Japanese


NEC days of Ryuichi TAKAHASHI


Ryuichi TAKAHASHI joined the NEC corporation on the April of 1981. He started working for NEC as a researcher of the R & D unit after 2 months training. At the beginning of 1983, when he was 28 years old, in the second year of his NEC days, Ryuichi TAKAHASHI was sent to the semiconductor unit, which set the tone for the rest of his life.



29 years old Ryuichi TAKAHASHI in 1983. He was in the semiconductor unit.

Ryuichi TAKAHASHI ( the very left ) playing the flute in a festival of the R & D unit





Publications



1981-1988


A book:

  1. Ryuichi TAKAHASHI: "Computer system design automation in NEC Corporation"(in Japanese) pp.217-221 in a book entitled VLSI design silicon-compilation edited by Masaharu IMAI and Naoshi SUGIYAMA, Science-forum (1988)





A conference:

  1. Ryuichi TAKAHASHI, Takeshi YOSHIMURA and Satoshi GOTO: "A VLSI Architecture Evaluation System," IEEE, ACM, Proc. ICCD 1986 :International Conference on Computer Design, pp.60-63(1986)





What did yo do in the semiconductor unit?

The semiconductor unit asked the R & D unit for a help to design the first Japanese truly original 32 bit microprocessor V60. Investigating the pipelining was Ryuichi TAKAHASHI’s job for the VLSI development division of the semiconductor unit.





The first Japanese truly original 32 bit microprocessor
NEC V60 (μPD70616)


Process

CMOS 1.5µm

Transistor count

375 000

Clock speed

16 MHz, 3.5MIPS

Addressing mode

21 types

Instruction set

original 273

Floating-point capability

IEEE-754

Virtual memory

4G Byte



What did you do after that?

Ryuichi TAKAHASHI noticed that some high level design automation system is necessary to compete with Intel. He decided to develop an original hardware description language (HDL) for the microarchitecture design.

Ryuichi TAKAHASHI developed two new technologies for the high level design system:

  1. Super high speed on demand functional simulation method

  2. Hardwired microcode machine technology

which are described on the paper for the ICCD 1986.



How was the ICCD’86?

It was held at Rye Town Hilton, Rye Brook, New York since October 6th till October 9th in 1986. More than 400 engineers attended. ICCD’86 was notable for the 18 presentations by 7 companies about their 32 bit microprocessors. It was the very years when the semiconductor technology grew enough to integrate the entire components for a 32 bit microprocessor on a single chip. Intel presented 80386 by 4 papers. Motorola introduced 68030 by 1 paper. IBM introduced Micro/370 by 2 papers. DEC introduced MicroVAX by 3 papers. RISC had already appeared. There were 3 papers about MIPS. It was very impressive for me to hear Intel saying that 80386 is a CISC!

Intel architecture with 8 registers could compete with RISC with 32 registers by register renaming to manage output dependency and antidependency. For example, Pentium had a reorder buffer with 40 entries.



Where did you complete that system?

After developing the prototype system in the R & D unit, Ryuichi TAKAHASHI was sent to the communication unit to test that system in a real environment. An automaton description language design automation (A^2DL-DA) was the result.







The Abiko plant for the communication unit of the NEC corporation in 1989