CompuServe Thread

#Pentium bug

4 messages in this thread
#174336From: Grant BoucherNov 26, 1994 4:40 PM
Gotta love that Pentium math bug…lost all scientific developer confidence… "Oh, it's fixed now…" yeah right. Grant Boucher – Amblin Imaging "It's hard to believe grown-ups do this for a living."
#174452From: E.R. BujtasNov 27, 1994 7:14 AM
Grant: This is NOT the first time that Intel had floating point problems. They seem to have a history of screwing-up their FPU's. Recall the early 486's. ..Andy
#174679From: Malcolm O'BrienNov 28, 1994 2:54 AM
Andy, Hey! The 287 was wonky too! I remember an article in Byte about how certain 287s would fail slowly bit by bit over months until the error got large enough for you to notice it. By then you'd been calcing all your Lotus sheets wrong for a month or more. Maybe it was a _smart_ move to step on the FPU to create the SX 486s! 😀 Malcolm
#174690From: John GagerNov 28, 1994 3:59 AM
Hehe!! I saw that on the news the other day myself. Now Intel will have to add to their advertisements "Intel's Pentium Processor: When computational accuracy is no big deal." <g> // John (Warp 4040 & Emplant Quadra 840) \X/ Mercury@ins.infonet.net