With the advent of the US Securities and Exchange Commission's (SEC's) Rule 15c3-5 on pre-trade risk due to take effect in a matter of weeks, the industry seem to have settled on the use of low-latency field programmable gate arrays (FPGAs) to make the necessary risk calculations.
However, we are only at the tip of the iceberg in adopting this architecture. It is not until any typical programmer can program these devices easily that one can say that the industry has accepted the technology, an industry contact told me recently.
In the meantime, the use of FPGAs will be a relatively expensive proposition in terms of purchasing, deploying and managing them from both a hardware and human perspective. Finding programmers well versed in the hardware description language (HDL) used by FPGAs and keeping them fat and happy in this competitive environment is going to strain the personnel budget.
There is light at the end of the tunnel. Vendors are working hard to either hide or eliminate the need to code in HDL.
Processor giant Intel is continuing to work on its Many Integrated Core (MIC) architecture that will integrate up to 32 cores on to a single die, according to vendor officials. The immediate benefit of the new architecture is that Intel multi-core and many-core platforms share the same programming model, letting C or C++ source code be compiled for either platform depending on which works best.
Microsoft researchers are also on the case with their Project Kiwi, which looks to model digital circuits with concurrent programs using a standard library in C# for multi-threaded programming among other things, according to vendor officials.
Then there are always the smaller vendors, such as British Columbia-based Secodix, that believe that they have built a better mousetrap. Secodix is betting that its eXtreme Processing Unit (XPU), which supports C++ programming on FPGA architecture, will have prospective clients knocking on its door.
Expect to see similar products to pop up in the months and year to come while the industry wrestles on a de facto standard.
Waters Wavelength Podcast Episode 75: An Update on the Julia Programming Language; AI & Alternative Data; Digital Currencies
Julia Computing's Viral Shah talks about the programming language he helped create and what's ahead for it. Then James and Anthony talk about the pairing of AI & alternative data, digital currencies, and Game of Thrones.Subscribe to Weekly Wrap emails