MapD, an NVIDIA-funded startup, has created a new type of database to handle terabytes of data and impressive visualizations that relies on GPUs instead of the traditional AMD or Intel chips. Todd Mostak, Founder and CEO of MapD, decided to build a SQL database that relies on graphics processors when he was completing his thesis that analyzed a huge number of tweets.
MapD was the winner of GPU Technology Conference Early Stage Challenge $100K Prize in 2014. The startup has built a high-speed GPU in-memory database that brings interactivity to big data. It can, for example, track more than a billion tweets worldwide at a time – and provide real-time visual analysis of the data. They have a Twitter demo on their website.
To date, they have received $2 million in funding from Google Ventures and NVIDIA and they are currently raising a Series A round of nearly $10 million.
“With GPUs getting more memory, Mostak realized that the Intel and AMD part of the database can now take a backseat to the graphics processors because the GPUs finally have the memory capabilities to store the data as its being processed. MapD machines still use a CPU for some processing, but it’s a secondary processor instead of the primary one, and as the memory in GPUs increases, the CPUs will recede further.”