Back to top

Heterogeneous Acceleration

Build for AMD Versal® Adaptive SoCs and FPGAs

Leverage our electronic system design service expertise and harness the power from cloud to edge.

Fidus Awarded AMD Premier Adaptive Compute Partner of the Year

Unlock performance on Versal™ with
AMD’s 2023 ACP Partner of the Year

We extend our congratulations to Fidus who has consistently provided expertise across a range of AMD products, including their expertise in designing with Versal devices. Fidus is a key contributor to the success of AMD, and we are proud to recognize them as a 2023 Adaptive Compute Partner of the Year” – Rhett Whatcott, Director, Global Training & Enablement at AMD. Read press release –>

Access NA's largest team of AMD-certified engineers

New 2nd gen Versal Portfolio

Leverage Fidus’ deep expertise with the AMD Versal family to overcome the challenges of high-tech electronic system design development:

  • Advanced Integration: Utilize AMD Versal SoCs’ superior capabilities for adaptive, domain-specific architectures that excel beyond traditional CPUs, GPUs, and FPGAs.
  • Streamlined Development: Benefit from our strategic partnership with AMD as North America’s longest Premier Partner. Our up-to-date knowledge of the latest silicon, tools, and software roadmaps drives efficiency.
  • Technical Certification: Our large team of AMD-certified professionals (largest in North America) ensure your solutions are implemented swiftly and effectively, significantly shortening time to market.
  • Risk Reduction and Acceleration: With Fidus, expect a reduction in development risks and a boost in speed for:
    • Developing high-efficiency AI kernels,
    • Managing massive data transfers using the Network on Chip (NOC),
    • Designing specialized custom hardware.

Our expertise spans the full Versal portfolio

AMD-versal-AI-EDHE-GEn-2

Versal AI Edge Series Gen 2

This series delivers an adaptive technology platform that combines high AI inference performance, low latency, and power efficiency for edge applications. Fidus optimized a Software Defined Radio Design using AI Engines leveraging the AMD VEK280:

  • RLS Decision Feedback Equalization
  • MLSE Equalization
  • Hardware-in-the-loop (MATLAB®)
  • Golden vector comparison
AMD Versal Prime Gen 2 chip

Prime Series Gen 2

The mid-range series with medium density programmable logic, signal processing, and connectivity capability.

Fidus developed CXL solutions on Versal featuring:

  • CXL engines for Type 1 and Type 3 support
  • Verified using UVM, including very Design Systems’ CXL VIP
  • Targeted at PCIe Gen5 application
Versal AI Core Series

Versal AI Core Series

The high-compute series with medium density programmable logic and connectivity capability coupled with AI and DSP acceleration engines.
Fidus conserved precious logic resources, enhanced design flexibility, and reduced power consumption using the AI Engines in a RADAR application:

  • Space Time Adaptive Processing (STAP)
  • Custom Kernel development
  • Cuboid data movement and processing
  • Hardware-in-the-loop (MATLAB®)
  • Golden vector comparison

Read more in this whitepaper.

AMD Versal Premium Series chip

Versal Premium Series

The high-end, high bandwidth series, rich in networking interfaces, security engines, and providing high compute density. Fidus developed robust custom Versal-based hardware for Mil/Aero applications featuring:

  • High density compute capabilities
  • Rugged 6U VPX Form Factor
  • SOSA™ aligned
  • Multiple high-speed memory interfaces
  • High-speed transceiver links

Versal Technical Webinar Series

Fidus have recently launched a series of webinars focused on best practices designing and working with the Versal platform.

 

application of AMD Versal Adaptive SoC to Radar Space Time Adaptive Processing (STAP)

Application of AMD Versal Adaptive SoC to Radar Space Time Adaptive Processing (STAP)

In this presentation released in conjunction with AMD, Fidus’ Versal experts discuss the application of AMD Versal Adaptive SoC to Radar Space Time Adaptive Processing (STAP) for detecting moving targets against cluttered backgrounds in space, highlighting the development approach, performance, and advantages of using AMD’s AI engines for this purpose.

Click here to read.

  • AIE (AI Engine) implementation and custom Kernel authoring
  • Network-on-Chip (NOC)
  • DCMAC (100/200/400/600GE)
  • SLR (Super Logic Region) restrictions and data transfer
  • Proficiency with high performance IO: PCIe Gen4/Gen5/CXL
  • HBM memory applications
  • Clock and timing
  • Complex signal DSP processing
  • Thermal simulations
  • XPIO interface usage and multiplexing
  • MATLAB-based development, verification, coding, targeting, and re-verification
  • PCIe Gen4 x16, 64-bit configuration, TLPs (Transaction Layer Packets)
  • Vitis™ and Vivado™ experience

Our Versal experience is second to none.

Trust us to deliver reliably and first-time-right. That’s why AMD is a customer of ours too!

Contact us