Few queries about MapD



Since we are using CPU-Based installation of MapD, I just wanted to know few things regarding this ,

  1. What will be the time constraints when loading data from HDFS or any other database ?
  2. I feel MapD is acting like an extra layer between database i.e Hadoop and visualization tool, so will it be the cost effective?
  3. Can we control when to use cpu or gpu power to access data ?
  4. Does MapD immerse provide feature of automatic refresh for dashboard?
  5. Can we have real time data movement to MapD and immerse accordingly ?
  6. What will be the approx cost for GPU based server?

Any help appreciated.



  1. This is a pretty general question, so assuming bulk movement and a reasonable level of machinery involved at both ends of process, the likely area taking up time will be Network time to transfer data from one instance to the other assuming 1Gbe network. If you have faster network the bottleneck will likely move to extraction process.

  2. I think we would need to discuss use case more before this conversation can go any where useful.

  3. Yes, it is possible to direct queries to cpu or gpu on demand

  4. yes

  5. yes

6, you will need to have that conversation with sale. sales@mapd.com