TMSC’s 3nm Update: N3P and N3X on Track with Density and Performance Gains

TSMC disclosed A major roadmap update for the N3 (three-nanometer class) family of process technologies was announced at the 2023 North American Technology Symposium this week. As TSMC’s final high-performance node based on FinFET transistors, N3 will last for years, optical shrink to enhance performance of N3P, N3E, and performance focus for HPC applications that tolerate high leakage and power Includes multiple versions including the N3X of .

Mass production of TSMC’s N3 (also known as N3B) process technology is already underway, but the node uses extreme ultraviolet lithography on up to 25 layers and can also use EUV double patterning, making it a viable option. becomes a particularly expensive node for As a result, TSMC expects the majority of customers to use N3E, which allows him to use EUV on up to 19 layers, does not use double-patterning EUV, has a wider process window and higher yields. N3E, which will go into mass production in the second half of 2023, will also be the basis for TSMC’s further 3nm evolution.

N3P is the first step in that evolution. This technology is primarily an optical reduction of N3E, with several other improvements that allow 5% performance improvement for the same leakage, 5% to 10% power reduction for the same clock, and 4% higher transistor density. is characterized by A mixed chip consisting of 50% logic, 30% his SRAM and 20% analog circuitry.

As an optical shrink of N3E, N3P preserves its design rules and allows chip designers to reuse N3E IP in new nodes. This is very important as IP design houses such as Ansys, Cadence and Synopsys already have a lot of IP for N3E chips. Optical shrink, on the other hand, means increased density for all kinds of transistors and circuits, including SRAM. SRAM is the type of circuit that has struggled to shrink in recent years (especially bad with his modern SRAM-heavy designs). N3P He expects to be ready for mass production in 2024.

Following N3P, TSMC plans to further expand the N3 family and branch out into high-performance computing applications such as CPUs and GPUs with N3X. This manufacturing process is projected to provide at least 5% higher frequencies compared to N3P and also allow significantly higher voltages. This further increases the clocks at the cost of higher overall leakage.