From cperciva@sfu.ca Wed Jan 7 18:12 EST 1998 Return-Path: Received: from banzai.cecm.sfu.ca by caillou (5.x/SMI-SVR4) id AA20053; Wed, 7 Jan 1998 18:12:16 -0500 Received: from rm-rstar.sfu.ca (root@rm-rstar.sfu.ca [142.58.120.21]) by banzai.cecm.sfu.ca (8.8.7/8.8.7) with ESMTP id PAA25111 for ; Wed, 7 Jan 1998 15:12:18 -0800 (PST) Received: from fraser.sfu.ca (fraser.sfu.ca [192.168.0.101]) by rm-rstar.sfu.ca (8.8.7/8.8.7/SFU-4.1H) with SMTP id PAA00272 for ; Wed, 7 Jan 1998 15:12:14 -0800 (PST) Received: by fraser.sfu.ca (950413.SGI.8.6.12/SFU-2.6C) id PAA07978 for isc@cecm.sfu.ca (from cperciva@sfu.ca); Wed, 7 Jan 1998 15:12:12 -0800 From: Colin Andrew Percival Message-Id: <199801072312.PAA07978@fraser.sfu.ca> Subject: Pi calculation To: isc@cecm.sfu.ca Date: Wed, 7 Jan 1998 15:12:11 -0800 (PST) X-Mailer: ELM [version 2.4 PL24] Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII Content-Length: 1328 Status: OR It's not actually a record, but you might want to add it to your list anyway. Using the BBP formula, I calculated that from the two hundred billionth (2*10^11) hexit (8*10^11 th bit) until the two hundred billion and twentieth (2*10^11+20) hexit (8*10^11+83 th bit), Pi is 3E6FBDAC38A97197785ED. Two calculations, one starting at 2*10^11, and the other starting at 2*10^11-1, calculated this. The calculation starting at 2*10^11 computed it was between 3E6FBDAC38A97197785ED647666822B1 and 3E6FBDAC38A97197785ED701AA1F6329, while the calculation at 2*10^11-1 computed it was between 3E6FBDAC38A97197785ED0D269603BD0 and 3E6FBDAC38A97197785EDC76A4D44350. These calculations were performed concurrently between September 1997 and January 1998, using idle time slices under Windows 95 on a Pentium 200 (87.1%) and a Pentium 133 (12.9%). The terms were added in ranges of 2^34, and the two calculations were checked against each other. The two calculations matched exactly (apart from rounding errors). This was actually a test run to make sure that my code for collecting results from two computers was working properly; for the next run I'll use a slightly faster implementation (40% faster) and Bellard's formula (40% faster) and more computers to calculate the trillionth hexit (four trillionth bit). Colin Percival