Accelerating Drug Development (R&D) with High Performance Compute
Pharmaceutical companies are never without pressure – pressure to be first to market, pressure to keep treatments affordable and pressure for the next “blockbuster” drug. With the average of 10 years from drug discovery to approval, the stakes are high. Disrupting the economics of drug development by deploying innovative technologies is vital to accelerate and expand research and development.
The key is the ability to acquire, process and transform data into actionable insights. Pharma is leaning in on technologies such as HPC and AI (High Performance Compute and Artificial Intelligence) to delve into immense amounts of data, find new medical insights and use these insights to develop new treatments. Let’s take a look at how organizations are accelerating drug development with high performance compute.
Applying Artificial Intelligence in Drug Development (Backed by High Performance Compute)
Just as previous generations developed machines to mass-manufacture dr, the new generation of scientists rely on AI performance to speed development and lower costs. Development and testing of new drugs creates petabytes of data at each stage, and that data in turn needs to be analyzed for useful information. Machine Learning in AI allows you to take that data, look for patterns and call out those that matter. Before AI, that was all done manually. While this is one example, as seen below AI can be utilized in multiple stages of the pharmaceutical development path for maximum impact.
AI and HPC : Why AI Needs High Performance Compute
Because AI is capable of processing data very quickly, it needs hardware that can keep up by processing trillions of calculations per second or more. While AI and HPC each provide increased speed and processing, they need each other for maximum results. While HPC supports AI models better than traditional systems, AI can be used to line up and optimize processing of your workloads.
How HPC Enables AI Performance
HPC clusters typically have 16 to 64 nodes with each running a minimum of 2 CPU’s. High Performance Compute (such as HPE Superdome Flex) therefore provides higher processing power than traditional infrastructure. HPC also has faster memory and storage resources contained within each node enabling higher speeds and larger capacity than traditional systems.
Let’s look at a few specific examples where HPC can help AI’s contribution to pharmaceutical product development.
- Specialized Processors – GPU’s like NVIDIA can be utilized to process specific AI algorithms. For example, they are often used to help analyze images and taking information from those images for studies.
- Data Growth – As data continues to expand, creating the space needed to properly process that data becomes important. AI itself needs a lot of processing power and adding data expands that requirement. Traditional architectures simply cannot keep up.
- Parallel Processing – Within an HPC System, a node is a self-contained unit. Because of this, you can use these units working together or create different tasks for each, allowing both improved efficiency and speed.
- Cost – HPC can help with costs in several ways
- HPC scale-up architecture allows you to more easily expand compute power in blocks, when extra processing is needed.
- HPC systems like HPE Superdome Flex or others can be managed within IaaS and Hybrid Cloud environments, so you can pay as you go and avoid upfront capital costs. This is especially helpful for organizations that want to leverage compute power without the need to purchase large amounts of capital equipment. Take a look at our blog for other Benefits of HPE Superdome Flex.
Simply Better Collaboration: HPC and AI Convergence
Drug development project requires collaborators that are outside your four walls, from academic partnerships, contract research organizations, and multiple sites within your own company. For pharma teams to collaborate, you need to secure and easily share data. HPC allows you to easily provision technology and create output results in a desired format needed by other collaborators. In addition, HPC brings more built-in security like silicon root of trust and machine learning which help you stay one step ahead of information security threats.
AI may not actually be alive but if it were, its perfect match would be HPC. The two technologies can aid each other and interrupt the trend of Eroom’s Law. HPC gives AI the speed it needs to increase intelligence and AI gives HPC the intelligence it needs to make smarter workload and security decisions. Those interested in staying competitive in pharmaceutical research and development will pay attention to evolving HPC cluster technologies and cluster management, to ensure that their infrastructure keeps them on top of the massive data needs of the drug development process. Comport has experts to facilitate infrastructure design, implementation and management to support these mission-critical advancements. Contact us today to discuss your latest project and accelerate your drug development with high performance compute!
“Epic, “Cogito”, and “Hyperspace” are registered trademarks of Epic Systems Corporation.
“IRIS” is a registered trademark of InterSystems Corporation.
Author:Bill Flatley, Field CTO for Healthcare
Bill is responsible for technical strategies and recommendations for Comport’s Healthcare clients. His extensive experience includes four healthcare systems in leadership roles supporting Clinical Applications, Digital Health, and Office of the CIO as the primary liaison between IT and the business.
