Their experimental results are unconvincing. Authors themselves admit, "we are aware that using a D flip-flop as a basic memory element is not optimal". But "we

Their experimental results are unconvincing. Authors themselves admit, "we are aware that using a D flip-flop as a basic memory element is not optimal". But "we chose a D flip-flop as a memory element since it enables embedding of the memory in a standard synthesis performed with Synopsys and Innovus". I think Logic-in-Memory should be memory first and logic second. Otherwise, we would have already used SRAM instead of DRAM. Using a D flip-flop means it really isn't DRAM-like memory. And the only use of using a standard synthesis with Synopsys is for writing a paper. It's not like DRAM vendors synthesize with Synopsys.