AMD Expands Alveo Portfolio with World’s Fastest Electronic Transaction Accelerator Card in a Thin Form Factor
— AMD Alveo UL3422 Accelerator Card Gives High-Frequency Traders an Edge in the Race for Fastest Trade Execution While Lowering Barriers to Entry —
SANTA CLARA, Calif., Oct. 14, 2024 – AMD (NASDAQ: AMD) today announced the AMD Alveo™ UL3422 accelerator card, the latest addition to its record-breaking accelerator family1 designed for ultra-low latency electronic trading applications. The AMD Alveo UL3422 provides traders, market makers and financial institutions with a slim, rack-space and cost-optimized accelerator card designed to be quickly deployed in a variety of servers.
Powered by AMD Virtex™ UltraScale+™ FPGAs, the Alveo UL3422 accelerator card features a novel transceiver architecture with hardened and optimized networking cores, custom-built for high-speed trading. It enables ultra-low latency trade execution, with sub-3 nanosecond FPGA transceiver latency and breakthrough “tick-to-trade” performance that is unachievable with standard off-the-shelf FPGAs1 .
“In the increasingly competitive world of high-speed trading, speed is the ultimate advantage. The Alveo UL3422 card provides a lower cost entry point while still delivering cutting-edge latency performance, making it viable for adoption by companies of all sizes looking to stay competitive in the ultra-low latency trading space.”
—— Yousef Khalilollahi
Vice President and General Manager of the Adaptive Computing Group at AMD
New slim form factor for cost-effective deployment
The Alveo UL3422 accelerator card is available in a slim full-height, half-length (FHHL) form factor and is designed to fit into a variety of server and co-located switch data centers.
The Alveo UL3422 accelerator card reduces port density, onboard memory, and connectivity options compared to its predecessor, but still uses the same AMD Virtex UltraScale+ VU2P FPGA to achieve ultra-low latency.
As a result, the Alveo UL3422 delivers equivalent performance in half the size of the existing Alveo UL3524 accelerator card, which holds the current STAC-T0 benchmark tick-to-trade performance world record.1 The Alveo UL3422 's slim, full-height, half - length form factor enables financial institutions to cost-effectively optimize compute density and rack space.
Ecosystem solutions and fast transaction paths
The Alveo UL3422 accelerator card is designed to enable rapid deployment by leveraging available infrastructure ecosystem solutions and reference designs , giving transactional developers the edge they need to quickly get their designs to market.
It is supported by a growing network of ecosystem partner solutions that provide IP and development frameworks to enable rapid implementation of trading solutions.
Exegy
Exegy, a provider of end-to-end front-office trading solutions, supports the AMD Alveo UL3422 card through its development framework (nxFramework), a hardware and software development environment designed to efficiently build and maintain ultra-low latency FPGA applications for the financial industry.
Hypertec
Hypertec, a provider of hardware, cloud and value-added solutions for the financial services industry, has worked closely with AMD to certify its HF X410R-G6 server to support the Alveo UL3422 accelerator card, making it the first 1U server fully optimized for the card.
Xelera Technologies
Xelera Technologies, a software provider for high-speed network technology and machine learning (ML) applications, has partnered with AMD to help address the latency deficiencies of machine learning algorithms in high-frequency trading. With Xelera Silva, users can take advantage of real-time, machine learning-based trading decisions while also using XGBoost, LightGBM, CatBoost, and other advanced models.
The Alveo UL3422 supports traditional FPGA flows with the AMD Vivado™ Design Suite and comes with a set of reference designs and performance benchmarks , enabling FPGA designers to quickly explore key metrics and develop custom trading strategies based on specifications.
AMD also provides developers with the open source and community-supported FINN development framework, which enables low-latency AI models to be deployed in high-performance trading systems. The FINN project uses PyTorch and neural network quantization technology to reduce the size of AI models while maintaining accuracy. The FINN compiler can generate quantized neural network (QNN) hardware IP blocks that can be used with AMD FPGAs.
The AMD Alveo UL3422 accelerator card is available now and is shipping in volume to financial services customers worldwide.
Want to buy this product
Please reply in the background "Alveo UL3422 purchase" to obtain the contact information.
Slide to view ↓
Recommended Reading
AMD Technology Day- Beijing Station invites registration
November 20, more information will be available soon
京公网安备 11010802033920号