home *** CD-ROM | disk | FTP | other *** search
- Description and use of this utility.
-
- Michael Joseph
- Email: qmick@pipeline.com
-
- 2 Files included are:
-
- P5_TEST.EXE The test utility written in Clipper 5.2
- README.TXT This document file.
-
-
- This utility is FREEWARE you do not have to pay anybody for it.
-
- You may distribute it as much as you want, you must only include this
- readme file in unaltered condition.
-
- This utility is designed to run on any IBM compatible PC and has no
- special memory requirements, it is however expressly designed to
- test the Floating Point Unit (FPU) in the Intel PENTIUM processor.
-
- To install:
-
- a) Unzip the P5.ZIP file in any directory.
-
- b) Read this file.
-
- c) From the command line and in the directory where the you unzipped
- the P5.ZIP file type P5_TEST and then hit the Enter key.
-
- This will execute the FPU test and instantly return a result with
- a description of what the result indicates.
-
- You may want to note the result somewhere and then (unless you wish
- to run the test again) delete the P5_TEST.EXE and README.TXT files.
- _________________________________________________________________________
-
-
- Following is a direct quote from 'The New York Times' November 24
- article on page D5 (the business section) on which this utility is
- directly based.
-
- Begin quote from 'The New York Times' November 24 page D5
-
- +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
-
-
- Close, but Not Close Enough
-
- The owners of computers that use Intel's Pentium microprocessors
- have found that the chips sometimes do not perform division
- calculations accurately enough.
-
- The problems arise when the chip has to round a number in a
- preliminary calculation to get the final result, a task that all
- processors normally perform. In these cases, however, the
- Pentium's figures are exact to only 5 digits, not 16, not 16, as are those of
- other computer processors. The Pentium's error, while small, can
- be 10 billion times as large as those of most chips.
-
- Here is an example of the way the imprecise rounding changes the
- results of a calculation and the way the deviation from the expected
- result is calculated.
-
- PROBLEM
-
- 4,195,835-[(4,195,835/3,145,727) X 3,145,727]
-
- CORRECT CALCULATION
-
- = 4,195,835-[(1.3338204) X 3,145,727] = 0
-
- PENTIUMS CALCULATION
-
- = 4,195,835-[(1.3337391) X 3,145,727] = 256
-
- DEVIATION
-
- 256/4,195,835=6.1 X 10E-5, or 61/100,000
-
- Source: Cleve Moller, The Mathworks Inc.
-
- +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
-
- End Quote
-
- This utility executes the calculation shown on the first line,
-
- If the result is not zero it will tell you so.
-
- If the result is zero it will tell you also.
-
- The author would like to state for the record:
-
- Should your Pentium not produce zero as an answer when you run this utility
- as described above, this only indicates the FPU is not performing
- this calculation correctly and by inference, others of the same type.
-
- What you choose to do should your Pentium produce an answer other than zero
- is up to you.
-
-
-
-
-
-
-
-
-
-
-
-