The Need for Variable Precision DSP Architecture
By Suhel Dhanani, Altera Corp.
Programmable Logic DesignLine (05/05/10, 02:57:00 AM EDT)
Two fundamental trends driving the electronic infrastructure business are the need for higher data (and increasingly mobile) bandwidth and the need for higher resolution video. While somewhat distinct, these trends are overlapping in that video is the primary driver for the ever-increasing data bandwidth requirements. Cisco estimates that �all forms of video will account for close to 90% of consumer (Internet) traffic by 2012�.�
The need for increased data bandwidth inherently implies higher performance and higher precision data processing. More and higher resolution data needs to be processed, while maintaining stringent power and cost specifications. Conceptually Figure 1 shows that while processing performance/precision needs are going up�total system cost and power targets remain the same.
Figure 1. Increasing Processing with Strict Power and Cost Budgets
Video processing provides a very compelling case study of this trend. We have seen the video infrastructure gearing up to handle the processing of high definition (HD) video (up from standard definition or SD) and now even 4K (3D) video resolutions. As this transition takes place�the number of pixels that need to be processed per frame as well as the color depth (the pixel resolution) increases, thus enabling higher image quality.
To read the full article, click here
Related Semiconductor IP
- HBM4 PHY IP
- eFuse Controller IP
- Secure Storage Solution for OTP IP
- Ultra-Low-Power LPDDR3/LPDDR2/DDR3L Combo Subsystem
- MIPI D-PHY and FPD-Link (LVDS) Combinational Transmitter for TSMC 22nm ULP
Related Articles
- Reusable architecture is DSP framework
- Understanding the reuse of a DSP architecture for different designs (Ceva)
- Embedded DSP Software Design Using Multicore a System-on-a-Chip (SoC) Architecture: Part 2
- A MAC-less Neural Inference Processor Supporting Compressed, Variable Precision Weights
Latest Articles
- Making Strong Error-Correcting Codes Work Effectively for HBM in AI Inference
- Sensitivity-Aware Mixed-Precision Quantization for ReRAM-based Computing-in-Memory
- ElfCore: A 28nm Neural Processor Enabling Dynamic Structured Sparse Training and Online Self-Supervised Learning with Activity-Dependent Weight Update
- A 14ns-Latency 9Gb/s 0.44mm² 62pJ/b Short-Blocklength LDPC Decoder ASIC in 22FDX
- Pipeline Stage Resolved Timing Characterization of FPGA and ASIC Implementations of a RISC V Processor