FERN: Fast Explicit Reaction Networks SVNCode

Little cleaning up

This version has been tested to give essentially the java results up to log t = -5, but diverges after that.

Fixed problem that checkAsy(Fminus[k], Y[k], dt) should be checkAsy(FminusSum[k], Y[k], dt). Because of this error it was not doing the asymptotic check correctly. Now gives essentially same results as java version up to log t = -6 for test case.

Added more diagnostic output on timesteps

cleaned up some things. This now agrees with the java example up to log t = -7 in the 3-4 decimal place up through Ar36, but quite different at log t = -6. Timestepping seems to be similar to factor of two in the two cases.

Modified explicitgpuDouble.cu to convert cuPrintf statements to printf statements in the kernel. Requires compute capability 2.0, so compile with nvcc -arch sm_20 explicitgpuDouble.cu. Without the -arch sm_20 flag the printf statements in the kernel will give compile errors.

Changed the three places where populations are updated to do serial. Now, at least for the first 100 timesteps, the CUDA and Java codes give almost exactly the same results. Thus, there seems to be an error in the parallel implementation of the population update algorithm.

update

update

Minor changes