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
Intel RapidCAD-1

RapidCAD

A 486DX in a 386 socket, with a dummy FPU chip to match.

Photo: Konstantin Lanzet (with permission), CC BY-SA 3.0, via Wikimedia Commons

RapidCAD is a two-chip upgrade set from Intel that puts an i486DX processor into a board built for the 80386. The second, dummy chip fills the old 387 socket and supplies the error signal the system expects. It was sold mainly to speed up floating-point work such as computer-aided design.

Share on WhatsApp

  • An i486DX core packaged to fit an Intel 80386 socket
  • RapidCAD-2 is a dummy FPU that only supplies the FERR signal
  • Floating-point work runs up to 80% faster than a 386DX
  • No level 1 cache, so integer gains stay around 10% typical
  • CHMOS 0.8 µm, 800,000 transistors, 132-pin CPGA

What it is

RapidCAD is a matched pair of Intel chips sold as a drop-in upgrade for machines built around the 80386 processor and its 80387 floating-point unit. The main chip, RapidCAD-1, is an i486DX in a special package that fits the 386 socket. Because that part already carries its own floating-point hardware on the die, the second chip, RapidCAD-2, is a dummy that sits in the 387 socket and only provides the FERR (floating-point error) signal older boards expect. Intel pitched it at computer-aided design, where floating-point maths dominates, and set a suggested list price of US$499.

How it came about

The set was aimed at owners of 386DX systems who wanted 486-class maths without replacing the motherboard. Intel packaged its existing i486DX core so it could drop into the older socket, and supplied a companion part to occupy the vacant 387 slot and keep the board's error signalling intact. Marketing centred on CAD workloads, which is where the RapidCAD name comes from.

Inside the chip

RapidCAD-1 carries an i486DX core, so the integer and floating-point units sit on one die and run instructions in the same number of cycles as a standard i486DX. RapidCAD-2 has 275,000 transistors but does no arithmetic; it exists to drive the FERR line. Two things hold the package back. There is no level 1 cache, the small fast memory normally built into a 486, and all traffic still crosses the narrow 386 bus. Both parts run at 5 V and 16–33 MHz.

How it was made

RapidCAD-1 is built on a CHMOS 0.8 µm process and holds 800,000 transistors in a 132-pin ceramic PGA package. RapidCAD-2 uses CHMOS as well, with 275,000 transistors in a 68-pin ceramic PGA. Both are rated for 5 V operation across 16–33 MHz. The sSpec codes are SZ624 for RapidCAD-1 and SZ625 for RapidCAD-2.

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

Where it went

RapidCAD went into existing 386DX desktop motherboards as a processor upgrade, with the main chip in the CPU socket and the dummy part in the 387 maths socket. Its intended workload was computer-aided design and other floating-point heavy tasks. Integer code gains little, typically about 10% over a 386DX and at best 35%, while floating-point work runs up to 80% faster. One report puts the overall gain at 20 to 170% against ordinary 386DX boards.

Versions and family

The set has two parts. RapidCAD-1 is the processor, a 132-pin CPGA carrying the i486DX core with 800,000 transistors. RapidCAD-2 is the 68-pin dummy FPU with 275,000 transistors. Both share the 16–33 MHz clock range and 5 V supply. Related Intel upgrade products include the 80486 OverDrive and the Pentium OverDrive.

Why it matters

RapidCAD showed that a 486-class core could be fitted to a 386 board by repackaging the silicon and faking the companion maths chip. Its value was narrow but real: CAD users got much faster floating-point maths from hardware they already owned. Integer performance barely moved, so the product is remembered as a floating-point upgrade rather than a general speed-up.

Problems and reception

The main criticism is that integer performance improves only slightly, because the design lacks level 1 cache and remains tied to the 386 bus. Reported speed-ups vary widely, from a modest 10% to as much as 170% depending on the workload and the report. The set was offered at a manufacturer's suggested list price of US$499.

Questions people ask about RapidCAD

Why does RapidCAD need two chips?

The i486DX already has its floating-point unit on the die, so no real maths coprocessor is needed. The second package is a dummy that fills the 387 socket and drives the FERR signal, which older motherboards require for correct floating-point error handling.

How much faster is it than a 386DX?

Integer code improves little, typically about 10% and at most 35%. Floating-point work gains much more, up to 80% faster in Intel's marketing, while one report puts the overall boost at 20 to 170% depending on the board and workload.

Why is integer performance not much better?

RapidCAD has no level 1 cache, the small on-chip memory a 486 normally uses to avoid slow memory accesses. It also still talks to the rest of the machine over the 386 bus, which limits how fast data can move.

What is inside the two packages?

RapidCAD-1 holds the i486DX core, 800,000 transistors in a 132-pin ceramic PGA. RapidCAD-2 holds 275,000 transistors in a 68-pin ceramic PGA but performs no arithmetic. Both run at 5 V and 16–33 MHz on a CHMOS process.

Good reading and sources

Where this page comes from

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