FPGA Insights has engaged in an exclusive interview with Piriya Chaichanasiri, FPGA Engineer at Vanta capital
Q1) Can you provide an overview of your organization and the services/products it offers?
In the past, I have been working at CERN in Switzerland (my FPGA skill started from there), and currently, I’m working for Vanta capita in financial technology.
Q2) Can you explain the benefits of using FPGAs over other types of processors?
The FPGA has many benefits such as parallel processing, digital processing, and a low latency level we work in the nanosecond range. An FPGA never fails for each rising edge
Q3) What are the most significant trends observed in the FPGA industry over the past year? How will these trends shape the industry’s future?
From my personal experience :
In the subway system, the most safety equipment is in FPGA. I never had a problem during my previous jobs. It’s very robust.
A lot of equipment in medical devices like proton therapy, and CT scans have to use FPGA because the accelerator of the particle is controlled and must be synchronized. I almost believe that FPGA is created as an accelerator of particles.
In High-Frequency trading, the faster will earn the best part of the cake in the stock market.
In the future, we can see more applications of FPGA in telecommunication, AI, imaging processing, Deep learning, and financial trading firms.
Q4) How do you see FPGA development evolving to meet the demands of modern applications and complex workloads?
I had a chance to work with VHDL(old school with notepad ++ and modelsim) and HLS. The tool provided by Xilinx Vitis_hls could help a new developer to work with C++. The approach it’s not easy but once you understand the link. Many possibilities are open.
Q5) Key drivers behind the increasing adoption of FPGAs in various applications and industries?
I did a personal project for object detection which I used Jetson Nano to display in real time. For such a small card, I can have only 10 ~ 15fps. So if you want a 60fps you will a supercomputer with an expensive GPU and very high power consumption. This thing could be replaced by an FPGA.
So the key drivers:
Parallel input, output, multi-thread, Low consummation, and low latency.
Q6) Sectors that stand to benefit the most from FPGA integration, and why?
Financial sector, A lot of financial sectors tried to replace C++ servers with an FPGA, because a server needs a CPU, NIC, RAM, and hard disk. However, the modern FPGA card has an Ethernet Sfp++, Blocks Ram, and DSP inside of the Chip reduces the latency just for electricity travel between each block.
Telecommunications, in the past I had to create a filter with RLC, then use the chips to filter the game change with the FPGA.
Once your ADC is connected to FPGA. You can create any filter with the code. (Tip: Use Matlab to generate a filter in VHDL)
AI, the matrix calculation in a few clocks is the benefit.
Some people do a mining Bitcoin with FPGA, but since BTC is a very hard-to-find token it’s not a trend anymore.
Q7) The role of FPGAs in accelerating AI applications and advancements expected in the near future.
The accelerator processing for any large datasets.
Q8) Ensuring the security and integrity of FPGA designs, especially in sensitive applications like finance and defense.
My previous project at CERN. I have to measure the cavity of the accelerator of the proton. Around the building, there is a lot of device that deliver a high voltage of ~100kV and some emit radiation on the rack however the FPGA work very well even the extreme condition. Even the big detector CMS is fully equipped with FPGA for data acquisition.
For finance, I still working on it. But I’m very confident any data will be in security.
Q9) Advice for students and professionals interested in pursuing a career in FPGA development to stay updated with the latest trends and technologies.
My advice would be to learn VHDL or Verilog for working in the telecommunications sector.
If you already know C++ you can try the tutorial Vitis_hls.
The FPGA has evolved you also need to learn how to use BRAM, URAM, DPS, HBM, Axis stream, #pragma, and many things it takes some time to understand those things it’s not easy as Raspberry Pi where you already have thousands of library. Each driver you have to create almost by yourself. Even chat GPT can’t help you.
Good luck.
