ElcomSoft announced the support of NVIDIA Fermi chips for accelerating password recovery even further. By using NVIDIA’s newest high-end video accelerators for its password recovery tools, ElcomSoft has achieved benchmark speeds demonstrating 20-fold gain in recovery speed compared to Intel current top of the line quad-core CPUs.
The new Fermi-based video boards exceed benchmark speeds of previous-generation NVIDIA Tesla solutions by doubling the password recovery speed of enterprise-grade Tesla solutions based on the previous-generation chipsets.
CPUs are not going away any time soon, but GPUs found in modern video cards such as NVIDIA GTX 480 are taking over. Depending on applications, CUDA-based GPUs can contribute more to the total performance of the PC than the central processing unit. In some applications, their contribution is orders of magnitude greater than the role of a modern CPU.
Until recently, GPUs were used almost exclusively for gaming and the drawing of some elements of the graphical user interface. The invention of CUDA-based architecture allowed GPUs to be used in other computational-heavy applications, such as password recovery tools manufactured by ElcomSoft.
The availability of hundreds of massively parallel processing units allowed ElcomSoft developers to offload major computations onto GPU units, increasing the performance of its password recovery tools some twentyfold compared to top of the line quad-core CPUs manufactured by Intel and AMD.
About NVIDIA Fermi
NVIDIA’s next-generation CUDA architecture code named Fermi is the latest and greatest GPU chipset. Adding a host of new features over the previous-generation chipset, Fermi boards double the number of available GPU units and increase memory throughput 30 to 50%. Fermi-based accelerators get a significant performance boost compared to older generation units. The latest NVIDIA GPUs set new benchmarks, bringing supercomputer performance to more homes and offices than ever before.