Specifications
Family
AMD (Xilinx) Kintex UltraScale+
Device
XCKU15P
MPNs (package/speed variants)
XCKU15P-1FFVA1156E · XCKU15P-1FFVA1156I · XCKU15P-2FFVA1156I · XCKU15P-2FFVA1760E · XCKU15P-1FFVE1517I
Logic Cells
1,143,450
Logic Units
1,143,450 (CLB-equivalent)
Process
16 nm FinFET+ (TSMC)
Max Fabric Clock
Up to 775 MHz (speed grade dependent)
Core Voltage
0.85 V nominal (VCCINT)
Memory
Block RAM + UltraRAM in the UltraScale+ memory hierarchy
DSP Slices
DSP48E2 slices for fixed-point and INT8 arithmetic
Transceivers
GTY transceivers up to 32.75 Gbps
PCIe
Integrated PCIe Gen3 x16 / Gen4-capable hard block (device dependent)
Memory Interface
DDR4 up to 2,666 Mb/s hard memory controller
Packages
FCBGA-1156 (FFVA1156), FCBGA-1517 (FFVE1517), FCBGA-1760 (FFVA1760)
Temperature Grades
E (0 to +100 °C) and I (-40 to +100 °C)
Configuration
SRAM-based; requires external configuration flash
Target Use
Wired/wireless infrastructure, video, test & measurement, AI inference
Overview
XCKU15P is the largest device in AMD's Kintex UltraScale+ family, bringing 1,143,450 logic cells and 16 nm FinFET+ silicon to the mid-range price band below Virtex UltraScale+.
Kintex UltraScale+ parts are the standard choice where Virtex-class logic density is not needed but GTY transceiver bandwidth is: they carry transceivers to 32.75 Gbps and a hard DDR4 memory controller to 2,666 Mb/s, so a single device can terminate line-rate network interfaces and drive a wide memory bus.
The XCKU15P is offered in three BGA packages (1156, 1517 and 1760 balls) across speed grades -1 and -2, in both extended (E) and industrial (I) temperature grades. AMD provides the Vivado toolchain for the UltraScale+ family.
Key Benefits
1.14M logic cells at a mid-range price point, 32.75 Gbps GTY transceivers, a hard DDR4 controller to 2,666 Mb/s, and a mature Vivado toolchain with long device availability for infrastructure designs.
Applications
Wired and wireless infrastructure line cards, broadcast and professional video pipelines, test-and-measurement instruments, radar and signal-processing front-ends, and INT8/DSP48E2-based AI inference accelerators.
Request a Quote — AMD KINTEX ULTRASCALE+ XCKU15P — 1.14M LOGIC CELL FPGA
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →