IMEC Develops Architecture for Flexible Forward Error Correction

IMEC has developed an innovative architecture for flexible forward error correction (FEC). The solution targets data transmission applications that need to combine flexibility, high throughput, and low power consumption. Examples are future wireless terminals and optical storage. IMEC’s FEC enables, on one processor, the turbo- and LDPC decoding of major communication standards. The technology is available for the industry either through a soft IP transfer, or through joint R&D projects.

IMEC flexible forward error correction (FEC)IMEC’s FEC solution supports both turbo- and LDPC coding, including multi-channel operation over different modes. It is the world’s first application-specific integrated processor (ASIP) for flexible FEC enabling both turbo- and LDPC coding for 3rd generation mobile phones (3GPP-LTE), wireless networks (IEEE802.11n, IEEE802.16(e)) and television broadcasting (DVB-S2/T2, GB20600). Support for other convolutional turbo- or LDPC codes can be enabled through assembly programming. A combined multiprocessor and ultra-wide SIMD (single instruction, multiple data) approach achieves scalability, high throughput and high energy efficiency.

The preliminary estimates for throughput and energy consumption show that IMEC’s FEC solution is competitive with solutions that separate turbo and LDPC decoding on dedicated hardware. The throughput that is achieved is between 0.07 and 1.25Mbps/MHz, with efficiencies from 0.3 to 0.5nJ/bit/iter in turbo mode and 0.08 to 0.1nJ/bit/iter in LDPC mode. The silicon area used by the flexible solution is comparable to the sum of multiple dedicated solutions.

The new flexible FEC fits in IMEC’s research strategy to design flexible components for data transmission. These are targeted at, amongst others, future mobile terminals; they combine high throughput, low power consumption, and a small footprint. Other IMEC components that follow this vision are a flexible RF transceiver and a flexible baseband chip. IMEC invites partners to collaborate in this research through its joint research programs. Industrial players can also profit from IMEC’s research by licensing the components.

FEC is used in all digital transmitters and receivers to ensure that the digital message is sent free of errors. When the transmitter sends a message, it encodes the bit stream, adding redundant data. These allow the receiver to detect and correct errors – within some bounds – without asking the transmitter for additional data. State-of-the-art FEC mainly uses 2 methods of FEC coding: turbo codes and LDPC codes. These are popular because they allow high-speed FEC encoding and decoding. But turbo codes and LDPC codes are complex, and decoding them puts a heavy computational load on the receiver. Therefore, until recently, FEC decoders for the different FEC methods were implemented as dedicated hardware blocks, focusing on minimum power consumption and area, and thereby sacrificing flexibility.

About IMEC
IMEC is a world-leading independent research center in nanoelectronics and nanotechnology. IMEC vzw is headquartered in Leuven, Belgium, has a sister company in the Netherlands, IMEC-NL, offices in the US, China and Taiwan, and representatives in Japan. Its staff of more than 1600 people includes more than 500 industrial residents and guest researchers. In 2007, its revenue (P&L) was EUR 244.5 million. IMEC’s More Moore research aims at semiconductor scaling towards sub-32nm nodes. With its More than Moore research, IMEC looks into technologies for nomadic embedded systems, wireless autonomous transducer solutions, biomedical electronics, photovoltaics, organic electronics and GaN power electronics. IMEC’s research bridges the gap between fundamental research at universities and technology development in industry. Its unique balance of processing and system know-how, intellectual property portfolio, state-of-the-art infrastructure and its strong network worldwide position IMEC as a key partner for shaping technologies for future systems.