Altera’s Total 28-nm DSP Portfolio: Fastest Path to Highest Performance Signal Processing
Altera’s total 28-nm digital signal processing (DSP) portfolio will put you on the fastest path to the highest performance signal processing. The portfolio includes four main components.
Variable-Precision DSP Architecture
An industry first, our variable-precision DSP architecture delivers precision and performance that you configure at compile time, on a block-by-block basis, for your application. You’ll find the variable-precision DSP block, which is backward compatible, in our 28-nm Stratix® V FPGAs.
- A single 28-nm variable-precision DSP block can support precisions ranging from 12x12 to 27x27.
- The precision of each block within a device can be set independently to support bit-growth in various designs, such as finite impulse response (FIR) and fast Fourier transform (FFT) implementations.
Timing-Driven Simulink Synthesis
Our DSP Builder Advanced Blockset can create fMAX and latency-optimized netlists automatically from MATLAB/Simulink design descriptions. We call this timing-driven Simulink synthesis.
Choose the right fMAX and latency for your system, and the DSP Builder tool will do the rest. It’ll even add in registers to increase fMAX or parallelize certain critical paths to meet latency constraints. This will save you weeks of hand-tweaking the resultant HDL code.
Complete 1080p Video Design Framework
We provide a complete 1080p video design framework, which includes our Video and Image Processing Suite of intellectual property (IP) cores. A winner of the 2009 EDN Innovation Award, this framework will help speed up your design cycle.
The framework includes:
- 18 video functions
- A streaming video interface standard
- More than a half dozen hardware-verified reference designs
- A range of video development kits
Comprehensive Floating-Point IP Cores
Our comprehensive portfolio of floating-point functions is the FPGA industry’s largest. It includes functions ranging from simple operators such as ADD/SUB/INVERT to complex matrix multiplication, matrix inversion, and floating-point FFTs.