In the semiconductor industry, Gordon Moore is undoubtedly one of the most important chip engineers in history. Moore’s Law, which states that the number of transistors that can be integrated into a chip doubles approximately every 18 to 24 months, has been regarded as a classic rule for semiconductor development.
Over the years, the semiconductor industry has continued to develop in accordance with Moore’s predictions. However, the rapid development of Moore’s Law requires not only pioneers like Gordon Moore but also countless outstanding chip design engineers who constantly pursue technological progress.
01. Legendary Chip Designer Jim Keller
In the world of chip design, there is one name that is known by all – Jim Keller. His legendary career and impressive achievements have made him a highly sought-after talent by major companies. Despite only having a bachelor’s degree, he has single-handedly shaken up Silicon Valley over the past 20 years.
Keller’s career began at DEC. In the 1960s and 70s, DEC was one of the world’s most successful computer manufacturers, and the Alpha21264 microprocessor that used RISC design was the fastest in the world at the time. Keller, who was still a newcomer to chip development, played a significant role in the development of the Alpha series.
Keller has always been an individual with a strong personality. When he was a trainee in chip development at DEC, someone asked him a hierarchical design question. Keller thought that half of what the person said made sense, while the other half was foolish. He then began to argue with the person, and an hour later, neither side had convinced the other. It was only after the person left that someone told Keller that he was Gordon Bell, the Chief Technical Officer of DEC.
In 1998, Keller joined AMD and participated in the design of the AthlonK7 processor. He quickly became the chief architect of the K8 processor, which defeated Intel’s 64-bit Itanium, allowing AMD to establish a foothold in the profitable server chip market for the first time.
Just as everyone thought Keller would make great strides at AMD, he unexpectedly left and went to the chip startup company Sibyte to lead the development of MIPS chips. A year later, Sibyte was acquired by the chip giant Broadcom, and Keller became its chief architect. However, Keller did not remain at Broadcom forever.
In 2004, Jim Keller left AMD to join a start-up company called PA-Semi, which was acquired by Apple four years later. Keller became a chip designer at Apple and led the development of the first A-series processors, including the A4 and A5, which were used in the iPhone 4 and iPhone 4S. Keller’s work at Apple was instrumental in the company’s success in the chip industry, and it also played a key role in Apple’s strategy to develop its own chips.
After leaving Apple, Keller returned to AMD as the Vice President and Chief Core Architect, where he led the development of a new microarchitecture, codenamed “Zen”. AMD was facing a major crisis at the time, and Keller’s practical approach of emulating the competition and abandoning the Bulldozer architecture became a key factor in AMD’s resurgence.
Following his success at AMD, Keller joined Tesla, where he was responsible for developing engineering technology for the company’s electric cars and led the Low Voltage Hardware department. Keller developed the FSD autonomous driving chip at Tesla, which was more than 20 times more powerful than the replaced Nvidia solution.
After his time at Tesla, Keller was hired by Intel as the Senior Vice President of the Silicon Engineering Group, where he led a team of 10,000 people. However, Keller’s time at Intel was reportedly not smooth, and he resigned in June 2020 after proposing innovative methods such as 3D stacked chips.
Keller’s goal at Intel was to simplify the development process of a large number of silicon products and to build a strategic platform for future product development. After leaving Intel, Keller became the CTO of Tenstorrent, a Canadian AI chip start-up, where he is currently focused on developing AI chips.
Jim Keller is known for his love of challenges and his desire to try different fields. Keller once said in an interview, “Tesla and Intel are another kind of the whirlwind, so you could say I jumped in and jumped out, and I really enjoyed playing with it.” Keller’s story continues, and he remains a highly respected figure in the tech industry.
02. Chief chip architect for major companies
Mike Clark, the Chief Architect of AMD Zen
Played a crucial role in the resurgence of AMD with the Zen architecture, which brought new designs and processes that resulted in a 52% increase in IPC performance, exceeding the originally planned 40% improvement. The credit for the success of Zen architecture is unparalleled, and there is ongoing debate as to who should be considered the father of Zen.
Although many people refer to Jim Keller as the “father of Zen,” when asked about it by the media, Jim simply stated, “I’m at best a crazy uncle.”
In reality, the true father of Zen architecture at AMD is Mike Clark, who has been with the company for nearly 30 years since his university graduation. He is currently a Fellow and AMD’s highest technical honor recipient.
In an interview with the media, when asked about Zen, Mike stated that he agrees with Jim Keller’s statement that there were many people involved in the development of Zen architecture. However, if someone must be singled out as the so-called father of architecture, Mike himself may be the best candidate because he has been involved with it from the very beginning since its inception in 2012, spending the past 10 years working closely with it.
Mike’s story is like a ten-year journey of sharpening a sword. In 1993, Mike graduated from the University of Illinois at Urbana-Champaign in the United States, receiving several job offers. However, he chose to join AMD because he believed it was a company that would truly allow him to pursue CPU design. Upon joining AMD, Mike began working on K5, their first basic design on x86. He also duplicated TLB, but he had no knowledge of x86 TLB and had to learn and reverse engineer how x86 TLB worked, which involved many setbacks.
