One of the more controversial producers of math coprocessors was ULSI System Technology. ULSI was founded in 1987 by George Hwang, the goal of the company was to design ‘system on a chip’ products. The company was based in San Jose, California.
In 1989 Alfred Chan joined ULSI, he came from Intel where he had worked on the 80387 design team. At ULSI he started working on a 387 clone. The controversy began when it turned out that Mr. Chan was in possession of Intel documents about the 80387. The documents in question were early target specifications for the coprocessor. This documentation was superseded by the version that was used for the production version of the 387.
When Mr. Hwang became aware of the documents he contacted Intel and offered to let Intel search the ULSI office for ‘Intel proprietary’ documents. He wanted to prove this way that his company had not used Intel documentation in the design process of its 387 clone. Intel however was not interested and filed a criminal lawsuit against Mr. Hwang and Mr. Chan for stealing trade secrets and confidential information. (and searched the ULSI office anyway)
Meanwhile the development of the 387 clone continued and in 1991 ULSI was ready to release their US83C87. ULSI was a fables chip designer, it relied on third parties for the production. ULSI’s math coprocessor was manufactured by HP which used its CMOS34 process for production at 1.2micron.
Intel immediately filed a new lawsuit, this time a civil suit for patent infringement. In December 1991 Intel was granted preliminary injunction halting shipments of the ULSI 83C87 coprocessor. Another court overruled the preliminary injunction in January 1992 and ULSI could resume distribution of its coprocessor. Intel tried again to halt shipments but was not successful in convincing the court.