We recently invested in faster GPUs and optimizing our processing. The results are obvious.
Here are some of the last few jobs that we have run. (The radius is in meters.)
16-19.3km/10-12mi jobs are now running in under 90 mins for most cases. (2-5x faster than before!)
3-5km/1.5-2mi jobs are running in 5-10 mins.
Cool! Can you share the specs on this new cnHeat server and GPU’s?
So I didn’t mean to mislead in my original post. We are built in the cloud (AWS). But we migrated our servers to a new platform in AWS for improved speed. Each GPU went from 2560 cores to 18176 cores. And we went from 16GB of GPU RAM to 48GB of GPU RAM.
All in all it’s a really nice boost in performance.
Any chance we could extend the distance a bit more now that you have more compute power? The 450m 3Ghz can have subs going out to 30KM due it’s high output power.