Titanium FPGAs

Designed for compute acceleration

Titanium FPGA Overview

You only have a few square millimeters to spare, and you need to pack in as much computing power as you can. Efinix’s next-generation Titanium FPGAs can help. Titanium FPGAs are fabricated on a 16 nm process, delivering high performance with the lowest possible power and a small physical size. They feature the innovative Quantum™ compute fabric that, with its enhanced compute capability, makes Titanium FPGAs ideal for embedded hardware acceleration applications. With a wide range of logic element (LE) densities from 35K to 1M, and compatibility with the Efinix RISC-V SoC cores, they can help you turn a tiny chip into an accelerated embedded compute system.

The Quantum compute fabric is made up of configurable tiles, the eXchangeable logic and routing (XLR) cell, that optimizes routing efficiency and speed while achieving high utilization ratios. The fabric also has highly configurable, embedded memory blocks along with dedicated, high-speed, DSP blocks. Together, these features deliver optimum performance for a wide array of applications from edge compute to industrial automation and video processing.

The 16 nm process node gives Titanium FPGAs a small footprint with low power consumption, making them ideal for highly integrated applications.

Titanium FPGA Block Diagram

Titanium FPGA Block Diagram

Ready for Computing

Titanium FPGAs are available in densities from 35K to 1 million logic elements and cover a broad range of applications:

Ti35 and Ti60

Designed for highly integrated mobile and edge devices that need low power, a small footprint, and a multitude of I/Os.

  • Mobile
  • Edge
  • AI IoT
  • Sensor fusion

Ti90, Ti120, and Ti180

Include 2.5 Gb embedded MIPI interfaces for multi-camera, high definition vision systems, edge computing and hardware acceleration.

  • Vision systems
  • Edge computing
  • Hardware acceleration
  • Machine learning

Ti240, Ti375, Ti550

Combines the compute density and transceiver interfaces for compute and industrial automation.

  • Industrial automation
  • Automotive
  • Adaptive acceleration
  • Fog computing

Ti750 and Ti1000

High-performance platform with the density and interfaces needed in the most demanding applications.

  • Communications
  • PCI Express accelerator card
  • FPGA-based server
  • Smart storage

Resource and Packaging Overview

Resources and Interfaces

Feature Ti35 Ti60 Ti90 Ti120 Ti180 Ti240 Ti375 Ti550 Ti750 Ti1000
Logic Elements (LEs) 36,176 62,016 89,812 119,750 176,256 236,888 370,137 533,174 727,056 969,408
10K RAM blocks (Mb) 1.53 2.62 7.34 9.8 13.11 19.37 27.53 39.65 54.07 72.09
DSP blocks 93 160 359 478 640 946 1,344 1,936 2,640 3,520
PLLs 4 4 10 10 10 10 10 10 10 10
GPIO 34 34 80 80 80 80 80 80 80 80
High-speed I/O 146 146 192 192 192 172 172 268 268 268
DDR4, LPDDR4 x32 x32 x32 x72 x72 2 x72 2 x72 2 x72
MIPI D-PHY 2.5 Gbps 4 RX
4 TX
4 RX
4 TX
4 RX
4 TX
3 RX
3 TX
3 RX
3 TX
3 RX
3 TX
3 RX
3 TX
3 RX
3 TX
16 Gbps Serdes x8 x8 x8 x12 x12 x16 x16 x16
25.8 Gbps Serdes x8 x8 x8
PCI Express Gen4 (16G) 1 x4 1 x4 1 x4 2 x4 2 x4 2 x8 2 x8 2 x8

Package Options

Package Pitch (mm) Size (mm) Ti35 Ti60 Ti90 Ti120 Ti180 Ti240 Ti375 Ti550 Ti750 Ti1000
64-ball WLCSP
0.4 3.5x3.4
100-ball FBGA
0.5 5.5x5.5
225-ball FBGA 0.5 8x8
225-ball FBGA 0.65 10x10
324-ball FBGA 0.65 12x12
484-ball FBGA 0.65 15x15
484-ball FBGA 0.8 18x18
625-ball FBGA 0.65 17x17
784-ball FBGA 0.8 23x23
1,156-ball FBGA 1.0 35x35

Efinity Software Support

The Efinity® software provides a complete tool flow from RTL design to bitstream generation, including synthesis, place-and-route, and timing analysis. The software has a graphical user interface (GUI) that provides a visual way to set up projects, run the tool flow, and view results. The software also has a command-line flow and Tcl command console. The software-generated bitstream file configures the Titanium devices. The software supports the Verilog HDL and VHDL languages. Read more

Titanium FPGAs are fully supported by the Efinity software, which has been optimized to take full advantage of the new features of the Quantum compute fabric. The result is highly efficient synthesis and placement that delivers compact and energy efficient designs.