Specifications
Ordering Part Number
A-UL3422-P16G-PQ-G
Product
AMD Alveo UL3422 Accelerator Card
FPGA
AMD Virtex UltraScale+ VU2P
Transceiver Latency
less than 3 ns
DSP Slices
1,680
Embedded Memory
256 MB total (76 MB block RAM + 180 MB UltraRAM)
On-board Memory
16 GB DDR4-2400
Host Interface
PCIe 4.0 x8 (x16 physical connector)
Network
2× QSFP-DD, up to 16× 10/25G ports
Max Power
120 W
Cooling
Passive (server airflow)
Form Factor
Slim full-height, half-length (FHHL)
Predecessor
AMD Alveo UL3524 — same FPGA, higher port density
Target Market
Electronic trading, market data, exchange infrastructure
Overview
The AMD Alveo UL3422 is an ultra-low-latency accelerator in a slim full-height, half-length form factor, deployed in exchanges worldwide. It is powered by the AMD Virtex UltraScale+ VU2P FPGA, purpose-built for electronic trading, and uses a breakthrough transceiver architecture to reach less than 3 ns latency for trade execution.
It is the successor to the Alveo UL3524 — the same FPGA, with the port density and card design streamlined into half the size, so firms can fit more accelerator ports per server and per rack unit. 1,680 DSP slices, 256 MB of embedded memory (76 MB block RAM plus 180 MB UltraRAM), 16 GB of DDR4-2400 and dual QSFP-DD deliver the memory and I/O budget real trading strategies need.
QS Compute supplies the Alveo UL3422 (A-UL3422-P16G-PQ-G) with volume pricing and 15-day sample lead time for qualified projects.
Key Benefits
less than 3 ns transceiver latency — world-record trade execution class performance; half the card size of the UL3524 for higher rack density; 16 GB DDR4 + 256 MB embedded memory for full order-book and strategy state; dual QSFP-DD scaling to 16× 10/25G ports.
Applications
Ultra-low-latency trade execution and order-book processing; market-data feed handling at exchange colocation; latency-sensitive risk and pre-trade checks; any workload where the transceiver path latency dominates.
Request a Quote — AMD ALVEO UL3422 — ULTRA-LOW-LATENCY TRADING ACCELERATOR CARD
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →