Andes Technology Corporation Launches New DSP+CPU Core Offering 134 Percent Higher Performance than Competing Alternatives for Low-Power SoC Designs

Facebook
Twitter
LinkedIn

AndesCore™ D1088 CPU+DSP Combination Delivers Better Performance, Using Less Power with Half the Code Size of Competitive Offerings

【HSINCHU, Taiwan】Andes Technology Corporation, the leading Asia-based supplier of high-volume, low-power 32-bit embedded CPU cores, today announced the D1088, a 5-stage pipeline processor with integrated DSP offering 130 SIMD (single instruction, multiple data) instructions. When implemented in a 90nm low power process, the D1088 delivers 588 DMIPS, 134% higher than competing offerings. Measured using the popular Whetstone floating-point benchmark, the D1088 achieves 92% better performance. When running popular and comprehensive (over 200) DSP libraries, the D1088 is 116% faster with half the code size. Even with the above advantages, the D1088 still achieves smaller die area and less power per MHz than competing offerings.

“We’re proud of the results achieved with the D1088 DSP+CPU core,” said Dr. Charlie Su, Chief Technology Officer and Senior VP of R&D at Andes Technology Corporation. “With the increasing numbers of designs performing signal processing for smart sensors, image processing, motion detection, audio and video, the D1088 will provide a wide range of benefits, including better compute performance and power savings, as well as additional functionality. The D1088 easily interfaces with both AHB and AXI bus for simplified use in SOC devices. It comes with an optional memory management unit to support the Linux OS and an optional memory protection unit to support real time operating systems (RTOS).”

About the D1088
Tightly integrated Integer and DSP processor architectures are not new, but most were designed for applications where power was not as much a constraint as it is today. The new D1088 was designed with low power in mind. It contains functionality to enhance efficiency and reduce application code size to lower both power and cost. For example, to significantly boost the computational efficiency in matrix, filtering, Fourier Transform, and statistics functions, it can execute 4-way 8-bit, or 2-way 16-bit SIMD instructions in a single-cycle latency. In addition, for multimedia applications, the D1088 also supports 64-bit add, subtract, and multiply mixed computation.
For voice application, the D1088 offers left shift, right rounding and shift, most significant word, 32×32 multiply and specially designed 32-bit instructions to replace the lengthy 64-bit computation. To reduce code size and increase efficiency, the D1088 provides a Zero Overhead Loop instruction to offload loop branching.  To enhance parallel computational capacity, the D1088 provides left and right shift, minimum, maximum, and absolute value, besides traditional SIMD instructions such as add, subtract, and multiply.

Availability
D1088 is available immediately for customer uses and is fully supported by AndesSight™ Integrated Development Environment (IDE) tool chain.