R
Virtex-5 Family Overview
Virtex-5 FPGA Logic
550 MHz Integrated Block Memory
•
•
•
On average, one to two speed grade improvement over
Virtex-4 devices
•
•
•
•
Up to 16.4 Mbits of integrated block memory
36-Kbit blocks with optional dual 18-Kbit mode
True dual-port RAM cells
Independent port width selection (x1 to x72)
Cascadable 32-bit variable shift registers or 64-bit
distributed memory capability
Superior routing architecture with enhanced diagonal
routing supports block-to-block connectivity with
minimal hops
−
−
Up to x36 total per port for true dual port operation
Up to x72 total per port for simple dual port operation
(one Read port and one Write port)
−
−
Memory bits plus parity/sideband memory support for
x9, x18, x36, and x72 widths
Configurations from 32K x 1 to 512 x 72
(8K x 4 to 512 x 72 for FIFO operation)
•
Up to 330,000 logic cells including:
−
−
−
−
Up to 207,360 internal fabric flip-flops with clock enable
(XC5VLX330)
Up to 207,360 real 6-input look-up tables (LUTs) with
greater than 13 million total LUT bits
Two outputs for dual 5-LUT mode gives enhanced
utilization
•
Multirate FIFO support logic
−
Full and Empty flag with fully programmable Almost Full
and Almost Empty flags
Logic expanding multiplexers and I/O registers
•
•
•
•
Synchronous FIFO support without Flag uncertainty
Optional pipeline stages for higher performance
Byte-write capability
550 MHz Clock Technology
•
Up to six Clock Management Tiles (CMTs)
Dedicated cascade routing to form 64K x 1 memory
without using FPGA routing
Integrated optional ECC for high-reliability memory
requirements
Special reduced-power design for 18 Kbit (and below)
operation
−
Each CMT contains two DCMs and one PLL—up to
eighteen total clock generators
Flexible DCM-to-PLL or PLL-to-DCM cascade
Precision clock deskew and phase shift
Flexible frequency synthesis
•
•
−
−
−
−
Multiple operating modes to ease performance trade-off
decisions
−
−
−
−
−
Improved maximum input/output frequency
Fine-grained phase shifting resolution
Input jitter filtering
Low-power operation
Wide phase shift range
550 MHz DSP48E Slices
•
•
•
25 x 18 two’s complement multiplication
Optional pipeline stages for enhanced performance
Optional 48-bit accumulator for multiply accumulate
(MACC) operation with optional accumulator cascade
to 96-bits
•
Differential clock tree structure for optimized low-jitter
clocking and precise duty cycle
•
•
32 global clock networks
Regional, I/O, and local clocks in addition to global
clocks
•
Integrated adder for complex-multiply or multiply-add
operation
•
•
•
Optional bitwise logical operation modes
Independent C registers per slice
SelectIO Technology
•
•
•
Up to 1,200 user I/Os
Fully cascadable in a DSP column without external
routing resources
Wide selection of I/O standards from 1.2V to 3.3V
Extremely high-performance
−
Up to 800 Mb/s HSTL and SSTL
(on all single-ended I/Os)
Up to 1.25 Gb/s LVDS (on all differential I/O pairs)
−
Works in conjunction with SelectIO technology to
simplify source-synchronous interfaces
•
•
•
True differential termination on-chip
Same edge capture at input and output I/Os
Extensive memory interface support
DS100 (v5.0) February 6, 2009
www.xilinx.com
Product Specification
3