Home
Name Modified Size InfoDownloads / Week
README.txt 2023-05-05 1.2 kB
data.mat 2023-05-04 2.9 GB
Totals: 2 Items   2.9 GB 0
data.mat is a  MATLAB® formatted data (v7.3) containing SRAM startup data from 345 standalone commercial-off-the-shel (COTS) SRAM chips (of size 4Mbit). To load the data in MATLAB® workspace, use the load() command. The data is stored in a variable named "sram", which is a nested structure array. To access the startup data from a specific chip in a specific operating condition, follow the below syntax-
x = sram.<manufacturer_tag>.<part-number_tag>.<chip_tag>.<operating_condition>
You can find the manufacturer name, part-number tag, and operating condition details in the referenced paper. With the above command, x will obtain a 256Kx20 16-bit unsigned integer array. We took multiple measurements of SRAM startup data to analyze the noise impact and each column of the array contains the data from a single measurement (each SRAM contains a total of 256K addresses). As each address contains 16-bit data, we packed each address data in a 16-bit unsigned integer.

If you use this data, please reference the following article:
B. M. S. Bahar Talukder, Farah Ferdaus, and Md Tauhidur Rahman, “A Non-invasive Technique to Detect Authentic/Counterfeit SRAM Chips,”  in ACM Journal on Emerging Technologies in Computing Systems, 2023. 
Source: README.txt, updated 2023-05-05