Monthly Archives: October 2008

Programming a Matrix Multiplication for GPUs with CUDA

CUDA makes it possible to program the GPU with the language C. This article will show you the steps to code a matrix multiplication routine in CUDA:

  • allocate memory on the GPU with cudaMalloc or cudaMallocPitch (for aligned memory allocation)
  • move data to the GPU with cudaMemcpy2D
  • select the kernel domain, write the kernel and run it
  • move results back from the GPU to the host with cudaMemcpy2D
  • free resources with cudaFree

Palit GeForce 9800 GT Super 1Gb Graphics Card

I you need a powerful graphics card to play 3D games and in the same time a video card that can be used in a HTPC (Home Theater PC), Palit has perhaps the right product with this GeForce 9800 GT 1Gb. This graphics card is equiped with an exceptionally quiet dual slot active
cooler and is available for about $170.

GeForce 9800 GT specifications:
– GPU: G92b (65nm) / 600MHz / 754-million transistors
– Memory: 1Gb GDDR3 / 256-bit / 1.8GHz effective
– Shader Processors: 112
– Textures Units: 32
– ROPs: 16
– DirectX 10 / OpenGL 2.1 / PhysX

Palit GeForce 9800 GT Super 1Gb graphics card


NVIDIA GeForce Graphics Cards Used to Break WIFI Passwords

With graphics cards such the GeForce GTX 280 and parallel brute force based algorithms coded in CUDA (GPGPU techniques), WiFi’s WPA (RC4 based) and WPA2 (AES based) encryption systems are now crackable and the time to crack them compared to CPU based algorithms is reduced by a factor of 100 (or 10,000 %).


Update (2008-10-16)