ElcomSoft adds AMD Radeon 7000 Series support
ElcomSoft updated its Phone Password Breaker and Wireless Security Auditor with OpenCL support, adding support for the latest AMD Radeon Graphics and ensuring universal acceleration hardware compatibility.
The two products are now using updated acceleration engines, further increasing the speed of password recovery with compatible AMD video cards. In addition, Elcomsoft Wireless Security Auditor can now employ several new types of attacks, exploiting potential wireless password weaknesses introduced by the human factor.
“By employing OpenCL in our products, we’re getting future-proof access to all the acceleration features available in current and future AMD hardware, add support for the latest AMD hardware and increase password recovery speed”, says Andrey Belenko, ElcomSoft Chief Security Analyst. “We could have handled NVIDIA accelerators the same way, via OpenCL, but at this stage, NVIDIA CUDA software gives us that last bit of extra speed.”
As an immediate benefit, ElcomSoft products gain compatibility with AMD RadeonTM HD 7000 Series hardware, a state-of-the-art graphic accelerators currently including AMD RadeonTM HD 7750, 7770, 7950, and 7970 models.
In addition, the use of OpenCL allowed ElcomSoft to additionally optimize the performance of its password recovery tools executing on the AMD platform to gain a further performance benefit of up to 50% compared to older versions of its software. At this time, acceleration algorithms for the NVIDIA platform are still relying on the proprietary CUDA software for extra speed delivered by that platform.
Additional enhancements have been made to Wi-Fi password recovery algorithms in Elcomsoft Wireless Security Auditor. The newly added types of attacks target potential password weaknesses introduced by the human factor. “Today’s Wi-Fi networks are all using cryptographically sound encryption”, says Vladimir Katalov, ElcomSoft CEO.
“With protocol-enforced minimum password length of 8 characters (WPA/WPA2 PSK), brute-force attacks make little sense. Relying on dictionary attacks and the human factor greatly increases the chance of timely recovery – or, rather, revealing security weaknesses in existing wireless network infrastructure,” he added.
Background
Attacking long, complex cryptographically strong passwords is a lengthy and time-consuming operation. During the last few years, ElcomSoft has been employing sheer processing power provided by today’s gaming video cards to accelerate password recovery. Today’s high-end video cards made by NVIDIA and AMD allow gaining the speed benefit of several dozen times compared to CPU-only operation.
NVIDIA was first on this market, releasing a CUDA application programming interface to enable developers use NVIDIA graphical acceleration hardware in non-graphical applications. CUDA provides universal compatibility among the many generations of NVIDIA graphical accelerators.
The same cannot be said about AMD. Its Stream SDK was lagging behind the releases of latest hardware, making it impossible to implement proper support for newer, more powerful GPU units in ElcomSoft password recovery products.
Recently, AMD released a new APP SDK allowing developers enable the use of AMD GPUs to accelerate applications beyond just graphics.
Available as part of an independently maintained open framework, Open Computing Language (OpenCL), the new SDK allows developing applications executing across heterogeneous platforms mixing CPUs, GPUs and other processors. OpenCL gives any application access to the graphics processing unit for non-graphical computing.