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
Die shot of Silicon Graphics GE25A Geometry Engine graphics accelerator chip (VF4339-0001 GEOMETRY ENG GE25A).

Geometry Engine

Early VLSI vector processor for 3D graphics

Photo: Pauli Rautakorpi, CC BY 3.0, via Wikimedia Commons

The Geometry Engine is a custom vector processor chip designed at Stanford University in the early 1980s. It accelerated three-dimensional graphics calculations and formed the hardware basis for the first Silicon Graphics workstations.

Share on WhatsApp

  • Designed by Jim Clark and Marc Hannah at Stanford
  • First demonstrated in 1981 as a 40-pin VLSI chip
  • Multiplies two four-element vectors every 15 microseconds
  • Powered the Silicon Graphics IRIS 1000 terminal in 1983
  • Performs approximately 6 million operations per second

What it is

The Geometry Engine is an application-specific integrated circuit built as a vector processor. It handles mathematical operations required for three-dimensional computer graphics. Jim Clark and Marc Hannah created the design at Stanford University under a contract from ARPA. The chip first appeared in 1981. It processes vertices and transformation matrices to render images. This device laid the groundwork for modern graphics hardware used in artificial intelligence and visual computing today.

How it came about

Researchers at Stanford University developed the chip to solve complex graphics problems. The project operated under funding from the Advanced Research Projects Agency. Clark and Hannah led the design effort. They demonstrated the initial version in 1981. The technology later moved into commercial products. Silicon Graphics used the engine in their earliest workstations. The IRIS 1000 terminal launched in November 1983. This product brought hardware-accelerated 3D graphics to the market for the first time using this specific chip architecture.

Inside the chip

The chip uses a very large scale integration approach. It fits into a 40-pin package. The processor multiplies two four-element vectors every 15 microseconds. One vector holds vertex coordinates in three-dimensional space. The other holds part of a transformation matrix. Clark proposed linking twelve of these chips in a pipeline. This system would handle matrix multiplications and clipping tasks. It also scales results to match display coordinates. This parallel structure allows high-speed processing of geometric data.

How it was made

The device is an early example of VLSI technology. It was implemented as a single 40-pin integrated circuit. The design focuses on vector multiplication capabilities. It serves as an ASIC tailored for graphics math. The fabrication details beyond the pin count and integration style are not provided in the available records.

Geometry Engine
A drawing made by EveryChip from what this page says. It is a diagram, not a picture of the real chip.

Where it went

Silicon Graphics adopted the chip for its computer workstations. The IRIS 1000 terminal relied on this hardware for graphics acceleration. It shipped in late 1983 as the company's first product. The system performs about 6 million operations per second. It handles line, character, and polygon clipping. Users employed these machines for computer-aided design and visualisation tasks. The engine enabled real-time rendering of three-dimensional scenes on dedicated terminals.

Versions and family

The initial implementation served as the foundation for a broader geometry system. A pipeline of twelve engines was the intended configuration for full matrix operations. The IRIS 1000 represents the primary commercial variant mentioned. The design evolved into the core technology for subsequent Silicon Graphics hardware generations.

Why it matters

This processor is considered a forerunner to modern tensor processors. It influenced hardware marketed for both graphics and artificial intelligence. By enabling hardware-accelerated 3D graphics, it changed workstation capabilities. The technology supported the growth of computer-aided design industries. Its architectural concepts persist in today’s graphics processing units and AI accelerators.

Questions people ask about Geometry Engine

Who designed the Geometry Engine?

Jim Clark and Marc Hannah designed the chip at Stanford University. They worked under a contract from ARPA to create a processor specifically for three-dimensional computer graphics tasks.

What was the first product to use this chip?

The Silicon Graphics IRIS 1000 terminal was the first commercial product. It shipped in November 1983 and used the Geometry Engine to provide hardware-accelerated 3D graphics capabilities.

How fast is the Geometry Engine?

The chip can multiply two four-element vectors every 15 microseconds. In system operation, it achieves a performance rate of approximately 6 million operations per second for graphics processing tasks.

Good reading and sources

Where this page comes from

The facts on this page come from the Wikipedia article "Geometry Engine", 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.