Source: tianji.com

    On June 30, Samsung Electronics officially announced that it had started mass production of 3nm chips in Huacheng factory in South Korea. Unlike FinFET chips used in previous generations, Samsung uses GAA transistor architecture, which can greatly improve the power and efficiency of the chips.

    Compared with the previous 5nm process, the power consumption of the new generation 3nm process is reduced by 45%, and the performance is improved by 23% while the area is reduced by 16%. Samsung also announced that the second generation 3nmgaa manufacturing process is still under development. The next generation process will reduce the power consumption of the chip by 50%, improve the performance by 30% and reduce the area by 35%. Samsung Electronics said that its GAA transistor chip will be used in the field of high-performance, low-power computing, and plans to expand to mobile processors. According to foreign media, Shanghai Pansi semiconductor, a Chinese virtual currency miner chip design company, will become the first customer of Samsung Electronics‘ 3nm process, and Qualcomm has also booked the capacity of Samsung‘s 3nm process.

    Samsung Electronics‘ 3nm chip adopts GAA architecture to effectively improve power by reducing power supply voltage and enhancing the ability of driving current. In addition, Samsung has also applied nano chip transistors in semiconductor chips of high-performance smart phone processors. Compared with nanowire technology, the former has wider channels, higher performance and efficiency. Samsung customers can customize the power consumption and performance indicators they need by adjusting the width of nano chips.

    For comparison, TSMC and Intel will start mass production of 3nm process chips in the second half of this year and the second half of next year respectively. In fact, it is no wonder that Samsung has chosen such a radical approach. It has worked so hard to mass produce 3nm chips in front of TSMC, and to achieve reverse superovulation. As we all know, TSMC is the most advanced chip foundry in the world, and its technology has always been at the top of the world. However, Samsung, which started relatively late, is indeed weaker than TSMC in terms of technology.

    The most intuitive thing is that during the 5nm process, the Xiaolong 8 manufactured by Samsung‘s 5nm process overturned early, forcing Qualcomm to hand over all subsequent orders to TSMC, while the Xiaolong 8+ manufactured by TSMC is far better than the Xiaolong 8 manufactured by Samsung‘s 5nm process in terms of performance and power consumption. This is mainly because the difference in transistor density between Samsung‘s 5nm process and TSMC‘s 5nm process is 35%. Therefore, even NVIDIA announced that the next generation of 40 Series graphics cards would be built by TSMC‘s N4 process, without Samsung.

    A good Samsung can‘t accept this gas, so Samsung planned early that it must lead TSMC at 3nm and achieve reverse superelevation. Samsung has been in the dilemma of yield rate before. According to rumors, the yield rate of Samsung‘s 4nm process is only 35%, which is too low to be accepted by Qualcomm and NVIDIA. Therefore, Samsung thoroughly investigated the capital flow used to improve the yield rate internally.

    However, TSMC was not idle and announced early that it would plan to achieve mass production of 2nm chips by 2025. The arms race between the two giants in the chip industry is about to begin. In other words, it has never stopped, but now it is full of gunsmoke.

Disclaimer: This article is transferred from other platforms and does not represent the views and positions of this site. If there is infringement or objection, please contact us to delete. thank you!         中恒科技ChipHomeTek

搜索

复制