1 1.1 ross 2 1.1 ross Package Overview for TestFloat Release 2a 3 1.1 ross 4 1.1 ross John R. Hauser 5 1.1 ross 1998 December 16 6 1.1 ross 7 1.1 ross 8 1.1 ross TestFloat is a program for testing that a floating-point implementation 9 1.1 ross conforms to the IEC/IEEE Standard for Binary Floating-Point Arithmetic. 10 1.1 ross TestFloat is distributed in the form of C source code. The TestFloat 11 1.1 ross package actually provides two related programs: 12 1.1 ross 13 1.1 ross -- The `testfloat' program tests a system's floating-point for conformance 14 1.1 ross to the IEC/IEEE Standard. This program uses the SoftFloat software 15 1.1 ross floating-point implementation as a basis for comparison. 16 1.1 ross 17 1.1 ross -- The `testsoftfloat' program tests SoftFloat itself for conformance to 18 1.1 ross the IEC/IEEE Standard. These tests are performed by comparing against a 19 1.1 ross separate, slower software floating-point that is included in the TestFloat 20 1.1 ross package. 21 1.1 ross 22 1.1 ross TestFloat depends on SoftFloat, but SoftFloat is not included in the 23 1.1 ross TestFloat package. SoftFloat can be obtained through the Web page `http:// 24 1.1 ross HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/SoftFloat.html'. 25 1.1 ross 26 1.1 ross TestFloat is documented in three text files: 27 1.1 ross 28 1.1 ross testfloat.txt Documentation for using the TestFloat programs 29 1.1 ross (both `testfloat' and `testsoftfloat'). 30 1.1 ross testfloat-source.txt Documentation for porting and compiling TestFloat. 31 1.1 ross testfloat-history.txt History of major changes to TestFloat. 32 1.1 ross 33 1.1 ross The following file is also provided: 34 1.1 ross 35 1.1 ross systemBugs.txt Information about processor bugs found using 36 1.1 ross TestFloat. 37 1.1 ross 38 1.1 ross Other files in the package comprise the source code for TestFloat. 39 1.1 ross 40 1.1 ross Please be aware that some work is involved in porting this software to other 41 1.1 ross targets. It is not just a matter of getting `make' to complete without 42 1.1 ross error messages. I would have written the code that way if I could, but 43 1.1 ross there are fundamental differences between systems that I can't make go away. 44 1.1 ross You should not attempt to compile the TestFloat sources without first 45 1.1 ross reading `testfloat-source.txt'. 46 1.1 ross 47 1.1 ross At the time of this writing, the most up-to-date information about 48 1.1 ross TestFloat and the latest release can be found at the Web page `http:// 49 1.1 ross HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'. 50 1.1 ross 51