Xilinx Versal ACAP | BeiLuo Distributor

About Versal ACAP

Versal ACAP is the industry's first adaptive compute acceleration platform, combining scalar engines, adaptable engines, and AI engines for next-generation computing.

As an authorized distributor of Xilinx, BeiLuo provides a comprehensive range of Versal ACAP products. These products are widely used in industrial control, automotive electronics, consumer electronics, and communication equipment. We offer a complete product line with multiple series to meet various application requirements. All products are genuine and original, backed by professional technical support and comprehensive after-sales service.

Selection Guide: How to Choose the Right Versal ACAP

Product List

2 products available

Part Number ↕	AI Engines ↕	DSP Engines ↕	Logic Cells ↕	Application Processor ↕	Real-time Processor ↕	Package ↕	Stock Status	Action
XCVM1802-2LSEVSVD1760	400 (133 INT8 TOPS)	1,968	1,852,000	Dual-core Cortex-A72	Dual-core Cortex-R5	SVD1760 (45x45mm)	In Stock	View Details
XCVE2802-2LSEVSVD1760	400 (133 INT8 TOPS)	2,952	2,780,000	-	-	SVD1760 (45x45mm)	In Stock	View Details

Need Versal ACAP Products?

Download Datasheet Request Samples Get Quote Contact FAE

Frequently Asked Questions

What is an ACAP and how is it different from FPGA?

ACAP (Adaptive Compute Acceleration Platform) differs from traditional FPGA: 1) Scalar engines - dual-core Cortex-A72 application processors, 2) Adaptable engines - next-gen FPGA fabric, 3) AI engines - vector processors optimized for ML, 4) DSP engines - high-performance signal processing, 5) NoC - network-on-chip for high-bandwidth connectivity. ACAP provides 10-100x performance improvement for AI and signal processing compared to FPGAs.

Use ACAP for AI inference, 5G, and data center acceleration. Use FPGA for traditional logic applications.

What are AI Engines and what can they do?

AI Engines are vector processors designed for machine learning: 1) Architecture - 2D array of VLIW SIMD processors, 2) Performance - up to 8 TOPS per AI Engine, 3) Data types - int8, int16, bfloat16, float32, 4) Programmability - C/C++ programming with Vitis, 5) Connectivity - direct connection to PL and memory. AI Engines excel at matrix multiply, convolution, and FFT operations common in AI and signal processing.

Use AI Engines for AI inference and vectorizable signal processing. Use PL for custom dataflow.

How do I program Versal ACAP?

Versal programming uses Vitis and Vivado: 1) Vitis - high-level application development, 2) AI Engine tools - graph compiler and kernel development, 3) Vivado - PL design and system integration, 4) PetaLinux - embedded Linux for scalar engines, 5) Libraries - Vitis AI and DSP libraries available. Programming model involves dataflow graphs for AI Engines and traditional RTL/HLS for PL.

Start with Vitis for application development. Use AI Engine tools for kernel optimization.

What is the Network on Chip (NoC)?

The NoC provides high-bandwidth connectivity: 1) Architecture - packet-switched network connecting all engines, 2) Bandwidth - up to 1 TB/s aggregate bandwidth, 3) QoS - quality of service for critical traffic, 4) Isolation - secure channels between domains, 5) Configuration - software-configurable at boot. NoC eliminates traditional routing congestion and simplifies system design.

Use NoC for high-bandwidth data movement between engines. Configure QoS for real-time traffic.

What are the power and thermal considerations?

Versal power and thermal characteristics: 1) Power - 20-60W depending on device and utilization, 2) Thermal - requires heatsink and airflow, 3) Power domains - independent control of engines, 4) Dynamic power - clock gating and power gating supported, 5) Estimation - use Xilinx Power Estimator for analysis. Premium devices have higher power density requiring careful thermal design.

Plan thermal design for 60W worst case. Use power estimation tools early in design cycle.