Naive OpenCL benchmarks..

Used for selecting GPU devices for ATI
export DISPLAY=:0.0
export DISPLAY=:0.1
For switching Nvidia and ATI see previous blog ..
For fixing MersenneTwister on Nvidia see my blog

Timing:Assure to use
clEnqueueNDRangeKernel(queue, kernel, 1, NULL, global_work_size, NULL, 0, NULL, &event);
clWaitForEvents(1, &event);
    long long profiling_start,profiling_end;
    clGetEventProfilingInfo(event, CL_PROFILING_COMMAND_START, sizeof(long long), &profiling_start, NULL);
    clGetEventProfilingInfo(event, CL_PROFILING_COMMAND_END, sizeof(long long), &profiling_end, NULL);

My tests
=====
core i7 920 gtx 275 ati 5850 ati 4850
linux
./MersenneTwister -q -t --device cpu -x 1000000
cpu
2000000                  0.259                    7.72201e+06
w7
2000000                  0.328564                 6.08709e+006  
gpu
2000000                  1.582                    1.26422e+06 
w7
2000000                  0.66877                  2.99057e+006    
r770
2000000                  2.911                    687049
gt200 (190)
2000000                  1.371                    1.43962e+06
195
2000000                  0.675                    2.96296e+06
195 w7
2000000             0.733789                 2.72558e+006
./MatrixMultiplication --device gpu -x 2048 -y 2048 -z 2048 -t -q

cpu
44sec
cpu w7
2048x2048                2048x2048                56.4574
gpu
2048x2048                2048x2048                1.34713 
w7
2048x2048                2048x2048                0.960548
rv770
2048x2048                2048x2048                3.80943   
gt200 (190)
2048x2048                2048x2048                1.77975
195
2048x2048                2048x2048                0.641499
195 w7
2048x2048                2048x2048                0.60038
w7 oc
2048x2048                2048x2048                0.55

Updated for 195 drivers

ATI FORUMS
========

2.66GHz core 2 Duo running Ubuntu 9.04, and I have a Radeon 4350 GPU.  I'v
ati.txt
$ ./MatrixMultiplication --device gpu -x 2048 -y 2048 -z 2048 -t -q
MatrixA                  MatrixB                  Time(sec)               
2048x2048                2048x2048                95.0832                 

$ ./MatrixMultiplication --device cpu -x 2048 -y 2048 -z 2048 -t -q
MatrixA                  MatrixB                  Time(sec)               
2048x2048                2048x2048                81.8752                 

$ ./MersenneTwister -q -t --device cpu -x 1000000
Generated Numbers        Time(sec)                Numbers/sec             
2000000                  0.506                    3.95257e+06             

$ ./MersenneTwister -q -t --device gpu -x 1000000
Generated Numbers        Time(sec)                Numbers/sec             
2000000                  4.133                    483910  

Phenom X4 9650 and Radeon 5770 on vista 32bit SP2.
MatrixMultiplication.exe --device cpu -x 2048 -y 2048 -z 2048 -t -q
MatrixA             MatrixB            Time          KernelTime
2048x2048      2048x2048        140.612        139.926

MatrixMultiplication.exe --device gpu -x 2048 -y 2048 -z 2048 -t -q
MatrixA             MatrixB           Time          KernelTime
2048x2048      2048x2048         3.3639         0.826539

MersenneTwister.exe -q -t --device cpu -x 1000000
Generated Numbers     Time            kernelTime    Numbers/sec
2000000                      0.976359     0.1577           2.04843e+006

MersenneTwister.exe -q -t --device gpu -x 1000000
Generated Numbers     Time             kernelTime   Numbers/sec
2000000                      1.89666        0.0705628    1.05449e+006