aiqbench analysis
Cognition's use of an AI agent (Devin) to engineer a specialized GPU lattice siever has accelerated the factorization of RSA-260, achieving a new benchmark at significantly lower cost than previous CPU-based methods. This development highlights how generative AI is directly optimizing hardware-specific implementations of classical algorithms like GNFS, rather than proposing new mathematical breakthroughs in number theory. For the industry, this demonstrates a practical pathway where AI agents can bridge gaps between domain experts to accelerate computational records in cryptography and high-performance computing.
Cognition GPU lattice siever factors RSA-260 at 10x lower cost
Cognition’s research team factored RSA-260, a 260 digit number, setting a new record for the largest publicly solved RSA Factoring Challenge problem and achieving a factorization at 10x lower cost than the previous state of the art, using a GPU lattice siever. RSA-260 sets a benchmark for breaking the RSA cryptosystem, exceeding the previous record of RSA-250 set in February 2020.
Cognition’s research team recently factored RSA-260, a 260-digit number, establishing a new record for the largest publicly solved RSA Factoring Challenge problem and surpassing the previous RSA-250 solved in February 2020, the company says. This feat was achieved using a new GPU lattice siever built by Cognition’s “Devins,” enabling factoring at ten times lower cost than previously public methods. The team estimates hyperscalers could now factor RSA-1024 numbers for around $30 million. Cognition’s development of a GPU lattice siever enabled the factorization of RSA-260, a 260-digit number, by significantly reducing computational costs compared to previous methods. The team achieved this feat not through algorithmic breakthroughs in number theory, but by focusing on hardware acceleration and efficient implementation of the well-established general number field sieve, or GNFS, according to Cognition. Prior public records for GNFS factorization relied exclusively on central processing units for the computationally intensive lattice sieving stage; Cognition’s innovation lies in successfully adapting this process for graphics processing units. The impetus for this project stemmed from a personal interest in factoring numbers, according to the lead researcher, who dedicated several weeks to optimizing a job scheduler for disaggregated compute resources. This led to the prompt given to Cognition’s…
Excerpt shown under the publisher's RSS terms · 199 words · full text and figures are on the publisher's site
