Home / Component / CPU / How to crack a password? use a HD5770, not your CPU

How to crack a password? use a HD5770, not your CPU

We all love our graphics cards, playing the latest games and being able to crank the eye candy is one of life’s little luxuries. There are other uses for video cards however, and one of them showcases the bruteforce capabilities for cracking passwords.

Vijay Devakumar, from India said that GPU’s are now being used for other purposes, aside from gaming, such as handling highly parallel duties. It isn’t easy however coding for a graphics card.

Devakumar has used his lower end HD5770 to crack passwords quickly and painlessly. He has a demo showing how to hack Windows logon passwords for instance. He also compares against a free CPU based security tool Cain and Abel.

AMD HD5770: password cracking powerhouse ... no really.

To give an indication of the brute force speed differences, a password with five characters took Cain 24 seconds to crack, rating 9.8 million passwords per second. ighashgpu found the same thing in less than one second, generating over 3.334 billion passwords before landing on the correct one.

Six character passwords prove harder to crack, taking Cain software 1 hour and 30 minutes to break while ighashgpu got it in 4 seconds. A seven character password, took Cain four days to crack while taking only 17 minutes and 30 seconds via ighashgpu.

Vijay Devakumar then tried a nine character password, which would have taken Cain around 43 years to work out. ighashgpu took 48 days.

Obviously the HD5770 is no power house, not when compared to cards such as the HD6990 or GTX580, but it just shows how much better the GPU architecture is for brute force routines. The HD6990 would be 3 or 4 times faster than the HD5770 at the same task.

KitGuru says: The moral of the story? keep your passwords as long as possible.

Become a Patron!

Check Also

Samsung’s RDNA 2 powered Exynos SoC outperforms Snapdragon 8 in GPU tests

Recently, Samsung unveiled its new Exynos 2200 chip with embedded RDNA 2 graphics thanks to …