At K6, Mike was responsible for integrating NexGen. When K7 came around, he became the main microcode designer for that generation. When it was time to develop the Greyhound (K9) core, Mike became the chief architect for K9. Now, he is the leader of the entire Zen core architecture division, responsible for the entire roadmap of Zen.
During the development of Ryzen, AMD faced many challenges. In 2006, AMD announced its acquisition of GPU company ATI for $5.6 billion. While this decision made AMD the first company to have both high-performance CPUs and GPUs, it also led to economic difficulties.
After the acquisition, ATI’s GPU market share began to decline. Meanwhile, Intel introduced the Tick-Tock strategy in 2006, which meant updating the manufacturing process every two years. When Intel released the Core2 series processors based on this strategy, its new core architecture allowed it to match or even surpass Athlon K8’s performance.
Although AMD subsequently launched the Phenom series processors from the K10 architecture to the Bulldozer architecture, most of these products could only attract consumers with their cost-effectiveness and more cores, and all of them were ultimately defeated by Intel.
At that time, AMD faced financial difficulties, and CPU development funds was frozen and isolated from other businesses. It was a difficult period, and the chip market required a new product every year, and AMD had to keep up. Mike admitted that one of the biggest challenges was keeping the team together, as many people chose to leave AMD, and Mike had to spend a lot of time convincing everyone that AMD would succeed.
Mike was named Zen, and he was there from the very first day in 2012, experiencing the pain and joy of Zen’s birth. Some people came and went, but he was with Zen from start to finish.
Jeff Wilcox, the Chief Chip Designer of Apple’s M1
played a crucial role in the development of Apple’s first custom-designed chip for Mac, which was released in November 2020. Later, Apple introduced the M1 Pro and M1 Max, which introduced the System-on-Chip (SoC) architecture to the Pro system and achieved breakthroughs in memory bandwidth, efficiency, and capacity.
As the Head of Mac System Architecture, Wilcox supervised Apple’s transition to custom-designed chips, including the company’s M1 chip. He was responsible for all architectural design, signal integrity, and power integrity for the Mac system. With the M1 chip, he led the Mac team to transition to Apple’s custom-designed chips and developed the SoC and system architecture for the T2 co-processor.
However, in January 2022, Wilcox announced on social media that he decided to leave Apple after an incredible eight-year journey and seek another opportunity. He expressed his immense pride in all his accomplishments at Apple, including the successful transition to Apple Silicon for the M1, M1 Pro, and M1 Max SoC and system. He will miss all his Apple colleagues and friends but is looking forward to his next journey. Currently, Wilcox has joined Intel as an Intel Fellow and CTO of the Design Engineering Group, focusing on client SoC architecture design.
Intel’s Chief Engineer, Murthy Renduchintala
Was responsible for managing, designing, and optimizing hardware architecture, whether at Intel or AMD, to earn the title of “Chief Architect,” one must have the respect of the semiconductor industry.
However, not all Chief Architects can lead chip companies to great success, as planning chip roadmaps is no easy task. In 2020, Intel announced that its Chief Engineering Officer, Murthy Renduchintala, would be leaving, likely due to delays in Intel’s planned roadmap.
Dr. Renduchintala left Qualcomm as Vice President and Co-President of Qualcomm CDMA Technologies (QCT) to join Intel in 2015, where he initially led Intel’s new business unit, the “Technology, Systems Architecture, and Client Group (TSCG).”
In 2016, he sent a memo to other senior leaders at Intel, outlining an urgent plan to address “competitive gaps,” and he was subsequently promoted to the position of Chief Engineering Officer.
Renduchintala was once considered a candidate to succeed Brian Krzanich as Intel’s CEO. Upon his arrival, he received a signing bonus of $10 million and a salary of over $25 million. By 2019, Renduchintala was one of the highest-paid executives at Intel, earning an annual total compensation of approximately $26.88 million.
From design, and engineering, to manufacturing, he was in charge of almost all hardware at Intel. Intel stated that Renduchintala’s team brought together all major technology, engineering, and manufacturing functions at Intel, including semiconductor process technology, manufacturing, and operations, system and product architecture, IP development, design and on-chip system engineering, software, and security, as well as management of Intel Labs.
When Renduchintala joined Intel, his task was to help Intel expand beyond the PC-era central processing unit (CPU) to a larger market. However, Intel soon sold the business to Apple for $1 billion.
At that time, Intel’s 10nm process technology had basic indicators even stronger than TSMC’s 7nm process technology, but Intel believed that 10nm was not the best node. To make up for the delay in 10nm chips, Intel was determined to work hard on 7nm to ensure timely delivery.
However, Murthy Renduchintala made serious mistakes in leading the team to develop 10nm and 7nm processor architectures, resulting in a delay of about 6 months in Intel’s 7nm process technology development progress from the original plan, and a delay of about a year in product launch time.
As a result, Intel’s Chief Engineering Officer Murthy Renduchintala announced his resignation shortly after announcing the delay.
