Skip to main content Skip to Footer

LATEST THINKING


Become a GPU “powerhouse”

Data fuels the digital economy. The leaders harness growth processing units (GPUs) to attain desired outcomes and get ahead.

Accelerate and scale

Data, data everywhere… Business needs scalable solutions to process the ever-growing volume and harvest useful insights.

GPUs—computer chips that rapidly perform mathematical calculations—offer an answer.

Today’s GPUs power an array of artificial intelligence (AI)-driven tools, such as data analytics, image processing and deep learning.

But how does a given business effectively apply GPUs to overcome challenges and identify opportunities?

Success requires IT and the rest of the business to talk to each other in the same language. Through Accenture Labs’ research, we see ways to capitalize on GPU acceleration.

Meet today’s GPU

Gaming industry companies were among the early adopters, using GPU chips to run visually-intensive video games on computers, consoles and mobile phones.

The US military also recognized the potential of GPUs to quickly process high-resolution satellite imagery and kick-start AI research.

Over the past decade, GPUs surpassed their central processing unit (CPU) counterparts in a few key areas:

Processing power: GPUs calculate faster than CPUs, which improves the ability to build better analytical models and gain insights.
Memory bandwidth: At about 732 gigabytes currently, GPUs have significantly more memory bandwidth than CPUs (102 GBs), and thus faster processing capacity.
Efficiency: GPUs are said to be 10 times more efficient than CPUs in both performance versus power consumption and performance versus cost.

Acquiring, unlocking GPU capability

The problem-solving capability of CPUs differs from GPUs, so companies will need to continue using both, depending on the nature of the task to be performed.

Those ready for GPU acceleration have two options: Delivery through the cloud, or building on-premise.

Every major cloud provider has recognized the demand for GPUs and responded with cost-effective options.

Regardless of how GPUs are sourced, you can get started by identifying complex and time-consuming data operations and processes that could be GPU-accelerated.

Importantly, make sure you price out how and where to access GPU processing capacity and determine the return on investment for each approach.

Seek an edge

Business leaders looking for that next competitive edge should consider GPU acceleration.

Whether used to speed big data processing or prepare for a future of AI-related technologies, the opportunities for generating better outcomes through GPUs are a smart bet.

Read more about how GPU solutions can power growth and how to map out your next steps.

SUGGESTED CONTENT