Our Vision
At MLE we share the vision of technology leaders such as IBM and Xilinx: General purpose compute architectures are running out of steam. Domain-specific compute architectures have become a necessity to deliver higher compute performance.
This drives compute-hungry applications such as Autonomous Vehicles or Datacenter Analytics or Deep Neural Network algorithms for Edge Inference to adopt Field-Programmable Gate-Arrays (FPGA), either as highly integrated, single-chip Adaptable Compute Acceleration Platforms (ACAP) or as reconfigurable compute companions tightly coupled to powerful CPUs via links that deliver massive bandwidth at very low latency.
Your Challenge
lies in a gap between what FPGA technology makes possible to do and what a typical engineering team can do within today's tight deadlines. This Design Productivity Gap makes FPGA technology difficult and risky for new adopters. World-wide, only a small portion of engineers has the unique skill set to master the complexity of FPGA programming:
- Filling over 100 million gate-equivalents with correct functionality
- Configuring hundreds of pins, each one to meet one of the many different I/O standards
- Parameterizing configurable I/O transceivers for at ten’s of gigabits to guarantee signal integrity
- Fully use of heterogeneous compute elements such as Lookup Tables (LUT), Digital-Signal Processing (DPS) blocks or modern Math Engines (ME)
- Implementing rich software for powerful embedded multi-core CPUs in AMP or SMP
- Inter-Process Communication (IPC) between multiple CPUs with different Operating Systems,
- Speaking various languages including classical Hardware Description Languages, VHDL and Verilog, and High-Level Languages C, C++, or SystemC
- Thinking in streaming algorithms to deliver energy and cost-efficient custom high-speed processing blocks for Data-in-Motion
Our Mission
is supporting customer projects with deep expertise and hands-on design services, offering pre-validated FPGA subsystems of FPGA blocks integrated with (open source) software, applying and promoting novel FPGA design methodologies for increased design productivity, including High-Level Synthesis, and fostering FPGA education via strong relationships with teaching and research engagements at Universities.
Our Offerings
- Functional Safety and Security (SIL/ASIL design flows, Trusted Execution Environments)
- High speed connectivity for storage and networking (10/25/100 GigE, TCP/UDP/IP, SATA, SAS, NVMe)
- Tools and methodologies for algorithm acceleration (Xilinx SDx), accelerators for ARM-, POWER-, and x86-based processing (PCIe NTB, OpenCAPI SNAP)
- Multi disciplinary design teams with expertise starting from domain-specific compute architectures over RTL design to electronic systems implementation
Our Applications and Markets
Over the last 30 years FPGAs have demonstrated significant values in applications which need highly-flexible programmable silicon across a wide span of markets and technologies - from consumer to cars to the cloud. MLE has successfully delivered solutions for classical or high-speed connectivity, software and system acceleration, Digital Signal Processing or Deep-Learning.
- Datacenter storage, network protocol and algorithm acceleration in the cloud
- Autonomous Driving and ADAS in automotive markets
- Test & Measurement with high-speed data acquisition or recording
- Image processing and sensor fusion with camera, Radar, Lidar sensors
