{"id":103157,"date":"2022-02-01T15:36:44","date_gmt":"2022-02-01T13:36:46","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/metod-identifikaczii-sistemy-polzovatelya-na-osnove-informaczii-o-gpu"},"modified":"2022-02-01T15:36:44","modified_gmt":"2022-02-01T13:36:46","slug":"metod-identifikaczii-sistemy-polzovatelya-na-osnove-informaczii-o-gpu","status":"publish","type":"post","link":"https:\/\/prohoster.info\/en\/blog\/news\/metod-identifikaczii-sistemy-polzovatelya-na-osnove-informaczii-o-gpu","title":{"rendered":"User system identification method based on GPU information","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"
Researchers from Ben-Gurion University (Israel), the University of Lille (France), and the University of Adelaide (Australia) have developed a new technique for identifying user devices by analyzing GPU performance parameters in web browsers. The method, named 'Drawn Apart', is based on the use of WebGL to create a GPU performance profile, significantly enhancing the accuracy of passive tracking methods that operate without cookies and do not store identifiers on the user's system. <\/p>\n
Previous methods considering rendering characteristics, GPU, graphics stacks, and drivers for identification were limited to distinguishing devices only at the level of different graphics card models and GPUs, serving merely as an additional factor to increase the likelihood of identification. A key feature of the new 'Drawn Apart' method is that it does not restrict itself to differentiating between GPU models but seeks to identify differences among identical GPUs of the same model, caused by variations in the manufacturing process of chips designed for massive parallel computations. It is noted that the variations that arise during production allow for the creation of unique fingerprints for the same device models.
<\/center> <\/p>\n
It turned out that these differences can be detected by counting the number of execution blocks and analyzing their performance in the GPU. Primitive tests used to differentiate different GPU models included checks based on a set of trigonometric functions, logical operations, and floating-point calculations. To identify differences in identical GPUs, the number of concurrently executed threads while executing vertex shaders was evaluated. It is hypothesized that the observed effect is due to variations in temperature settings and power consumption among different chip instances (a similar effect was previously demonstrated for CPUs\u2014identical processors executing the same code exhibited different power consumption). <\/p>\n
Since operations via WebGL are executed asynchronously, it is not possible to directly use the JavaScript API performance.now() for measurement of their execution time. Thus, three tricks were proposed for measuring time: <\/p>\n
\n
onscreen \u2014 scene rendering on an HTML canvas with measurement of the callback function's execution time, set through the Window.requestAnimationFrame API and called after the rendering is complete.\n
offscreen \u2014 using a worker and rendering a scene in the OffscreenCanvas object with timing of the convertToBlob command execution.\n
GPU \u2014 rendering in the OffscreenCanvas object, but using the provided timer in WebGL to measure the duration of command set execution on the GPU side. <\/ul>\n
During the creation of the identifier on each device, 50 checks were performed, each covering 176 measurements of 16 different characteristics. An experiment collecting information on 2500 devices with 1605 different GPUs demonstrated a 67% increase in the efficiency of combined identification methods with the addition of support for Drawn Apart. Specifically, the combined method FP-STALKER averaged identification within 17.5 days, while in conjunction with Drawn Apart, the identification period increased to 28 days.
<\/center> <\/p>\n
\n
The accuracy of separation for 10 systems with Intel i5-3470 (GEN 3 Ivy Bridge) chips and Intel HD Graphics 2500 GPU in the onscreen test was 93%, while in offscreen \u2014 36.3%.\n
For 10 systems with Intel i5-10500 (GEN 10 Comet Lake) and an NVIDIA GTX1650 graphics card, the accuracy was 70% and 95.8%.\n
For 15 systems with Intel i5-8500 (GEN 8 Coffee Lake) and Intel UHD Graphics 630 GPU \u2014 42% and 55%.\n
For 23 systems with Intel i5-4590 (GEN 4 Haswell) and Intel HD Graphics 4600 GPU \u2014 32.7% and 63.7%.\n
For six Samsung Galaxy S20\/S20 Ultra smartphones with Mali-G77 MP11 GPU in the onscreen test, the identification accuracy was 92.7%, while for Samsung Galaxy S9\/S9+ smartphones with Mali-G72 MP18 \u2014 54.3%. <\/ul>\n
<\/center> <\/p>\n
It was noted that the accuracy was affected by the GPU temperature, and for some devices, a system reboot led to distortion of the identifier. When using the method in combination with other indirect identification methods, the accuracy can be significantly increased. There are also plans to enhance accuracy through the use of compute shaders after stabilizing the new WebGPU API. <\/p>\n
Companies like Intel, ARM, Google, Khronos, Mozilla, and Brave were notified about the issue back in 2020, but the method's details have only now been revealed. Researchers also published working examples written in JavaScript and GLSL, which can operate with and without screen output. Additionally, datasets for classifying extracted information in machine learning systems have been released for Intel GEN 3\/4\/8\/10 based GPU systems. \n Source: opennet.ru<\/a> <\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"
![\"User](\"\/wp-content\/uploads\/2022\/02\/947fc4477ff99c14ca7854e36e4ad976.jpg\")
<\/center> <\/p>\n![\"User](\"\/wp-content\/uploads\/2022\/02\/c57096ca8d5833df92cd7f992f19a289.jpg\")
<\/center> <\/p>\n![\"User](\"\/wp-content\/uploads\/2022\/02\/e5462f5b761dc6e16a43ac00d8f0b7aa.jpg\")
<\/center> <\/p>\n