R8000
MIPS IV microprocessor chipset for scientific computing workstations and servers.
The R8000 is a superscalar microprocessor chipset developed by MIPS Technologies, Toshiba, and Weitek. Released in June 1994, it was the first implementation of the MIPS IV instruction set. Designed for high-performance floating-point calculations, it powered Silicon Graphics workstations and supercomputers before being replaced by the R10000.
- First implementation of the MIPS IV instruction set architecture.
- Superscalar design issues up to four instructions per cycle.
- Powered 50 of the top 500 supercomputers in late 1994.
- Fabricated using a 0.7 μm CMOS process by Toshiba.
What it is
The R8000 is a microprocessor chipset that implements the MIPS IV instruction set architecture. It was developed jointly by MIPS Technologies, Toshiba, and Weitek. The chipset includes the R8000 integer processor, the R8010 floating-point unit, Tag RAMs, and a streaming cache. Introduced in June 1994, it aimed to deliver supercomputer-level performance in a smaller microprocessor form factor. It was designed specifically for technical and scientific computing tasks requiring heavy floating-point throughput.
How it came about
Development began in the early 1990s at Silicon Graphics. The goal was to replace large, discrete-component supercomputers with cheaper RISC microprocessors. MIPS Computer Systems announced initial details in April 1992. After Silicon Graphics acquired the company, development moved to its subsidiary, MIPS Technologies. The launch was delayed from 1993 to mid-1994. The first 75 MHz version appeared on 7 June 1994. A faster 90 MHz model followed in mid-1995. High costs and a narrow market limited its adoption. The cheaper and faster R10000 largely replaced it in January 1996.
Inside the chip
The R8000 is a superscalar processor capable of issuing four instructions per cycle. It executes instructions in program order using a five-stage integer pipeline. The design separates integer and floating-point operations to improve efficiency. The R8000 handles integer tasks, while the R8010 unit processes floating-point instructions. A decoupled instruction queue allows limited out-of-order execution for floating-point operations. This structure helps mitigate latency from the external streaming cache. The integer unit features two execution units, a shift unit, and address generators. The floating-point unit uses an iterative algorithm for division and square roots.
How it was made
Toshiba manufactured both chips using its 0.7 μm VHMOSIII process. This triple-layer metal CMOS technology supported the high transistor counts. The R8000 die contains 2.6 million transistors and measures roughly 300 mm². The R8010 contains 830,000 transistors. Both chips sit in 591-pin ceramic pin grid array packages. They operate on a 3.3 V power supply. The R8000 dissipates 13 W at 75 MHz. The external Tag RAMs use a 0.7 μm BiCMOS process to reduce cycle times for cache lookups.
Where it went
Silicon Graphics integrated the R8000 into several high-end systems. These included the Power Indigo2 workstation, Power Challenge server, and Power Onyx visualization system. The chipset also powered the Power ChallengeArray cluster. In November 1994, fifty of the top 500 supercomputers used R8000 processors. The highest-ranked systems were four Power Challenges, each containing eighteen R8000 chips. The processor delivered strong floating-point performance, achieving a SPECfp92 rating of 310. Its integer performance was more modest, with a SPECint92 rating of 108.
Versions and family
The initial release operated at 75 MHz. A faster 90 MHz variant became available in mid-1995. The chipset relies on external components, including two Tag RAMs and a streaming cache ranging from 1 to 16 MB. The streaming cache uses commodity synchronous static RAM. This modular approach allowed system builders to adjust memory capacity. The R8000 was succeeded by the R10000, which offered better general performance at a lower cost. No other major variants of the R8000 itself were produced.
Why it matters
The R8000 demonstrated that microprocessors could match the performance of traditional supercomputers. It helped make high-performance computing more accessible via deskside workstations. Its design influenced later superscalar processors by decoupling integer and floating-point pipelines. Although its market share was limited by cost, it held a significant presence in the TOP500 supercomputer list upon release. It paved the way for subsequent MIPS designs that further integrated these functions.
Problems and reception
The chipset was popular within its intended niche of scientific computing. Users valued its tremendous floating-point throughput. However, the high price of US$2,500 per chip restricted broader adoption. Critics noted its modest integer performance compared to its floating-point capabilities. The complex multi-chip design also added to system cost and complexity. These factors led to its relatively quick replacement by the more balanced R10000.
Questions people ask about R8000
What does the R8000 chipset include?
The chipset consists of the R8000 integer microprocessor, the R8010 floating-point unit, two Tag RAM chips, and an external streaming cache. This separation allows specialized handling of different instruction types.
Who manufactured the R8000 chips?
Toshiba fabricated both the R8000 and R8010 chips. They used their 0.7 μm VHMOSIII process, which is a triple-layer metal complementary metal–oxide–semiconductor technology.
Why was the R8000 replaced so quickly?
The R8000 had a high cost and a narrow market focused on scientific computing. The R10000, introduced in January 1996, was cheaper and generally offered better overall performance, leading to its rapid adoption.
What was the clock speed of the R8000?
The first R8000 processor ran at 75 MHz and was released in June 1994. A faster version operating at 90 MHz became available in mid-1995 for newer Silicon Graphics systems.
Good reading and sources
The facts on this page come from the Wikipedia article "R8000", 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.
Like R8000
All CPUs and microprocessorsMore from MIPS Technologies
All of them
WorldR4200
Low-power MIPS III microprocessor for embedded systemsCPU · MIPS · 1993