Interested in becoming a Labs partner?
Contact our team today to learn more about the Labs Partner Program.
-
Products
The ICON range is an overclocked, liquid-cooled server platform for Intel CPUs. Generating ultimate performance for low-latency use cases.
ACE New The ACE range is a high-performance, overclocked, liquid-cooled server platform built for AMD CPUs. Engineered to deliver ultra-low latency & maximum power.
FLEX The FLEX range focuses specifically on unique FPGA use cases such as high density FPGA servers. - Why Us
- Our Story
Orthogone Technologies
The Testing
By integrating Orthogone's ULL FPGA IP solutions with Blackcore's overclocked high-performance servers, we are enabling electronic trading firms to achieve new levels of efficiency and competitiveness. With increasing market volatility and fierce competition in electronic trading, HFT firms must continuously push the boundaries of technology to stay ahead. Blackcore integrated an AMD Xilinx X3533PV into multiple test systems in our lab. Secure remote access was then given to Orthogone to install their framework onto the FPGA card and run multiple test scenarios.
The Results
Combining Orthogone’s ULL FPGA Framework with Blackcore’s overclocked high-performance servers (ICON 3100- RL+ and G3 SPR-M) allows sub-700ns full TCP/IP loopback latency performance. It represents a latency gain of about 8% compared to a standard server (Dell/EMC PowerEdge R750). As detailed in their benchmark report, the performance of the joint solution is self-evident in terms of latency and consistency:
- 699ns Full loopback, minimum RTT latency for 64B frame (6B TCP/IP payload).
- 798ns Full loopback, 90th percentile RTT latency for 64B frame (6B TCP/IP payload).
Finally, latency improvements are still possible with readily available FPGA solutions such as the AMD-Xilinx UL3524 and UL3422, which can improve end-to-end latency by approximately 25ns for all payload sizes. Likewise, latency optimizations are still possible with the use of PCIe Gen 5 and beyond in future hardware platforms.
“FPGAs are notoriously hard to code for, so having access to a simpler pre-made framework allows firms at the start of their FPGA journey immediate access to ultra low latency performance. Blackcore again demonstrates its ability to add performance to existing products out-of-the-box. The partnership between these technologies allows firms to enter the market at a competitive level earlier and start iterating on their algorithmic logic without getting bogged down re-writing FPGA building blocks.’’
Platform
Benchmark
Version
Date
Actions
Meet our Labs Partners
This program combines the skillsets of team Blackcore and our Partners - together we will innovate to deliver the best solutions for our electronic trading clients.
Our Credentials
