SITE MAP  CONTACT US

Volume 27 (Sep 30, 2004)

The Challenges and Passion Behind the Creation of the V850 Series (1/2)

Part 3: The Development of an IP Core for System LSI Chips

It all began with a single article...

The emergence of the V850 Series 32-bit single-chip RISC microcontroller was epoch-making in that it accelerated the transition from 16-bit to 32-bit and CISC to RISC. As the 32-bit single-chip microcontroller was gaining widespread attention as the world's highest-performance microcontroller when the press release introducing it was issued in 1994, its future potential was still unknown at that time.

It was sometime after the publication of the press release that Hiroaki Kaneko, a key player on the design team, was asked to write an article for Nikkei Electronics based on the theme, "A 32-bit CPU core created especially for built-in applications." Kaneko's article, titled "A 32-bit evaluation microcontroller incorporating all of the functions of ICE," ran in the December 5, 1994, edition of the magazine (Photo 1). The V850 Series-which had been developed using the ASIC method from the design stage-actually fulfilled all of the requirements for an IP core capable of acting as an ASIC core because its cell-based design used a library identical to the one used for ASICs. At the time, however, no one could have predicted that this article would become the catalyst for the future development of the V850 Series.

Striving to equip an ASIC with a V850 core

Shortly after the article was published, the sales division received an inquiry from a customer who had read it. Although full details of that inquiry are unknown, it is known that he expressed interest in a photograph that had accompanied the article. That photograph was of the V851 chip with its built-in V850 CPU core, peripheral circuitry and various types of memory including RAM and ROM (Photo 2). Filled with hope and anticipation that perhaps this could lead to a breakthrough in the negotiations that had just gotten off the ground for the V850 Series, Ikutaro Okuda-who was in charge of microcontroller applied technologies-and Kaneko decided to pay this customer a visit.

The visit took place at the beginning of 1995. Contrary to what they had expected, the customer was not interested in a custom microcontroller. Instead, he wanted to know if they could find a way to develop an ASIC with a built-in CPU. Without the least bit of hesitation, the two replied, "That should be no problem at all since the V850 Series is developed using the same flow and library as for ASICs." The two had absolutely no doubt that the V850 core met all of the requirements necessary for being built into an ASIC. Upon returning to the office, they asked the ASIC team for support. The other members who were assembled to work on the project were just as optimistic as Okuda and Kaneko. Yoshikazu Sakurai, who originally belonged to the microcontroller team, brought to the project the challenging spirit of a true engineer. This was NEC Electronics' first full-fledged system LSI project. The engineers on the customer side were passionate about developing this new product and were clearly enthusiastic about having the chance to collaborate with NEC Electronics. Although this should have been an enjoyable project for all of the engineers involved, that wasn't exactly the case...

Discovering cultural differences

It was now the spring of 1995. Along with the beginning of the new season, the much-anticipated project was about to start. However, dark clouds soon began to gather over this project that everyone had expected to go smoothly. As discussions with the customer progressed, the team members found question after question popping into their heads. Okuda and Sakurai finally realized that this was because there was no consensus between the microcontroller and ASIC teams.

At times, the team members even found themselves bewildered by differences in the usage of seemingly simple and straightforward terms. For example, when someone from the microcontroller team heard the term development tool, he would immediately think of in-circuit emulators or compilers. For the ASIC team, however, the term would bring to mind CAD tools. Another such example was the term buffer. The microcontroller team would use the term to mean a memory buffer, yet the ASIC team would use it to describe an I/O buffer. These differences created serious barriers and led to mutual misunderstandings. This was only natural, however, since the project was the first collaborative effort between the two teams. Whereas the microcontroller team was attempting to increase performance and reduce production costs through manual circuit design at the transistor level and through the creation of software programs, the ASIC team was focused on saving development time and labor through automation of the design process using EDA tools and the HDL design method. These differences in the usage of certain terms and the direction of design turned out to be cultural in nature. When the two teams found themselves participating in the same project, the differences between the software-centered culture of the microcontroller team and the hardware-centered culture of the ASIC team were first brought to light.

At one point, the customer asked Okuda, "How can we actually verify the timing guarantee?" Being from the microcontroller field, Okuda found that when he thought about this from an ASIC point of view, he was unsure whether or not his ideas were correct. He told the customer he would need more time to think. Even within a group of professionals, it is only natural for problems to arise when you bring together people from two completely different fields of specialization. Okuda comments, "Since our customer completely understood our situation, his comments and questions were a great help to us." Okuda returned to his office and asked the microcontroller and ASIC team members for their input. After giving it careful consideration, Okuda concluded that microcontroller techniques should be used for situations that could not be dealt with using ASIC techniques. Consequently, this choice turned out to be indispensable in making it possible to carry out the project according to schedule.

| 1   2 |

AdvancedParametric