3 ms·
We love what MapD is doing. We operate in different spaces. We have no visualization platform nor intention of making one. Instead of trying to solve problems i
by felipe_aramburu 10y ago
We love what MapD is doing. We operate in different spaces. We have no visualization platform nor intention of making one. Instead of trying to solve problems in GPU RAM we focus on solving problems that require disk based storage. So whereas MapD will be able to perform faster queries on problems that fit in GPU Ram we are focused on problems that require orders of magnitude more storage than what can fit into GPU RAM. Think dumping a datalake into a database to make it all queryable.
- makmanalp 10y agoSo more out-of-core stuff. Sweet, thank you!