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
Hexagon

Hexagon

A family of digital signal processors and neural processing units by Qualcomm.

Hexagon is a brand for digital signal processors and later neural processing units designed by Qualcomm. First introduced in 2006 as QDSP6, it powers Snapdragon chips in mobile devices and networks.

Share on WhatsApp

  • Introduced in 2006 as QDSP6 by Qualcomm.
  • Uses 4-way multithreaded VLIW architecture.
  • Clock speeds range from 400 to 2000 MHz.
  • Added floating-point support in version 5.
  • Shipped over 1.2 billion cores by 2011.

What it is

Hexagon serves as the brand name for a series of digital signal processors created by Qualcomm. Originally launched in 2006 under the code name QDSP6, meaning sixth generation digital signal processor, the architecture has evolved to include neural processing capabilities. It functions as a specialized co-processor within system-on-chip designs. The primary goal is to handle complex calculations efficiently while consuming minimal energy. This makes it suitable for running audio, video, and artificial intelligence tasks without draining battery life in portable gadgets.

How it came about

The lineage began with QDSP5 before transitioning to Hexagon in 2006. Early versions focused on integer arithmetic, adding floating-point support in version five. By 2013, Qualcomm introduced dynamic multithreading to improve performance. Subsequent iterations integrated vector extensions for advanced imaging and computer vision. Later generations added dedicated tensor accelerators to boost machine learning workloads directly on the device.

Inside the chip

The architecture uses a very long instruction word format, packing four instructions into a single packet. These execute in parallel across multiple execution units. It employs hardware-assisted multithreading to switch between tasks rapidly. Data buses are thirty-two bits wide. The design supports single instruction, multiple data operations for processing several data points simultaneously. Privilege levels manage access to system resources during execution.

How it was made

The cores are integrated into Qualcomm's system-on-chip products. They are typically found in the multimedia or modem subsystems of Snapdragon processors. User-programmable cores reside in the multimedia section, while modem cores remain controlled by the manufacturer.

Where it went

These processors appear in smartphones, cars, and wearable technology. They also operate within cellular network infrastructure components. Developers use them for audio processing, cryptography, and AI acceleration. Software libraries allow third parties to run specific algorithms on the DSP instead of the main CPU.

Versions and family

Six major versions of the QDSP6 architecture were released starting in 2006. Versions progressed through V1 to V6, appearing in Snapdragon models from 800 to 820. Earlier QDSP5 cores remained in low-end products. Later iterations included the Hexagon Tensor Accelerator and Vector Extensions to enhance specific computational tasks.

Why it matters

By 2012, Qualcomm had shipped billions of these DSP cores inside its systems. At the time, this volume made it the most widely deployed DSP architecture globally. The integration of AI engines into mainstream mobile chips relied heavily on this hardware foundation. It enabled on-device processing for photography and voice recognition features.

Problems and reception

It notes that the original hypervisor was closed-source, but an open-source alternative called MiniVM was released in 2013. Third-party software vendors eventually integrated their tools to support the platform.

Questions people ask about Hexagon

What does Hexagon do?

It is a digital signal processor that handles multimedia and AI tasks efficiently. It offloads work from the main CPU to save power and improve speed in mobile devices.

Is Hexagon open source?

The core architecture is proprietary. However, Qualcomm released an open-source hypervisor called MiniVM in 2013 for certain versions. Compiler support exists in LLVM and GCC.

How does it process instructions?

It uses Very Long Instruction Words. Four instructions are grouped into a packet and executed in parallel every clock cycle using multiple execution units.

Where is it used?

It is embedded in Qualcomm Snapdragon chips. These chips power smartphones, automotive systems, wearables, and parts of cellular network equipment.

Good reading and sources

Where this page comes from

The facts on this page come from the Wikipedia article "Qualcomm Hexagon", 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.

  • Qualcomm Snapdragon: Raimond Spekking, CC BY-SA 4.0, source
  • Tensor Processing Unit: Norman P. Jouppi, George Kurian, Sheng Li, Peter Ma, Rahul Nagarajan,, CC BY 4.0, source

The weekly digest

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