MIPS, ARM pushing RISC/DSP cores in Taiwan
EE Times: MIPS, ARM pushing RISC/DSP cores in Taiwan | |
Mike Clendenin (07/28/2005 10:12 AM EDT) URL: http://www.eetimes.com/showArticle.jhtml?articleID=166403298 | |
TAIPEI, Taiwan Highlighting the increasing demands on consumer devices, especially portable systems, MIPS Technologies Inc. and ARM Ltd. are ratcheting up their efforts to persuade designers here to use combined RISC/DSP cores in the next wave of designs. Reckoning that more than 50 percent of commonly executed DSP tasks are modestly complicated, MIPS believes that much of this work can be handled in a high performance RISC-based host processor with DSP instruction sets, such as multiplier-accumulator (MAC). “Over time you see more and more applications moving over into this space,” said Chris Cavigioli, MIPS’ market development manager for multimedia, during the Embedded Systems Conference-Taiwan. “We see applications in audio, such as music decompression in MP3 players, voice-over-IP, which is typically the speech coding algorithms, and fax and data modem algorithms and JPEG image decompression.” Bringing together the worlds of RISC and DSP is nothing new. But as notions like the “digital home” and “connected computing” begin to gain traction, demand for devices to carry out complex control and signal processing intensive tasks is growing. Taiwan hopes to be at the forefront of the convergence between the PC realm and the consumer electronics world where these devices will operate. But working with both RISC and DSP architectures poses some programming difficulties for developers. First, they often must juggle distinct development environments: While RISC devices are control-oriented and C-friendly, DSP engines require difficult, data- and math-intensive machine code "inner loop" programming. In addition, code development must be done in a heterogeneous multiprocessor environment, with at least two or more compute elements. In its approach to combining the functions, Cavigioli said MIPS is not out to create the world’s best DSP, but simply to offer what’s good enough. With a die penalty of about 3 percent, he said MIPS can double its performance of real-time signal processing tasks and capture about 80 percent of the more commonly executed, less intensive DSP tasks. Because of increasing system complexity, stemming from a need to handle more intensive algorithms, for instance, and also due to the need to meet the demand for lower development costs, designers are using RISC/DSP engines. “Now if you look at RISC with DSP versus a single processor, a dedicated signal processor can offer better performance. But a RISC with a DSP capability can give you much greater flexibility and greater ease of design reuse and also faster time to market,” said John Cornish, marketing vice president of ARM's Processor Division. ARM’s approach also includes adding DSP instructions, including MAC, and single-instruction, multiple-data (SIMD). Cornish said ARM’s next-generation cores will include a hybrid 64- or 128-bit SIMD architectures and include integer and single precision floating point capability, which will be important for high performance multimedia applications. Bernard Cole, editor of Embedded.com, contributed to this report.
| |
- - | |
Related News
- ParthusCeva Announces Architecture Standard for Hybrid DSP/RISC-Based System-on-Chip for ARM Environment
- ARM And Mediatek Partner To Bring RISC Technology To Taiwan
- Taiwan's Macronix licenses ARM's 16/32-bit RISC processor cores
- Taiwan's Macronix licenses ARM's 16/32-bit RISC processor cores
- Taiwan's ITRI picks MIPS' RISC processor core for cable-modem designs
Breaking News
- Ubitium Debuts First Universal RISC-V Processor to Enable AI at No Additional Cost, as It Raises $3.7M
- TSMC drives A16, 3D process technology
- Frontgrade Gaisler Unveils GR716B, a New Standard in Space-Grade Microcontrollers
- Blueshift Memory launches BlueFive processor, accelerating computation by up to 50 times and saving up to 65% energy
- Eliyan Ports Industry's Highest Performing PHY to Samsung Foundry SF4X Process Node, Achieving up to 40 Gbps Bandwidth at Unprecedented Power Levels with UCIe-Compliant Chiplet Interconnect Technology
Most Popular
- Cadence Unveils Arm-Based System Chiplet
- CXL Fabless Startup Panmnesia Secures Over $60M in Series A Funding, Aiming to Lead the CXL Switch Silicon Chip and CXL IP
- Esperanto Technologies and NEC Cooperate on Initiative to Advance Next Generation RISC-V Chips and Software Solutions for HPC
- Eliyan Ports Industry's Highest Performing PHY to Samsung Foundry SF4X Process Node, Achieving up to 40 Gbps Bandwidth at Unprecedented Power Levels with UCIe-Compliant Chiplet Interconnect Technology
- Arteris Selected by GigaDevice for Development in Next-Generation Automotive SoC With Enhanced FuSa Standards
E-mail This Article | Printer-Friendly Page |