In this article, we analyzed the stability, power, performance, and area of 6T-SRAMs using a promising scaling booster, i.e., source/drain patterning (SDP) scheme for the 3-nm technology node based on 3-D TCAD simulation. SDP scheme allows to decrease the spacing between transistors by downsizing the source/drain epitaxy. Proposed 5-nm and 3-nm SDP-SRAM (SDP-SRAM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">5</sub> and SDP-SRAM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> ) are compared with conventional 5-nm node SRAM (Conv-SRAM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">5</sub> ), quantitatively. Unlike a Conv-SRAM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">5</sub> , SDP-SRAMs have more design margins according to the two fin position parameters: fin location adjustment (FLA) and separation pitch adjustment (SPA). The optimized layout has the maximum FLA by significantly decreasing the back-end-of-line (BEOL) bitline and internal node capacitance. Read and write static noise margins were comparable (<; 10 mV) between the SDP-SRAM and Conv-SRAM. The SDP-SRAM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">5</sub> improved the read access time 18% compared to the Conv-SRAM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">5</sub> in low-power (LP) applications, but the SDP-SRAM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> degraded it by 11% in high-performance (HP) applications because of the narrow M2 pitch. Both the SDP-SRAM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">5</sub> and SDP-SRAM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> improved write access time about 10% compared to a Conv-SRAM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">5</sub> in LP applications. The cell area of the SDP-SRAM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</sub> decreased about 0.64 times compared to a Conv-SRAM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">5</sub> without severe degradation of performance. The SDP scheme provides flexible layout designs enabling a high-density SRAM, promising for 3-nm node logic applications.