Cyrix was an US-American company, which manufactured microprocessors. It was created 1988 by former coworkers of Texas Instruments and was one fabless company, possessed thus no own production plants. Instead one left on behalf with SGS Thomson (ST) and IBM, later then with national Semiconductor and at the conclusion with TSMC manufacture.
The first products of the company were mathematical coprocessors for 80386-CPUs. one brought the first adequate x86-Mikroprozessor to 1992 then with the Cx486 on the market. CCUs with extended current savings techniques and Upgrade CCUs for 386er-PCs were sold.
one brought 1995 with the Cyrix 5x86 and the Cyrix 6x86 two very interesting CCUs on the market:
The Cyrix 5x86 (code name M1sc) was a Upgrade CCU for 486er-Hauptplatinen and was based on the Design of the 6x86 (M1).
The Cyrix 6x86 (code name: M1) however had been developed for the base 5 and so pin-compatibly to Intel the Pentium.
It pay attention that the Integer achievement of the 6x86 was so far superior Intel the Pentium that a CCU with only 133 MHz was faster than a Pentium 166. Therefore the so-called P-Rating (PR) was introduced, so that the CCU were comparable.
As reaction to Intel the Pentium MMX, Intel the Pentium II and the AMD K6 (everything with MMX technology), placed Cyrix to 30. May 1997 the Cyrix 6x86MX (M2) forwards. This CCU was based to a large extent on the old Cyrix 6x86, however with a larger L1-Cache, an improved jump forecast and with a MMX compatible technology (EMMI) was equipped.
All in all a quite good CCU, owing to the improvements within the Integer range to play trumps knew and again a higher achievement than the competition products with same clock frequency reached. For this reason again a P-Rating was used. A further change to the 6x86 was the use of clearly more flexible multiplicators: x2,5 became possible and thus a substantially larger pallet at clock frequencies.
Up to then it saw completely well technically actually out for Cyrix: One had a competitive CCU, a relatively high market share and many patents. Financially it ran less well: One made losses, which cost fight with Intel and AMD much money for years, and in the long run also the AMD K6 had been able to win many trailers, which pressed the sales figures of the 6x86MX.
Given in November 1997 then the assumption through national Semiconductor admits. With NatSemi in the back one went out with it that Cyrix could offer also in the future competitive products, with the Cayenne/Gobi Design was an appropriate CCU in the development.
On 15 April 1998 then the first disappointment took place: Cyrix presented the Cyrix MII on this day. This CCU was however not new, but it concerned only a renamed 6x86MX. First MII were the MII-300 and the MII-333. Surely these CCUs were not bad, architecture were however already 1 year old and against AMDs K6-2 and Intel Pentium II could one thereby - v. A. also because of the low clock rates - not exist. IBM took part in the mode feature of the change of name by the way not and sold the MII-300 and -333 further as 6x86MX-PR300 and 6x86MX-PR333.
Cyrix dropped back, the sales sagged. Now also it turned out that NatSemi more was interested in SoCs (system on chip) than in Desktop CCUs. That was naturally catastrophic for the development of a new Desktop CCU. In the middle of 1999 came then the message that NatSemi Cyrix sold on VIA Technologies (Taiwan). When first measure gave VIA the production of the MII CCUs at TSMC, and owing to a the Shrinks on 0,18 the clock rates could are easily increased, and one brought still the MII-366, MII-400 and MII-433 on the market.
At the beginning of of 2000 was then presented finally again a new CCU under the name VIA Cyrix III to the public for the first time. This CCU was already so long in the development that it had gotten equivalent three different code names: Originally with the development under the name "Cayenne" still begun as Cyrix still independent was. After the fusion with NatSemi the code name in "Gobi" was changed. But in that well two years of the affiliation to NatSemi the CCU could not be finished. Only after the over name through VIA Technologies the CCU under the code name became "Joshua" finished.
Due to the long development time one could actually expect a matured product, but the VIA Cyrix III was only on the paper a very interesting CCU for the base 370 and disappointed in first Vorabtests: Instabilities, high heat development and very low clock rates connected with completely utopian P-Ratings (there much too highly set) led to the fact that VIA Technologies the Joshua Design gave up completely and instead as VIA Cyrix III sold the Samuel developed by Centaur Technology.
As consequence from the whole one it was to be heard at that time that Cyrix was closed. Lately it gave on the part of VIA however a few suggestions that the Cyrix Design team would still exist and on new CCUs was worked. Which however at this rumor to is, is to be evaluated not really. The future will show it.
The CCU technology of Cyrix is only used at present into the Geode CCUs by national Semiconductor. The Geode SoCs of the Cyrix MediaGX is based and by national Semiconductor was easily improved on the Design. The SoC section was transferred in the year 2003 by AMD. AMD offers the same product under the name to Geode GX. In the meantime 2005 with that were published Geode LX far improved version of the old Cyrix MediaGX. The Geode of NX ONE likewise offered by AMD however completely of the AMD Athlon XP is based on the architecture.
On 24 October 2005 AMD has the x86-Technik von Cyrix to the Chinese Ministry for science and technology (Chinese Ministry OF Science and Technology - MOST) and to the University of Peking for the development of embedded and low power x86-Komponenten licensed. Thus the technology will in the future be used again more intensively by Cyrix and also developed further.
We found here 18 articles.
Index | Privacy | Terms Of Use | Sitemap | Feedback