with a Board Management Controller (BMC), which accepts IPMI 2.0 commands. Use it along with BittWare’s BittWorks II Toolkit to program the FPGA over USB, monitor card power and temperature, and re-program the onboard clocks. You’ll also be able to set points to shut down the card when it gets too hot, access JTAG, or access the software tools remotely.
Choosing a TeraBox FPGA server means knowing you are getting a pre-configured and tested solution. This includes setup and installation of your FPGA cards and associated hardware, your choice of operating system, and development tools. Your TeraBox arrives ready for use—giving your team more time for development and deployment.
The TeraBox 200DE supports many of BittWare’s Achronix, Intel and Xilinx FPGA-based PCIe cards. The table below lists system totals when populated with four cards*:
FPGA | Cards in Server | Memory | I/O | Processing | |
S7t-VG6 | Speedster7t | 2 | 4 banks GDDR6 (up to 8 GBytes) | 450 Gbits/sec 6x 100/50/40/25/10 GbE | 692K 6-input lookup tables (LUTs) 375 Mbits embedded RAM |
520N-MX | Stratix 10 MX | 2 | 2 banks DDR4 (up to 256 Gbytes) 2 banks QDRII+ (up to 1152 Mbits) 16 GBytes HBM2 | 250 Gbits/sec 4x 100/50/40/25/10 GbE | 2.1 million system logic elements 3,960 DSP blocks |
XUP-P3R | UltraScale+ VU9P | 2 | 4 banks DDR4 (up to 512 Gbytes) 4 banks QDRII+ (up to 2304 Mbits) | 250 Gbits/sec 4x 100/50/40/25/10 GbE Clock output, 1PPS output, trigger input/output | 2.5 million system logic cells Up to 6,840 DSP slices |
Our technical sales team is ready to provide availability and configuration information, or answer your technical questions.
"*" indicates required fields
White Paper FPGA Acceleration of Convolutional Neural Networks Overview Convolutional Neural Networks (CNNs) have been shown to be extremely effective at complex image recognition problems.
BittWare Partner IP NVMe Bridge Platform NVMe Intercept AXI-Stream Sandbox IP Computational storage devices (CSD) allow storage endpoints to provide computational storage functions (CSF) to
PCIe Gen4 data mover IP from Atomic Rules. Achieve up to 220 Gb/s using BittWare’s PCIe Gen4 cards, saving your development team when you need more performance than standard DMA. Features: DPDK and AXI standards, work with packets or any other data format, operate at any line rate up to 400 GbE.
BittWare Webinar Introducing VectorPath S7t-VG6 Accelerator Card Now available on demand: In this webinar, Achronix® and Bittware will discuss the growing trends of using PCIe