Data continues to grow significantly faster than the rate at which CPUs are improving. While adding more servers seems like a logical solution to the problem, scaling clusters out poses a new set of headaches: new networking bottlenecks, increased operational complexity, physical data center limitations for area, power, and cooling, and more.
At Speedata, we believe that domain-specific hardware accelerators will speed up workloads by factors previously thought impossible. So, we designed an accelerator for Data Analytics workloads to achieve just that.
Speedata’s Analytics Processing Unit (APU) is the first processor designed from the ground-up with the goal of accelerating Data Analytics workloads. The focus on accelerating data analytics, down to the level of the silicon, allows us to offer orders-of-magnitude acceleration compared to the alternatives, all without requiring code changes.
To learn more about our revolutionary architecture,
click on “Book a Demo” and we will email you a copy of our technical whitepaper.
This chart shows the average acceleration factor achieved across several TPC-DS queries when running queries on a CPU, GPU and APU. Individual queries and workloads often show acceleration factors greater than 100x, depending on the computational intensity of the specific workload.
To find out the expected speedup for your specific workloads,
get in touch with us by clicking on “Book a Demo”.