News
IndustryTSMC’s grip on advanced chipmaking will face competitionIndustryCadence joins European Chip Design Platform to support semiconductor startups and SMEsMaking chipsMarc Hijink on ASML’s Culture, Supply Chain, and GeopoliticsProcessorsIntel's post-Nova Lake CPUs take shape as Razor Lake desktop and laptop variants surfaceAI chipsHow Google’s New AI Chip Software Helps It Battle NvidiaIndiaMore IC Packaging Plants Emerge In IndiaIndustryUS Tightens China AI Chip Access as Nvidia, TSMC and Semiconductor Stocks SlideDesignAxelera AI moves into HPC with TitaniaMaking chipsChinese envoy says chip equipment export curbs threaten EU-China trade thawIndiaSEMICON India 2026 Day 2 Sees 25 MoUs, Strategic Collaborations Across Semiconductor Value ChainIndustryTSMC’s grip on advanced chipmaking will face competitionIndustryCadence joins European Chip Design Platform to support semiconductor startups and SMEsMaking chipsMarc Hijink on ASML’s Culture, Supply Chain, and GeopoliticsProcessorsIntel's post-Nova Lake CPUs take shape as Razor Lake desktop and laptop variants surfaceAI chipsHow Google’s New AI Chip Software Helps It Battle NvidiaIndiaMore IC Packaging Plants Emerge In IndiaIndustryUS Tightens China AI Chip Access as Nvidia, TSMC and Semiconductor Stocks SlideDesignAxelera AI moves into HPC with TitaniaMaking chipsChinese envoy says chip equipment export curbs threaten EU-China trade thawIndiaSEMICON India 2026 Day 2 Sees 25 MoUs, Strategic Collaborations Across Semiconductor Value Chain
EveryChip
AsAP

AsAP

University-developed many-core processor array for DSP tasks

AsAP is a chip architecture developed at the University of California, Davis. It uses an array of simple processors with asynchronous clocking to achieve high energy efficiency. The design targets digital signal processing applications through fine-grained parallelism.

Share on WhatsApp

  • AsAP 2 reaches 1.2 GHz clock speed
  • 167 processors in 65 nm CMOS design
  • Consumes 9.2 W for 1 trillion ops/sec
  • No global clock or phase signals required
  • Developed at UC Davis VLSI Lab

What it is

The Asynchronous Array of Simple Processors (AsAP) is a chip architecture created by researchers at the University of California, Davis. It consists of a two-dimensional grid of simple programmable processors. These units connect via a reconfigurable mesh network and use small local memories. The design aims for high performance and energy efficiency in digital signal processing tasks. It avoids global clock signals, using a globally asynchronous locally synchronous approach instead.

How it came about

Researchers in the VLSI Computation Laboratory developed the architecture. The first version, AsAP 1, was taped out in May 2005. It used 0.18 μm CMOS technology. The second generation, AsAP 2, followed with a 65 nm design. This later version included dedicated hardware for specific algorithms. Both chips were fully functional and demonstrated high clock rates for university-designed processors.

Inside the chip

The architecture uses task-level and fine-grained parallelism. Each processor has a simple core and small scratchpad memory. They communicate only with nearest neighbors to avoid long wires. This local connectivity improves scalability. The clocking scheme allows individual oscillators to halt when idle. This reduces power dissipation significantly. Processors can dynamically change voltage and frequency. Specialized units handle fast Fourier transforms and video motion estimation in the second generation.

How it was made

The first chip used 0.18 μm CMOS standard cell technology. It contained 36 processors. The second chip used 65 nm CMOS and held 167 processors. AsAP 2 processors reached 1.2 GHz at 1.3 V. At lower voltages, they consumed microwatts of power. The design achieved one trillion operations per second while dissipating only 9.2 watts. This efficiency remained constant across different workloads due to the MIMD architecture.

Where it went

The platform supports various digital signal processing applications. Mapped tasks include filters, sorting, and matrix multiplication. It also handles wireless LAN baseband processing for IEEE 802.11a/g standards. Video encoding blocks like H.264 CAVLC compression run on the array. A C compiler and automatic mapping tool simplify programming. The hardware manages inter-processor timing, separating it from software logic.

Versions and family

AsAP 1 featured 36 processors in a 6x6 grid. It operated at up to 540 MHz. AsAP 2 expanded to 167 processors with long-distance interconnects. It added dedicated hardware for FFT and Viterbi decoding. The second generation supported dynamic voltage and frequency scaling per processor. This allowed finer control over power and performance compared to the initial design.

Why it matters

The project demonstrated that university-designed chips could achieve high clock rates. AsAP 2 held the record for the highest clock rate in a university-fabricated processor. Its energy efficiency per operation remained stable across varying loads. This contrasted with many other architectures. The work highlighted the benefits of asynchronous design and fine-grained parallelism for signal processing.

Questions people ask about AsAP

What makes AsAP different from standard processors?

AsAP uses a globally asynchronous locally synchronous clocking scheme. It lacks global clock signals, phase-locked loops, or crystal oscillators. Individual processors halt completely when idle, saving power. This design simplifies clock distribution and improves scalability.

How many processors are in the AsAP chips?

The first generation, AsAP 1, contained 36 processors arranged in a 6x6 grid. The second generation, AsAP 2, increased this count to 167 processors. Both versions used a two-dimensional array structure with nearest neighbor communication.

What applications can run on AsAP?

The architecture supports digital signal processing tasks. Examples include fast Fourier transforms, Viterbi decoding, and JPEG encoding. It also handles IEEE 802.11 wireless LAN baseband processing and H.264 video compression blocks efficiently.

Good reading and sources

Where this page comes from

The facts on this page come from the Wikipedia article "Asynchronous array of simple processors", written by its contributors and used under the CC BY-SA 4.0 licence, and from its Wikidata record (CC0). EveryChip rewrote the text in its own words, with AI assistance, and compares every page with the article to make sure no passage is copied. It may be reused under the same CC BY-SA 4.0 licence with a link back.

It may contain mistakes, and figures change over time. Nothing here is investment or buying advice. If something is wrong, tell us. Last refreshed 2026-11-14.

Photo credits on this page

The small pictures above come from Wikimedia Commons under free licences. Some are cropped to fit.

The weekly digest

Every Monday: the chip of the day, new pages on EveryChip, and the conferences and anniversaries of the week ahead.