Cyrix was founded in early 1988 by Jerry Rogers and Tom Brightman.
Interestingly however is the fact that the two founders first named the company
APT Corporation. Rogers and Brightman had both worked for Texas Instruments
and had a couple of business ideas. One was producing x87 compatible math coprocessors
another producing a MAC clone.
They hired Bill Briggs to work with them on the business plans for the different ideas. This all happened in late 1987 and at the beginning of 1988. At the time they were located in the Galleria, Dallas right next to the office of Sevin and Rosen.
Sevin and Rosen
were venture capitalists; they brought together money to invest in promising
start-ups. They had already helped start ups Compaq and Lotus with venture money.
Sevin and Rosen invested about $4 million and APT Corporation issued about 10
million shares. These shares were devided between Sevin and Rosen Jerry, Tom
and other early employee’s.
Over the period of a few months APT Corp. hired a total of 17 employees and they got to work. The goal was to produce a coprocessor that was ten (!) times faster than the Intel 80387. To calculate their performance goal they simply took the Intel datasheet and divided the clock counts by ten. With this in mind the resulting design would easily be much faster that Intel’s math coprocessor. Only later the design team realized that with such a fast coprocessor the CPU would be the bottleneck. That is in fact what happened, the CPU was busy decoding instructions and communicating with the memory and de coprocessor, all the while using more clock cycles than the coprocessor needed to calculate its instructions. The coprocessor had to wait for the CPU for instructions and communication back the results.
The design team used new ideas for the functions of the math coprocessor. The implementation for Divide and Square root were based on ideas from David W. Matula. Transcendal work was inspired by Tom Brightman and Warren Fergusen.
The first design was code named L7 for Lucky Seven and also a reference to L.J. Sevin. Revision 1.0 of the two metal layer chip had some problems and worked only in the lab.
In the lab the engineers could run the chips while they were still on the wafer. With the use of a ‘probe’ card, a chip tester and a stepper the functions of the chip could be checked.
Shortly before the tape out of the first design, APT Corporation changed its name. Apparently Sevin and Rosen liked two syllable names that started with a "C" (Compaq & Cypress, companies they helped to set up earlier) and Cyrix was born!
They hired Bill Briggs to work with them on the business plans for the different ideas. This all happened in late 1987 and at the beginning of 1988. At the time they were located in the Galleria, Dallas right next to the office of Sevin and Rosen.
Sevin and Rosen
were venture capitalists; they brought together money to invest in promising
start-ups. They had already helped start ups Compaq and Lotus with venture money.
Sevin and Rosen invested about $4 million and APT Corporation issued about 10
million shares. These shares were devided between Sevin and Rosen Jerry, Tom
and other early employee’s.Over the period of a few months APT Corp. hired a total of 17 employees and they got to work. The goal was to produce a coprocessor that was ten (!) times faster than the Intel 80387. To calculate their performance goal they simply took the Intel datasheet and divided the clock counts by ten. With this in mind the resulting design would easily be much faster that Intel’s math coprocessor. Only later the design team realized that with such a fast coprocessor the CPU would be the bottleneck. That is in fact what happened, the CPU was busy decoding instructions and communicating with the memory and de coprocessor, all the while using more clock cycles than the coprocessor needed to calculate its instructions. The coprocessor had to wait for the CPU for instructions and communication back the results.
The design team used new ideas for the functions of the math coprocessor. The implementation for Divide and Square root were based on ideas from David W. Matula. Transcendal work was inspired by Tom Brightman and Warren Fergusen.
The first design was code named L7 for Lucky Seven and also a reference to L.J. Sevin. Revision 1.0 of the two metal layer chip had some problems and worked only in the lab.
In the lab the engineers could run the chips while they were still on the wafer. With the use of a ‘probe’ card, a chip tester and a stepper the functions of the chip could be checked.
Shortly before the tape out of the first design, APT Corporation changed its name. Apparently Sevin and Rosen liked two syllable names that started with a "C" (Compaq & Cypress, companies they helped to set up earlier) and Cyrix was born!
