As I'm slowly learning CUDA, I got some of the tests running. Those under 1_Utilities actually define the parameters (stats) of the card. There are the parameters for an NVidia Quadro 600:

Device 0: "Quadro 600"
  CUDA Driver Version / Runtime Version          5.0 / 5.0
  CUDA Capability Major/Minor version number:    2.1
  Total amount of global memory:                 1024 MBytes (1073283072 bytes)
  ( 2) Multiprocessors x ( 48) CUDA Cores/MP:    96 CUDA Cores
  GPU Clock rate:                                1280 MHz (1.28 GHz)
  Memory Clock rate:                             800 Mhz
  Memory Bus Width:                              128-bit
  L2 Cache Size:                                 131072 bytes
  Max Texture Dimension Size (x,y,z)             1D=(65536), 2D=(65536,65535), 3D=(2048,2048,2048)
  Max Layered Texture Size (dim) x layers        1D=(16384) x 2048, 2D=(16384,16384) x 2048
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total number of registers available per block: 32768
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  1536
  Maximum number of threads per block:           1024
  Maximum sizes of each dimension of a block:    1024 x 1024 x 64
  Maximum sizes of each dimension of a grid:     65535 x 65535 x 65535
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 1 copy engine(s)
  Run time limit on kernels:                     Yes
  Integrated GPU sharing Host Memory:            No
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Device PCI Bus ID / PCI location ID:           1 / 0
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

So as I was reading in a document (there are quite a few!) I have the maximum number of threads available: 1536. It's probably not that important for my test, but I was really wondering because the warp size is 32 and the number of cores is 96. All of that is just not well explained in the documentation (at least for what I read so far.) I actually see that I have 2 processors, 48 cores per processor.

So... if I understand properly I could run my alpha computation in 150 "cycles" on such a CUDA card (640x480 / 1024 = 300 which we can share on both processors.) Need to test that theory. That would probably be quite a bit faster than the SIMD implementation (which is already dead fast: about 3750 full frames per second.)

More soon. 8-)