dp sdk montecarlo and reduction examples

Accelerated Computing CUDA CUDA Programming and Performance
1

See attached
montecarlo uses floats by defualt. run with -type=double
reduction uses integers by default. run with -type=float, -type-double or -type-int

I’d suggest installing beta 2.0 sdk and rename the montecarlo and reduction projects and drop these new ones in.

see makefile or vs2005 project settings (custom build steps) for examples of building with -arch sm_13
dp_reduce_monte.zip (50.9 KB)

2

This will make a nice addition in the SDK indeed. Any chance you will add Kahan on GPU for the reduction exampe (as an option)? If you want I have some code implementing Kahan on GPU, it is just not parameterized like the reduction example, but it is working.