Many people believe that the future is GPGPU computing and that the parallelism of graphics cards can speed up a large number of tasks. However, researchers from Greece and the United States have come up with a GPU-assisted malware that is very difficult to detect by antivirus.
For the moment, users can rest assured, since the two pieces have been developed by a group of researchers belonging to an American university in order to demonstrate that the execution of this type of code using the graphics processor and ignoring the CPU implies a greater risk for the user, since the computational benefits of this type of programs are increased.
The GPU malware that were created
The first uses the GPU to decompress the malicious code, in this way, being compressed at the beginning and not being executed through the CPU, in addition, the specific algorithm can also be changed on a regular basis, which makes it even more difficult. its detection. And using the raw-power of the GPU, it can be used to hide code that would not be feasible on the CPU. Jellyfish and Demon Keylogger have been the two creations that have been published on GitHub after several months of work. Although at the beginning we mentioned that the benefit obtained is at the processing level, the reality is very different and cybercriminals could obtain other and very important benefits.
The problem with this is that these 2 malwares benefit from something that is not currently found: an analysis of the code execution using the computer’s GPU. Both are compatible with Nvidia and AMD graphics cards and it is more than likely that this study will help those responsible for security tools to improve and anticipate what the trend may be in the coming years.
At the moment this is not a problem at present but… as we well know, cybercriminals do not stop and we must be prepared for these since they could be found in the future.
Check also:
Silver Sparrow, A malware that affected macOS
PoS Malware is a Virus that steals financial data
[…] reads:Malware based on GPU, a possible future problemSpecter – Which problems this vulnerability […]