math coprocessors
- To: mathgroup at yoda.ncsa.uiuc.edu
- Subject: math coprocessors
- From: jdelany at sargas.ACS.CalPoly.EDU (Jim Delany)
- Date: Sat, 20 Oct 90 11:30:23 PDT
Richard Fateman's note on the pitfalls of relying on the Motorola floating point coprocessor prompts me to share the example I like to use: Mathematica (sun3.68881) 1.2 (November 6, 1989) [With pre-loaded data] ... In[1]:= x = 2^100; In[2]:= N[ Sin[x] ] Out[2]= 0.973158 In[3]:= N[ Sin[x], 40 ] Out[3]= -0.872183605 The difference between the hardware and software values is almost 2. Does anyone have a sharper example? Jim Delany ( jdelany at cosmos.acs.calpoly.edu ) Math Department, Cal Poly, San Luis Obispo CA 93407