The News: Groq, a compute accelerator supplier for artificial intelligence (AI) and machine learning (ML), announced that its technology has demonstrated the ability to accelerate COVID drug discovery compute speeds 333x faster compared to legacy GPU solutions at Argonne National Laboratory. Read the Groq Press Release here.
Groq Turbocharges COVID Drug Discovery at Argonne National Laboratory
Analyst Take: In early 2020, Argonne and Groq collaborated with a group of scientists to counter the SARS-CoV-2 virus by enabling accelerated drug discovery. Using AI initiatives enabled by GroqChip and the associated software GroqWare, Argonne scientists created ML models of the virus and used them to screen a database including billions of candidate drug molecules. These simulations enabled them to identify high-potential lead compounds for clinical therapy trials.
With this collaboration with the U.S. Department of Energy’s Argonne National Laboratory, I see the possibility of Groq attaining a sales and marketing breakthrough in promoting its GroqChip and GroqWare portfolio. The Argonne collaboration demonstrated that Groq’s technology can accelerate developer velocity enabling a widely distributed set of researchers to work together on updating the virus model and swiftly screening the four billion drugs in their database for effectiveness against the virus. I view this capability as playing a key contributing role in leading the accelerated discovery of antiviral drug combination candidates critical to battling the pandemic.
Argonne used GroqChip accelerator, a high-performance chip with 750Tops @900 MHz and 80 terabytes/sec of memory bandwidth, providing an architecture that enables more models to fit within its unified, coherent on-chip memory. In addition, Groq TruePoint technology maintains accuracy while taking advantage of lower precision, boosting the performance of Argonne’s workload.
By using GroqWare, Argonne was able to streamline its development flow that generated the ease of use needed to produce higher developer velocity. In turn, the accelerated platform enables scientists to iterate rapidly with new approaches and new models. Moreover, Groq’s deterministic architecture, able to extend single core performance across multiple chips, enables GroqWare to deliver consistent predictable performance across an array of ML models while eliminating overhead for context switching.
From my perspective, these capabilities taken together liberate researchers from the traditional slow, dynamic profiling process after deployment needed to optimize hardware that is subsequently implemented in production. For Argonne, accelerated AI capabilities rapidly adapt to and accommodate chemical structures that have not been seen previously and that have never been synthesized and do not exist in nature. As a result, GroqChip acceleration of AI gave Argonne the speed and flexibility that physics-based computation would have great difficulty achieving.
Groq Demonstrates Throughput Advantage over NVIDIA GPU Solutions at Argonne
At Argonne, Groq showed its solution delivered up to 333x performance in comparison to Argonne’s existing NVIDIA GPU solution, reducing the time to solution for these Argonne algorithms from days to minutes:
Given the Argonne National Laboratory’s well-established reputation and pedigree, I foresee Groq gaining swift market credibility in touting the performance (and security) credentials of its GroqChip product, gaining key differentiation against GPU market titan NVIDIA, the GPU-related ambitions of Intel and AMD, as well as upstarts like Graphcore and Esperanto.
Key Takeaways on Groq’s Argonne National Laboratory Collaboration
I expect that Groq will gain substantial sales and marketing momentum from its collaboration with Argonne National Laboratory in turbocharging COVID drug discoveries. I also anticipate that Groq will need to replicate and highlight similar performance milestones across a broad array of use cases and applications as well as industry verticals to catalyze more ecosystem-wide consideration of its GroqChip/GroqWare/Groq True Point proposition. Overall, Groq validates that it is developing disruptive technology that can address rapidly expanding AI and ML opportunities with the ability to drive powerful use cases such as COVID antiviral drug discovery.
Disclosure: Futurum Research is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.
Other insights from Futurum Research:
Image Credit: ttownmedia.com
The original version of this article was first published on Futurum Research.