I am using R for geostatistical analysis and spatiotemporal variogram calculations on a server. Although it has very high specifications but using variogramST takes a very long time (12 hours so far and still running) because I'm working on a relatively big data set. I have found some R packages that use CUDA. Is there any GPU support for geostatistics packages like gstat and spacetime?


No. Two students of mine have been working on implementing conditional simulation on a GPU; this work is found here. I'll give it a thought on how variogramST can be made faster. Have you thought about giving it a smaller data set, to begin with?

  • The data I'm using is a set of measurements in a certain region over 7 days, that's why it's a big data set. And I'm trying to study the behavior over the 7 days so I cannot reduce the size. I noticed it's very slow due to the "tunit" parameter in variogramST. When I set it to "days" it takes around 15 hours to finish. "hours" takes around 4-5 hours, and "mins" takes a few minutes – Bayan Nov 3 '15 at 11:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.