As business moves and changes ever-faster, complex planning scenarios, ad-hoc and what-if analyses all need to approach real time in execution. In today’s global competition, agility is essential. Speed matters. As data volumes expand, response time is increasingly critical to the performance of Enterprise Planning and Business Intelligence (BI) applications.
Many organizations are confronted with poor performance of their BI systems. The result: untapped business potential; drawnout preparation of essential forecasts, simulations, and reports; and slow decision-making in markets that demand increasingly rapid responses.
Added to this situation is the inability of computer systems’ central processing units (CPUs) to process increasingly voluminous and complex data at the speeds necessary for effective response. To avoid this scenario, strategically savvy enterprises are turning to a new and more powerful tool to boost the speed and performance of data analysis in a way that traditional CPUs cannot: graphics processing unit (GPU) accelerators.
Why GPUs Are More Effective Than CPUs
Today’s applications do not scale well with the increasingly huge data volumes that must be processed. Main memory (RAM) access is the principal limiting factor that prevents data-centric applications from scaling appropriately with available CPU processing power. Moreover, GPUs are much more effective than CPUs at scaling. Their development shows enormous generational leaps over the past few years that will continue. Since GPUs were originally designed for video processing, they are optimized for parallel data access, resulting in significantly higher memory bandwidth. This provides the potential to dramatically increase processing speed through unmatched operational thoughput and memory bandwidth.
Importantly, applications accelerated with GPUs benefit from next-generation GPU development out of the box, without changes on the software side. Cost and risk are reduced without impacting performance, and applications are “future-proofed” for the ongoing explosion of big data. (You want to learn more about the architecture of GPUs and how to use them? Read our free brochure to get more in-depth information).
Jedox GPU Accelerator – The EPM Performance Booster
Nearly a decade ago, Jedox was among the first software companies to recognize the potential of using GPU technology with its database application. What started as a research project rapidly evolved into a product.
The GPU Accelerator sits in the heart of Jedox. The Jedox in-memory database server is responsible for handling user queries, managing data, and calculating formulas of theoretically unlimited complexity. The GPU Accelerator “lifts off” from the server, taking action whenever the requested calculation will profit from GPU power.
But the tool is more than a co-processor. It also maintains the complete data storage of accelerated data cubes so that data updates are fully supported and propelled at greater speed. This feature underscores Jedox’s key distinguishing feature, its planning capabilities.
The GPU Accelerator Applied
The high query performance and computing power of the GPU Accelerator opens a host of areas in which to employ Enterprise Planning and BI applications. These include highly complex planning, forecasting, and simulation; real-time monitoring; real-time web analysis; and batch reporting. GPU technology can provide significant advantages in cases where no acceptable support is currently available due to long calculation times The Power of Parallel Processing
With data volumes proliferating alongside the need for rapid (and often complex) analysis of data, the advantages of GPU’s parallel processing over CPU processing are evident. With typical speed increases ranging from 5x to 20x depending on the application, GPU technology becomes a strategic tool to meet the growing challenge of “future data” and boost performance along with speed.
Risk free: Try the GPU Accelerator in the Cloud
With Jedox, there’s no need for specialized hardware to make use of GPU-powered processes: Test-drive the Jedox GPU Accelerator in the cloud. Microsoft Azure provides GPU power in a secure, easy-to-access cloud infrastructure environment where you can see how GPU technology facilitates agile decision-making.
The Jedox GPU Accelerator can be tested on Azure NC instances in combination with the freely available Jedox for Azure Bring Your Own License (BYOL) virtual machine image. Please note that the BYOL version is a pre-configured Jedox installation and requires a valid Jedox license. Moreover, while the Jedox for Azure BYOL virtual machine image is provided for free, running a virtual machine in Azure will be charged by the hoster. Pricing details will be displayed in the Azure portal when selecting the actual Azure instance.
Ready to boost the performance of your analytics application? This Knowledge Base article will guide your through the set-up of the virtual machine, installation of GPU driver, and how to enable the Jedox GPU Accelerator. Have fun!