Thomas R. Nicely

Thomas R. Nicely was a mathematician at Lynchburg College (now the University of Lynchburg) in Virginia who, in October 1994, discovered that an Intel Pentium processor was returning incorrect results in a long-running computation of Brun's constant — the sum of the reciprocals of the twin primes — and whose public disclosure of the discrepancy, after Intel's own tech support gave him no answers, forced one of the most consequential chip recalls in the industry's history.

Matters to this vault as the outside researcher whose incidental discovery — not a deliberate hardware test — surfaced the Pentium FDIV bug publicly, and as a case study in primary-source discipline: his own 1994 e-mail, preserved verbatim on his personal site, turned out to settle a dating question that a decade of secondary retellings (and a direct re-read of the field's best-known technical post-mortem) could not.

References

• claim-nicely-twin-prime-project-surfaced-pentium-fdiv-bug — his twin-prime project surfacing the bug in 1994, and his own 30 October 1994 e-mail confirming the exact report-to-Intel (24 October, call reference 51270) and disclosure dates.
• claim-nicely-2011-faq-corroborates-1994-dates-but-is-a-memory-reconstruction — his 2011 retrospective account, corroborating the two headline dates while carrying its own honest caveat about everything else in it.
• claim-intel-475m-charge-pentium-fdiv-recall — the corporate aftermath his disclosure set in motion.